# Federico Rao — full text

> Federico Rao is an independent digital consultant and full-stack developer based in Italy and working with clients worldwide. He designs and builds conversion-focused websites and web applications, AI and workflow automations, CRM and lead-funnel systems, and technical SEO foundations for service businesses, professional firms, and small product teams.

Projects are implemented end to end (architecture, interface, integrations, deployment, and handover) rather than split across separate suppliers. Typical clients already have customers and a defined offer and need the website, enquiry funnel, and automation around that offer to work better.

The site publishes only inspectable evidence: case studies describe what was built, and rankings, revenue, traffic, or time savings are not claimed without a measured baseline. Certifications, awards, or media appearances are not claimed unless independently verifiable.

Languages: English is the default and lives on the unprefixed URLs. Every marketing page also has an Italian (Italiano) version under /it/ with the same path (for example /it/about/ and /it/projects/studio-mazzoli/), linked with reciprocal hreflang; the Italiano section below lists them. Approved Italian translations of articles, when they exist, are published under /it/blog/. Enquiries are accepted in English and Italian.

Contact: info@federicorao.com or the form at https://federicorao.com/contact/.

This file contains the English editorial text of the indexable pages at https://federicorao.com/ (Italian versions under https://federicorao.com/it/ are listed in llms.txt). The curated index is https://federicorao.com/llms.txt.

## Federico Rao - Digital Consultant and Full-Stack Developer

URL: https://federicorao.com/

Conversion-focused websites, AI automations, and digital systems for service businesses. Based in Italy, operating globally.

### What this practice builds

Federico Rao is an independent developer and digital consultant working on three connected surfaces: the website that explains an offer, the funnel that captures qualified demand, and the automation layer that removes manual steps between an enquiry and a delivered service. Projects are implemented end to end, covering architecture, interface, integrations, deployment, and handover, rather than being split across separate suppliers who each own a fragment. The usual context is a service business, clinic, professional firm, agency, or small product team that already has customers and a defined offer, and needs the digital system around that offer to stop losing attention, enquiries, and operator time.

### Where a project usually starts

Most engagements begin from one observable problem: a site that does not explain what the business sells, a form that produces unqualified enquiries, a manual process repeated every week, or a set of pages that search engines never select. The first pass records what users and crawlers actually receive, including response codes, initial HTML, rendered content, journey steps, tracking, and the existing conversion path. Requirements are then separated into necessary, useful, and premature, so implementation follows the decision that has to improve instead of a feature list assembled before anyone looked at the current system.

### Custom web development

Development work covers marketing sites, web applications, storefronts, internal tools, and the content model behind them. Typical decisions concern rendering strategy, route structure, accessible interaction, image handling, form validation, third-party integrations, editing rights, and deployment. The output is meant to be maintainable by its owner: content edited without a developer, components reused rather than duplicated, and a build that can be inspected, tested, and redeployed. Stack choices such as React, Node.js, Python, or a headless CMS follow the required behaviour, hosting constraints, and operating cost, not familiarity or fashion.

### Automation and AI integration

Automation is applied to work that is repetitive, rule-describable, and already understood by the person doing it. Common examples include quote funnels, CRM routing, document generation, enrichment of inbound leads, scheduled reporting, and handoffs between tools that currently rely on copy and paste. Before a workflow is automated its inputs, owners, exception paths, approval points, and failure behaviour are written down, because an unattended process that fails silently is worse than a manual one. Language models are used where a draft, classification, or extraction is genuinely useful, and their output stays reviewable rather than published unattended.

### Technical SEO and search experience

Search work starts from what a crawler receives before JavaScript executes: status code, canonical URL, robots directive, title, description, headings, structured data, internal links, and meaningful text in the initial response. From there it covers indexation, duplicate routes, rendering parity, page type against search intent, information architecture, performance, and the measurement needed to tell whether pages are discovered, selected, and read. Technical eligibility is treated as a precondition rather than a promise: it removes the reasons a page cannot rank without claiming that a ranking will follow.

### Sector-specific starting points

Some sectors have a recurring shape, so the site keeps dedicated pages describing what changes in each: veterinary clinics, dental practices, law firms, e-commerce operations, fitness and wellness businesses, real-estate teams, restaurants, and Web3 or crypto products. Each of those pages covers the buying moment, the pages that carry the decision, the discovery surfaces that matter locally or in a marketplace, the follow-up workflow after an enquiry, and the regulatory or professional constraints that limit what may be claimed publicly. They exist to make scope concrete before a conversation, not to advertise industry expertise in general terms.

### Evidence, ownership, and limits

Useful evidence is inspectable: crawlable HTML, working interfaces, route output, documented content models, schemas, tests, dashboards, and public repositories where they exist. Client results that depend on confidential data are not published as proof, and rankings, revenue, traffic, or time savings are not asserted without a measured baseline. Delivery includes handover: responsibilities, routine editing, deployment, monitoring, credentials, data retention, and escalation should be clear before a system becomes business-critical. Software cannot repair an undefined offer, absent internal ownership, or a process that changes on every execution.

### How to start a conversation

A useful first message identifies the business, the current system, the outcome that should become observable, the constraint that matters most, and who decides. Access to analytics, Search Console, hosting, or an existing repository helps when it is available, and its absence is not a blocker for a first review. The reply is a scoped direction rather than an immediate quotation, because effort depends on what already exists. Requests that only ask for volume, thin pages at scale, or manufactured authority are declined instead of being priced.

## About Federico Rao - Digital Consultant

URL: https://federicorao.com/about/

About Federico Rao, a digital consultant specializing in web development, AI automation, CRM systems, and conversion-focused digital infrastructure.

### What the work covers

The practice connects full-stack development, user experience, technical SEO, automation, analytics, CRM workflows, and content operations. The connecting principle is not a particular tool. It is the design of a digital system that explains an offer clearly, captures the right actions, preserves usable data, and can be maintained after launch.

### How projects start

A project starts with business context, audience, current journey, constraints, ownership, and the outcome that needs to become observable. Existing tools and content are reviewed before proposing a rebuild. Requirements are separated into necessary, useful, and premature so implementation effort follows the actual decision rather than a fashionable technology.

### How work is evaluated

Relevant evidence can include crawlable HTML, page performance, accessibility, error rates, successful form completion, qualified enquiries, workflow completion, manual touches, and maintainability. The appropriate measures depend on the project. Traffic, rankings, revenue, or time savings are not guaranteed and should not be presented without a measured baseline.

### Editorial responsibility

Public articles identify an author and can include purpose, preparation method, source links, comparison disclosures, and AI-assistance disclosures. Generated content remains a draft until reviewed. Unsupported credentials, awards, appearances, tests, customer results, and product experience are excluded rather than used to manufacture authority.

### Limits and fit

Software cannot repair an undefined offer, absent operational ownership, unreliable source data, or a process that changes on every execution. The strongest fit is a project with a real problem, a responsible decision maker, accessible constraints, and willingness to measure what happens after implementation.

### Artifacts instead of vague claims

The most useful proof is inspectable: public project pages, working interfaces, route output, documented content models, schemas, tests, dashboards, repositories when they are public, and source-linked articles. Confidential work cannot always be exposed, so the site distinguishes a description of capability from independently verifiable performance evidence.

### Collaboration and ownership

Decisions need an owner on both sides. Access to existing tools, timely review, content approval, legal constraints, and operational knowledge affect delivery as much as code. Handover is part of implementation: responsibilities, routine editing, deployment, monitoring, credentials, data retention, and failure escalation should be clear before a system becomes business-critical.

### Technology choices

React, Node.js, Python, cloud platforms, databases, automation tools, commerce systems, and design software are implementation options rather than the offer itself. A technology is selected for the required behavior, team constraints, ecosystem, operating cost, security boundary, and maintainability. Familiarity does not justify adding infrastructure the project does not need.

## Federico Rao - Author and Digital Consultant

URL: https://federicorao.com/author/federico-rao/

Author profile for Federico Rao covering digital consulting, web development, AI automation, technical SEO, editorial methods, and disclosures.

### Areas of work

Work spans websites and web applications, content architecture, frontend performance, workflow automation, AI-assisted operations, CRM integrations, analytics, technical SEO, search-native rendering, structured data, and editorial systems. These areas are connected through implementation rather than presented as unrelated advisory categories.

### Editorial method

Articles begin with a stated reader problem and intended decision. Source links, limitations, comparison relationships, and AI assistance are disclosed when relevant. Time-sensitive product or market claims should be checked against dated first-party sources. A citation supports a specific claim; it does not automatically validate every conclusion around it.

### Review and correction

Automation can help organize research, produce a brief, suggest internal links, check required fields, and prepare a draft. Publication remains an editorial decision. Correction requests should identify the affected URL, disputed statement, and supporting source so the material can be reviewed and its modification date updated where necessary.

### Identity and authority

Public identity references are limited to profiles controlled by Federico Rao. No certification, award, publication, media appearance, customer result, benchmark, or hands-on product test is claimed without a verifiable source. A transparent absence of evidence is preferred to a stronger but unsupported biography.

### What readers can inspect

Service pages describe fit, process, deliverables, limits, and measurement. Project pages show the categories of systems represented on the site. The editorial policy documents AI and sourcing rules. The press record is reserved for independently verifiable appearances and remains explicit when no entry is available.

### Who, how, and why

The named author answers who is responsible for the page. Editorial purpose and method fields explain why a guide exists and how it was prepared. Sources and disclosure fields provide context where readers would reasonably expect it. These elements make responsibility visible; they are not a substitute for accurate, useful, and original material.

### Subject boundaries

Technical articles can explain implementation patterns, trade-offs, and observable behavior. They do not replace legal, medical, financial, security, accessibility, or compliance advice from an appropriately qualified professional. When a decision crosses one of those boundaries, the limitation should be stated and the relevant review treated as a dependency.

### Commercial context

The site offers digital consulting services, so some educational content can connect to a relevant service or contact path. That commercial purpose is not hidden. Comparisons require relationship disclosures, and an article should remain useful even when a reader does not request a proposal or purchase a referenced product.

## Editorial Policy and AI Disclosure

URL: https://federicorao.com/editorial-policy/

Editorial policy for AI-assisted drafts, human review, sources, corrections, comparisons, affiliations, and prohibited fabricated evidence.

### AI supports drafts, not accountability

AI may help with outlines, synthesis, formatting, metadata, internal-link suggestions, and first drafts. Imported or generated material remains a draft. Editorial responsibility, factual review, approval, and publication remain attributed to the named author or editor. Automation is treated as a production aid rather than evidence of expertise or accuracy.

### Sources and claims

Specific, time-sensitive, comparative, legal, medical, financial, and product claims should be checked against accessible and preferably primary sources. A source link does not automatically validate the interpretation around it. Unsupported claims are removed, narrowed, or identified as opinion, and a missing source is not replaced with a plausible-looking citation.

### Quality and duplication controls

The workflow can require editorial fields, apply quality scores, compare similarity, request source data, suggest internal links, and block automatic publication. These controls reduce avoidable mistakes but cannot replace reading the complete page, matching search intent, checking factual claims, and deciding whether the draft adds distinct value.

### Comparisons and relationships

Comparisons state the intended use case, criteria, evaluation date where relevant, first-party source links, limitations, and known commercial relationships. Affiliate links, sponsorships, or client relationships must be disclosed. Products are not ranked from invented testing, reviews, accounts, or usage experience.

### Corrections and prohibited evidence

Correction requests can identify the URL, disputed statement, and supporting source through the contact page. Material corrections should update the page and modification date. Fake reviews, customer results, accounts, interviews, credentials, awards, citations, tests, and media appearances are prohibited.

### Publication and updates

Draft status is not a public endorsement. A page becomes eligible for publication only after its required fields and review are complete. Updates should correct or improve material information rather than change a date alone. When a page can no longer be supported, it should be revised, consolidated, redirected, archived, or removed from indexing according to its remaining value.

## Press and Appearances

URL: https://federicorao.com/press/

Verified interviews, talks, podcasts, guest articles, and media references involving Federico Rao, with sources and disclosures for every entry.

### Current record

No verified appearance is listed at this time. The absence is stated directly because an empty record is more accurate than invented authority. A future entry must have a public source, identifiable publisher or organizer, date, format, topic, and direct URL before it appears here.

### Evidence standard

A self-authored social post, an untraceable logo, or a private claim is not treated as independent coverage. The linked page should show enough context to verify participation. If the material is removed or substantially changed, the entry should be reviewed and may be archived or annotated.

### Commercial transparency

Sponsored appearances, paid placements, affiliate relationships, client relationships, and promotional arrangements must be disclosed beside the relevant entry. A commercial relationship does not automatically invalidate a source, but readers need that context to evaluate the reference.

### What will never be added

Fabricated interviews, podcast appearances, event participation, publication logos, quotations, awards, nominations, or third-party praise are excluded. This page is not populated from AI-generated suggestions unless each suggested item is independently verified against the original publisher.

### Requests and corrections

Requests for an interview, technical comment, collaboration, or correction can be sent through the contact page. A useful message includes the outlet or event, proposed topic, format, expected date, publication terms, and a source link when correcting an existing record.

### How entries will be presented

A verified entry will identify the title, publisher or organizer, format, subject, publication or event date, direct source, and any necessary disclosure. Descriptions will summarize the context without turning participation into an endorsement. Broken links, changed titles, or disputed attribution will be reviewed rather than silently preserved.

### Why the empty state is public

Authority pages are easy to fill with ambiguous logos and synthetic biographies. Publishing the inclusion standard before any appearance exists makes the future record auditable and gives readers a stable place to check claims. It also prevents an AI-assisted content workflow from treating suggested appearances as facts merely because they sound plausible.

## Digital Services for Business

URL: https://federicorao.com/services/

Web development, AI automation, CRM, lead funnels, and digital consulting services for service businesses and growth-stage teams.

### Three services, one system

Web development, automation and AI, and technical SEO are sold separately but designed as one system. A site that explains an offer clearly still loses enquiries if nothing routes them; an automated funnel produces noise if the pages feeding it attract the wrong visitor; and search work is wasted on pages that cannot answer the question that generated the visit. Engagements normally start with whichever layer is currently blocking the others, and the remaining work is proposed only when there is a reason to do it. Nothing here requires buying all three.

### When web development is the right entry point

Choose development when the interface itself is the constraint: the offer is unclear on the page, the site cannot be edited without a developer, the stack is unmaintained, performance or accessibility blocks real use, or a required behaviour simply does not exist yet. Work covers information architecture, component design, content modelling, integrations, forms, admin surfaces, deployment, and handover documentation. The deliverable is a system whose owner can publish, measure, and extend it without a retainer, with the code and content model inspectable rather than locked inside a proprietary builder.

### When automation is the right entry point

Choose automation when demand already exists but the operation around it is manual: enquiries retyped between tools, quotes assembled by hand, follow-up depending on someone remembering, reporting rebuilt every month. The first step is a written map of the current process, its owners, its exceptions, and the point where a human must still approve. Automation then replaces the mechanical steps and leaves judgement where judgement belongs. Monitoring and a defined failure path are part of the build, because an unattended workflow that breaks quietly costs more than the manual version it replaced.

### When technical SEO is the right entry point

Choose search work when pages exist and are not being discovered, selected, or understood: indexation problems, rendering gaps, duplicate or competing routes, page types that do not match the intent behind the query, weak internal linking, or measurement that cannot answer whether anything improved. The review begins with the raw response a crawler receives and continues into architecture, intent mapping, structured data, performance, and reporting. Findings are tied to a URL, file, response, or observable search signal, so each recommendation can be verified rather than accepted on authority.

### Scope, ownership, and handover

Every engagement defines who owns the decision, what access is required, which constraints are fixed, and what will be observably different at the end. Content approval, legal review, and operational knowledge affect delivery as much as implementation, so they are scheduled rather than assumed. Handover covers credentials, deployment, routine editing, monitoring, data retention, and escalation. Ongoing maintenance can be agreed separately and is described in terms of what is covered, not as a promise of uninterrupted availability across every browser, provider, network, and third-party integration.

### Sector starting points

Where a sector has a recurring buying pattern, a dedicated page describes it: veterinary clinics, dental practices, law firms, e-commerce operations, fitness and wellness businesses, real-estate teams, restaurants, and Web3 products. Those pages set out the decision moment, the pages that carry it, the discovery surfaces that matter, the follow-up workflow, and the professional or regulatory limits on public claims. They are a faster way to scope a first conversation than a generic capability list, and they are written to be disagreed with when the specific business works differently.

### Limits and non-fit

These services do not include guaranteed rankings, guaranteed lead volume, purchased links, fabricated reviews or testimonials, undisclosed commercial relationships, or content produced at scale without editorial review. They also do not fix an offer that has not been defined, data that cannot be trusted, or a process nobody inside the business owns. Where those are the real constraint, the honest answer is that implementation should wait. A scoped review is the usual first step, and it can conclude that no build is currently justified.

## Custom Web Development

URL: https://federicorao.com/services/web-development/

Custom websites, web applications, SaaS platforms, and e-commerce builds designed for speed, SEO, conversion, and long-term maintainability.

### Who this is for

Custom web development is appropriate for service businesses, professional firms, product teams, and founders whose offer, workflow, integrations, or content model cannot be represented reliably by an off-the-shelf theme. A clear owner, accessible requirements, real content, and an outcome that can be observed make the engagement stronger.

### The problem being solved

Typical constraints include unclear user journeys, slow pages, duplicated content, inaccessible interactions, fragile plugins, manual publishing, disconnected forms, missing analytics, and architecture that makes every change expensive. The problem is defined before the framework because a modern stack does not compensate for an undefined audience or offer.

### What can be delivered

Scope may include information architecture, wireframes, interface implementation, responsive behavior, accessibility, CMS content models, authentication, dashboards, e-commerce, third-party integrations, analytics, technical SEO, deployment configuration, monitoring, testing, documentation, and handover. Deliverables are selected around the project rather than bundled as decorative features.

### Discovery and architecture

Business goals, audiences, journeys, content ownership, data, integrations, constraints, and acceptance criteria are mapped first. Existing systems are reviewed before proposing replacement. Necessary requirements are separated from useful enhancements and premature complexity, creating an architecture that can be explained, tested, and maintained.

### Design and implementation

Page hierarchy and critical interactions are validated before detailed implementation. Components are built with semantic markup, keyboard behavior, responsive layouts, explicit states, and reusable content patterns. Weekly or milestone-based review keeps decisions visible, but timing depends on scope, feedback availability, integrations, and content readiness.

### Search and performance foundation

Indexable pages can include crawlable initial HTML, stable canonical URLs, route-specific titles and descriptions, structured data, sitemap coverage, intentional robots directives, descriptive internal links, and image dimensions. Performance work considers loading strategy, Core Web Vitals, JavaScript cost, caching, and real user conditions rather than a single lab score.

### Measurement and validation

Validation can cover browser behavior, mobile layouts, keyboard navigation, form delivery, authentication, error paths, analytics events, crawlable output, build artifacts, and deployment smoke tests. Business measures may include successful enquiries or workflow completion, but traffic, rankings, conversion rates, revenue, and commercial outcomes are not guaranteed.

### Editing and ownership

Content can remain editable through a CMS or administration interface with fields and permissions designed before implementation. Editing freedom is balanced with layout, metadata, and accessibility constraints. Documentation should explain routine publishing, media handling, deployment, environment configuration, and the owner of ongoing maintenance.

### Limits

Software cannot create product-market fit, customer evidence, legal approval, content strategy, or internal ownership. Third-party services can change their APIs, pricing, availability, and policies. Security, privacy, accessibility, and regulated content may require specialist review beyond the implementation scope. These dependencies are made visible rather than hidden behind a launch date.

### When it is not a fit

A custom build is not appropriate when the only requirement is a disposable clone, copied interface, unlicensed asset library, or vague request with no accountable reviewer. It may also be unnecessary when an existing platform already meets the content, integration, reliability, and ownership requirements with a smaller configuration change.

### Common questions

An existing site can often be improved instead of rebuilt if its architecture supports the required content, performance, accessibility, analytics, and integrations. Technical SEO foundations can be included, but rankings cannot be promised. Editable content is modeled deliberately so routine updates do not damage page hierarchy or search metadata.

### Content readiness

A website needs approved facts, offer language, service scope, policies, imagery rights, and an owner for future updates. Placeholder copy can support prototyping but should not become permanent evidence. When content is imported from another system, redirects, canonical history, media references, and structured fields are mapped so migration does not create avoidable duplication or loss.

### Reliability after launch

Deployment is followed by checks for status codes, critical routes, forms, authentication, analytics, sitemap availability, console and server errors, and responsive behavior. Monitoring depth depends on the project. A maintenance agreement can cover updates and incidents, but no system can promise uninterrupted availability across every browser, provider, network, and external integration.

### Security and privacy boundary

Implementation can apply secure defaults, secret handling, permission boundaries, dependency review, and data minimization appropriate to the architecture. It does not create legal compliance by itself. Policies, consent, retention, regulated data, contracts, and high-risk security requirements may need specialist assessment and explicit acceptance criteria.

### The next step

A useful request describes the current system, audience, key problem, required integrations, content ownership, timing constraints, budget context, and the action that should become easier after launch. That information supports a focused discovery or proposal; it does not create an automatic commitment before feasibility and scope have been reviewed.

## Automation and AI

URL: https://federicorao.com/services/automation/

AI-powered automation, CRM workflows, quote funnels, integrations, and operational systems for businesses that want to scale without manual drag.

### Who this is for

Automation is strongest for teams with a repeated workflow, a defined trigger, reasonably consistent inputs, known exceptions, an accountable operator, and an outcome that can be measured. Useful candidates often exist in sales operations, marketing, administration, reporting, customer support, data processing, and content operations.

### The problem being solved

Manual handoffs create delay, duplicate entry, missed follow-up, inconsistent records, invisible exceptions, and reporting gaps. The objective is not to remove every person from a process. It is to assign predictable work to systems while keeping review, escalation, and accountability where judgment or impact requires them.

### Workflow discovery

Current steps, owners, tools, inputs, outputs, decisions, permissions, sensitive data, exception paths, and failure consequences are mapped before a tool is selected. Baseline evidence can include cycle time, manual touches, volume, error rate, waiting time, escalation rate, and the cost of unresolved exceptions.

### Deterministic rules before AI

A stable rule is preferred when the decision can be expressed reliably. AI becomes relevant for variable tasks such as classification, extraction, summarization, drafting, semantic matching, or analysis. This distinction reduces cost, makes failures easier to diagnose, and avoids using probabilistic output where exact behavior is required.

### What can be delivered

A delivery may include process maps, API integrations, CRM updates, webhooks, scheduled jobs, data pipelines, prompt and rule definitions, validation, approval queues, logs, alerts, retries, rate limiting, fallbacks, dashboards, operating documentation, and a monitored rollout. The exact combination follows the risk and data involved.

### Human review and controls

High-impact or uncertain actions need approval points, confidence thresholds, access control, audit trails, and manual escalation. The system should define what happens when an API is unavailable, data is incomplete, a provider rate limit is reached, an output fails validation, or an operator needs to correct the record.

### Existing tools and custom software

The current CRM, email platform, database, spreadsheets, forms, and supported APIs are considered first. Custom software is justified where existing tools cannot provide the necessary control, reliability, user experience, or operating cost. Adding a new platform without removing a real constraint creates more administration rather than leverage.

### Measurement

Success criteria are selected before implementation and may include cycle time, completion rate, manual touches, exception rate, data quality, response time, throughput, or operating cost. Estimates are separated from observed results. Time saved, revenue impact, and return on investment are not claimed without a baseline and measured post-launch evidence.

### Privacy, security, and providers

Automation can move personal, commercial, or confidential data between systems. Data minimization, access controls, retention, provider terms, secret management, logging boundaries, and deletion paths need deliberate treatment. Specialist legal or compliance review may be required for regulated decisions or sensitive categories.

### Limits

AI output is probabilistic, provider models and pricing can change, external APIs can fail, and source data can be incomplete. Monitoring and fallback behavior reduce risk but do not create perfect accuracy. A workflow with undefined ownership, constantly changing decisions, or unacceptable failure consequences may not be ready for automation.

### Common questions

Not every automation needs AI; predictable logic should usually stay deterministic. Existing tools can often be retained when their APIs and permissions support the workflow. Production failures are handled through validation, logs, retries, alerts, fallbacks, and escalation designed around the impact of an incorrect or delayed action.

### Data quality before model choice

An automation inherits the quality of its inputs. Duplicate contacts, inconsistent identifiers, missing consent, stale fields, inaccessible documents, and ambiguous labels should be measured before model selection. Validation and normalization may create more value than a more capable model when the workflow cannot distinguish a complete record from an unreliable one.

### Rollout and rollback

A safe rollout starts with a constrained workload, visible logs, known operators, acceptance examples, and a way to disable or bypass the automation. Parallel comparison with the existing process can reveal exceptions before full adoption. Rollback, record correction, provider substitution, and ownership after handover belong in the operating plan.

### Cost and rate boundaries

Usage-based APIs, automation platforms, proxies, storage, observability, and human review can all contribute to operating cost. Provider rate limits and quotas shape throughput. Estimates should state their assumptions, and public users should supply their own credentials where required rather than receiving an undisclosed subsidy or exposing an owner-managed secret.

### The next step

A useful request supplies one repeated process, example inputs and outputs, current tools, approximate volume, sensitive-data constraints, known exceptions, and the cost of delay or manual handling. That is enough to assess feasibility and design a limited pilot without promising a production outcome before the workflow has been observed.

## Technical SEO and Search Experience

URL: https://federicorao.com/services/seo/

Technical SEO, search-native rendering, content architecture, schema, performance, and analytics for businesses that need stronger organic visibility.

### Who this is for

Technical SEO is appropriate for businesses with a real offer and commercial website that is not being crawled, understood, selected, or converted as effectively as it should. Common contexts include service firms, clinics, agencies, founders, and content operations with indexation, rendering, architecture, performance, or measurement constraints.

### Baseline before recommendations

The first pass captures what search engines and users actually receive: response status, robots directives, canonical tags, raw HTML, rendered content, sitemap entries, internal links, structured data, mobile behavior, Core Web Vitals evidence, analytics, Search Console data when available, and the current conversion path.

### Crawlability and indexability

Checks cover robots rules, server and client noindex behavior, sitemap quality, canonical conflicts, redirects, soft 404s, duplicate routes, parameter handling, orphan pages, crawl depth, and whether important content depends on JavaScript execution. Technical changes are tied to a URL, file, response, or observable search signal.

### Search-native rendering

Public pages should provide stable titles, descriptions, canonical URLs, robots directives, structured data, and meaningful content in initial HTML. Client rendering can enhance the experience, but critical search signals should not conflict with server output. Non-200 responses should not rely on JavaScript to become valid pages.

### Money pages and intent

Commercial pages explain audience, problem, process, deliverables, limits, objections, non-fit, evidence, and a relevant next step. The page type follows search intent rather than a keyword alone. A service page, guide, comparison, case study, and tool solve different decisions and should not be generated from one interchangeable template.

### Content architecture

Topic clusters connect a commercial hub to distinct supporting guides through descriptive internal links. Briefs identify primary intent, audience, angle, source needs, related pages, overlap risk, and the action expected after reading. New content is blocked when it merely paraphrases an existing page or lacks a distinct reader decision.

### AI-assisted publishing controls

Automation can prepare outlines, drafts, metadata, source lists, internal-link suggestions, similarity fingerprints, and quality scores. Generated material remains draft-only until reviewed. Factual checks, source-to-claim alignment, originality, useful depth, disclosures, and cannibalization remain editorial responsibilities rather than tasks hidden behind volume.

### Structured data and entities

Schema can describe people, services, organizations, articles, breadcrumbs, profiles, and collection pages when the visible content supports those entities. Markup is not used to invent reviews, ratings, frequently asked questions, credentials, or relationships. Author and editorial-policy pages give entity claims a stable, inspectable destination.

### Performance and experience

Review includes loading strategy, JavaScript cost, image dimensions, layout stability, interaction responsiveness, mobile navigation, semantic elements, labels, and keyboard behavior. Lab tools provide diagnostic evidence, while field Core Web Vitals require real-user data. A score is treated as a signal to investigate rather than a guaranteed ranking lever.

### Measurement and iteration

Search Console, analytics, crawl checks, PageSpeed evidence, index coverage, query-to-page fit, leads, and lead quality can be monitored after deployment. The purpose is to test whether pages are discovered, selected, read, and connected to useful actions. More impressions without qualified outcomes do not automatically represent progress.

### Limits and non-fit

Technical corrections cannot guarantee rankings, traffic, citations, or revenue because competition, authority, usefulness, demand, and search systems also matter. The work is not a fit when the only objective is publishing thin pages at scale, fabricating authority, hiding commercial relationships, or pursuing tactics designed to deceive users or platforms.

### Common questions

New content is preceded by review of existing pages, intent, overlap, competitor page types, sources, internal links, and measurement. Articles can be drafted automatically with quality gates, but publication still requires review. Technical foundations improve eligibility and clarity without creating a promise of ranking.

### Competitor and SERP evidence

Competitor research is used to understand search intent, page type, coverage gaps, source expectations, and user objections. It is not permission to copy structure, language, proprietary data, reviews, or claims. Search results are time-sensitive and location-dependent, so observations should record the market, device or locale where relevant, and collection date.

### Internal links and consolidation

Each new guide should have a relevant path from a hub and should return readers toward a useful service, related guide, project, comparison, or contact step. Existing content is preferred when it can be improved. Overlapping drafts are merged, redirected, or withheld so publication does not create several weak pages competing for the same decision.

### External authority

Technical and editorial improvements operate on the site itself. Mentions, references, citations, reviews, and links from other communities require genuine participation and independent choice. Reddit, forums, directories, guest contributions, and digital public relations should not rely on fake accounts, coordinated manipulation, undisclosed promotion, or automated spam.

### The next step

A useful SEO request includes the domain, target market, core services, priority locations or languages, access to Search Console and analytics where available, known migrations, publishing capacity, and lead-quality definition. That evidence supports prioritization. A generic audit without business context can identify defects but cannot decide which opportunity matters most.

## Web, Automation and AI Projects

URL: https://federicorao.com/projects/

Review selected web development, automation, AI, e-commerce, learning-platform, and digital-product implementations by Federico Rao.

### How to read these records

Each project page describes an implementation: the problem it addressed, the architecture chosen, the workflows it supports, and the constraints that shaped it. Visible product behaviour, interface decisions, and technical structure can be inspected directly. Business outcomes that depend on client data are not published as proof, and none of these records should be read as a claim about revenue, growth, or performance for the organisations involved. Ownership is stated per project, including collaborative work and internal tools built without a client.

### ELMS Learn, enterprise learning platform

A learning management system built as fourteen independently deployable Spring Boot microservices behind a React frontend covering the educational lifecycle: enrolment, course delivery, assessment, mentoring, placement, and analytics. Services communicate over REST and Kafka event streaming, each owning its own PostgreSQL database to enforce bounded contexts. The frontend uses role-based routing across four user types with code splitting and a resilient try-API-then-mock development pattern. It is the largest distributed-systems record on the site and the clearest example of multi-service architecture rather than a single application.

### Acquamarina, commerce storefront

A bilingual luxury storefront for handcrafted goods, built around product browsing, cart, checkout, and authentication with a Mediterranean design system. State is handled with feature-scoped stores, forms are validated with schema-based typing, and the cart drawer uses spring animation. Full English and Italian localisation is part of the build rather than a later addition. It demonstrates commerce interaction design and internationalisation on a small, self-contained codebase.

### Atelier Engine, AI imagery workflow

A desktop-class tool for generating product photography variants for fashion inventory. A vision model analyses source photographs, prompts are constructed from that analysis, and generation runs in batches producing studio, close-up, and on-model variants. The backend streams progress over WebSocket; the interface provides prompt authoring, run management, and workflow control. Operational concerns dominate the design: CSV inventory batches, pausing and resuming runs, recovery after interruption, and append-only audit logging.

### MTRM, fashion trend analytics

A machine-learning pipeline that ingests social media text and images to score emerging demand signals in fashion. A custom text classifier trained on a demand lexicon handles language, an object-detection model identifies clothing items in screenshots, and scored output is served through an API over a document store. It is a research-oriented system: the interesting parts are the labelling approach, the scoring method, and the deployment shape rather than a polished consumer interface.

### Praxis, practice management

A focused management interface for professional clinics covering the full case lifecycle: creation, active handling, notes and review, archival, and reactivation. The interface is deliberately narrow, with two panels separating active from historical cases, file attachments, and inline notes. It was built as a minimum viable tool for immediate deployment, replacing spreadsheets and paper files for one practice, and it illustrates how small a system can be while still being useful.

### Third Eye, mobile audio product

A cross-platform mobile application for mindfulness and ambient sound mixing, layering ambient layers over master tracks through a custom interface. The client uses persisted state and GPU-accelerated graphics for mood-reactive visuals; the backend handles authentication, catalogue, mixer state, subscriptions, and content administration. It is the most complete mobile record here, covering onboarding, listening journeys, subscription handling, and the administrative surface needed to keep content current.

## ELMS Learn Case Study

URL: https://federicorao.com/projects/elms-learn/

Case study for ELMS Learn, an enterprise-grade learning platform with AI-assisted educational workflows and scalable architecture.

### System shape

Fourteen independently deployable Spring Boot services sit behind an API gateway, each owning its own PostgreSQL database so that bounded contexts are enforced by deployment rather than by convention. Services communicate through REST clients for synchronous calls and Kafka event streaming for asynchronous workflows such as enrolment, progress tracking, and notification fan-out. The React frontend spans four distinct user roles with route-level code splitting and a development pattern that falls back to mock data when a service is unavailable, so the interface remains workable while individual services are being rebuilt.

### What can be inspected

The record is useful for architecture decisions rather than for outcome claims: service decomposition and where the boundaries were drawn, authentication using token plus double-submit protection, gateway routing, real-time presence over WebSocket, and observability through metrics endpoints. It was collaborative work, so credit is shared. Nothing here asserts adoption, enrolment numbers, or commercial performance for any institution.

## Contact Federico Rao - Get a Project Proposal

URL: https://federicorao.com/contact/

Contact Federico Rao to scope conversion-focused websites, AI automation, CRM workflows, technical SEO, or digital consulting projects.

### What to include

Your business type, current challenge, target timeline, budget range, and the outcome you want to create.

### Next step

The ideal request leads to a focused proposal, technical direction, or a scoped discovery call.

## Free Initial HTML SEO Checker

URL: https://federicorao.com/tools/raw-html-checker/

Check title, description, canonical, robots, H1, structured data, text, and links delivered in the initial HTML response of a public web page.

### What the checker reads

The report looks for response status and content type, title, meta description, canonical URL, robots directive, H1 count, structured-data blocks, visible text volume, and links present in the raw response. It does not wait for hydration, click controls, accept consent prompts, log in, or render a browser session.

### Why initial HTML matters

Search engines can render JavaScript, but critical signals should be stable before and after rendering. Conflicting canonical or robots directives create ambiguity, non-200 responses may not be rendered, and delayed content can make diagnostics harder. Inspecting the first response provides a reproducible baseline rather than a complete SEO verdict.

### What the result cannot prove

A passed check does not prove that a page is useful, authoritative, accessible, fast for real users, eligible for a rich result, indexed, ranked, cited by AI systems, or able to convert visitors. Word and character counts are coverage prompts rather than ranking factors or writing targets.

### Privacy and operating limits

The submitted URL is sent to the server for the requested check and the fetched response is not retained by this tool. Only public HTTP or HTTPS pages should be submitted. Network rules, redirects, robots behavior, timeouts, unsupported content types, and remote-server protection can prevent a result.

### How to use the evidence

Review failures in context, compare the raw response with the rendered page, confirm server status, and inspect the exact route in Search Console when access is available. A missing element may require server rendering, static generation, metadata configuration, routing fixes, or a deliberate decision that the page should not be indexed.

### Next steps

Use the technical SEO service for a broader crawl, indexation, rendering, architecture, performance, and measurement review. The editorial policy explains how public guidance and AI-assisted material are sourced and reviewed. The topic hub connects the checker to implementation guides as they are published.

## Veterinary Websites and Digital Systems

URL: https://federicorao.com/read/veterinari/

Digital presence systems for veterinary clinics: trust-building websites, local SEO, appointment funnels, and patient acquisition workflows.

### How owners choose a clinic

Most veterinary enquiries start with an urgent or anxious question: an unfamiliar symptom, a second opinion, a vaccination schedule, or a new pet. Owners compare a small number of nearby clinics within minutes, and they decide on legibility rather than depth. What settles it is usually whether the clinic states its species and services, shows the people who will handle the animal, publishes opening and emergency arrangements without ambiguity, and makes the next action obvious. A clinic that answers those four questions above the fold rarely loses the comparison to a longer website.

### The clinic website

The site should carry one page per service line that owners actually search for, written in the words they use rather than clinical terminology alone: routine consultation, vaccination, sterilisation, dentistry, diagnostics, surgery, and species-specific care where the clinic accepts exotics. Each page states what happens during the visit, what preparation is required, how long it takes, and how pricing is communicated. Team pages carry names, qualifications, and registration details where professional rules require them. Practical information such as parking, waiting arrangements, and payment methods reduces the number of calls that exist only to remove uncertainty.

### Local discovery and opening information

Veterinary demand is overwhelmingly local and time-sensitive, so consistency between the site, the map listing, and directory entries matters more than volume of content. Address, phone number, and hours must match everywhere, including seasonal changes and holiday cover. Emergency and out-of-hours arrangements deserve their own explicit statement, because an owner who cannot find it at eleven at night will call whoever states it clearly. Structured data describing the practice, its location, and its opening hours helps search systems present that information accurately in results.

### Appointments, reminders, and follow-up

The booking path should ask only what is needed to schedule: animal, reason, urgency, and contact. Everything else can be collected later. After the visit, the useful automation is a reminder system tied to real clinical events, such as vaccination recalls, parasite treatment intervals, post-operative checks, and chronic condition reviews. These are welcome when they are specific and resented when they read as marketing. A single record per client and animal, kept in one system, is what makes the difference between a reminder that feels attentive and one that arrives after the pet has already been treated elsewhere.

### Consent, records, and communication

Client and animal records contain personal data, and in many cases sensitive commercial information about the practice itself. Any automation touching them needs a defined lawful basis, retention period, and deletion path, with marketing consent kept separate from clinical communication. Review requests should be sent only where professional and platform rules permit, and never conditioned on the outcome of treatment. Photographs of animals and clinical imagery require the owner’s permission before publication, and identifying details should be removed unless consent explicitly covers them.

### What to measure in a clinic

Useful measures are appointment requests, calls originating from the site, completed booking forms, recall response rates, and the proportion of new clients who return for a second visit. Traffic alone is not a clinic metric. None of this can compensate for capacity: a system that generates more requests than the practice can schedule creates a worse experience than the one it replaced. Clinical claims, success rates, and comparative statements about other practices are outside what a website should assert, and they are excluded from the pages built here.

## Law Firm Websites and Digital Systems

URL: https://federicorao.com/read/studi-legali/

Authority-led websites, clear practice-area journeys, qualified enquiry paths, and maintainable digital systems for law firms.

### How prospective clients evaluate a firm

Legal enquiries are usually triggered by a deadline, a dispute, a transaction, or a regulatory notice, and the person searching is rarely a lawyer. They arrive with a situation rather than a practice area, and they are trying to establish two things quickly: whether the firm handles this kind of matter, and whether it can be trusted with it. Firms lose that comparison by describing themselves in institutional language, listing practice areas without explaining what each covers, or hiding the individuals responsible behind a collective identity. Named professionals with stated qualifications convert better than an anonymous firm profile.

### Practice-area architecture

Each practice area should have its own page written for the situation, not the statute: what kinds of matters it covers, how an engagement typically proceeds, what the client is expected to provide, how fees are structured in general terms, and what the realistic time frames are. A single page listing every area of law forces the reader to self-diagnose and gives search systems nothing specific to select. Where the firm serves several cities or jurisdictions, that distinction belongs in the architecture as well, because competence and admission are location-bound in ways other sectors are not.

### Authority and publication

Published commentary is the most credible signal a firm can control, provided it is genuinely authored and clearly attributed. Useful material explains a change in law, a procedural step, or a common misunderstanding, with dates and sources, and states plainly that it is general information rather than advice on a specific matter. Professional conduct rules restrict comparative claims, guarantees of outcome, and certain forms of solicitation, and they differ by jurisdiction and bar association. The publishing system is built to respect those constraints rather than to maximise output.

### Enquiry qualification and conflicts

The intake form should capture the matter type, the parties involved, the jurisdiction, the deadline if one exists, and how the person prefers to be contacted. Collecting the counterparty early matters because a conflict check has to happen before substantive discussion, and an intake process that ignores this creates real professional risk. Routing rules can direct an enquiry to the right practice group automatically, but the decision to accept a matter stays with a responsible lawyer. Automatic acknowledgements should confirm receipt without implying that a relationship has been established.

### Confidentiality and document handling

Anything built for a firm has to assume that the data passing through it is privileged. That constrains where form submissions are stored, which third-party tools may see message content, how long records are retained, who inside the firm can access them, and what happens to an enquiry that is never converted into a matter. Document upload, if offered, needs access control, encryption in transit and at rest, and a defined retention and deletion policy. Convenience features that route client information through unvetted external services are excluded rather than mitigated.

### What to measure in a firm

The meaningful measures are qualified enquiries by practice area, conflict-clear consultations, the proportion of enquiries that become engagements, and the source of the matters the firm actually wants. Volume of contact-form submissions is a poor proxy, because unqualified enquiries consume partner time without producing work. A website cannot promise case outcomes, cannot imply a success rate, and cannot substitute for reputation built through referral and precedent. Where a claim cannot be evidenced or is restricted by professional rules, it is left off the page.

## Dental Practice Websites and Digital Systems

URL: https://federicorao.com/read/dentisti/

Trust-building websites, treatment information, booking journeys, local discovery, and responsible patient-acquisition systems for dental practices.

### How patients decide before calling

Dental enquiries split into two very different journeys. Routine care is chosen on proximity, availability, and price transparency, often in under a minute. Elective and higher-value treatment, such as implants, orthodontic aligners, or cosmetic work, involves weeks of comparison, and the deciding factors are the practitioner’s credibility, the clarity of the treatment explanation, and how the practice handles the first contact. A single website has to serve both without letting the routine path bury the elective one, which is why treatment pages and appointment pages are usually separated.

### Treatment pages that carry the decision

Each treatment should have a dedicated page describing what the procedure involves, how many appointments it typically requires, what recovery looks like, which alternatives exist, and how the practice communicates cost. Ranges and consultation-based pricing are acceptable; silence is not, because it forces the patient to call in order to obtain the one fact that determines whether they proceed. Practitioner profiles should state qualifications, registration numbers where required, and specific experience with that treatment. This is what distinguishes a practice page from a template that could belong to any clinic.

### Imagery, claims, and regulated content

Clinical photography is persuasive and heavily constrained. Before-and-after images require documented patient consent, must represent typical rather than exceptional results, and in many jurisdictions are restricted or prohibited in advertising for certain treatments. Testimonials referring to specific clinical outcomes carry similar restrictions. Health-claim rules generally prohibit guarantees, comparative superiority claims, and language that creates unrealistic expectations. Content is written to stay inside those limits by describing process and criteria rather than promising results, which also produces pages that survive review by a professional body.

### Booking, recalls, and the follow-up gap

The most common operational loss is an enquiry that arrives outside opening hours and is answered the next morning, by which time the patient has contacted another practice. An automated acknowledgement that confirms receipt, sets a response expectation, and offers a booking slot closes most of that gap without pretending to be a clinician. Beyond intake, the durable value is in recall automation tied to real intervals: hygiene appointments, orthodontic reviews, implant checks, and treatment plans that were quoted but never started.

### Reputation and patient data

Review platforms hold the public narrative, and the only sustainable response is a direct relationship: an owned patient list, a request workflow that complies with platform rules, and an internal route for complaints that resolves them before they become public. Patient records are health data under most privacy regimes, which means marketing consent must be separate from clinical communication, retention must be defined, and any tool that processes appointment or treatment information needs a documented basis. Analytics and advertising pixels on treatment pages deserve particular scrutiny for the same reason.

### What to measure in a practice

Track booked appointments by treatment type, consultation-to-treatment conversion, recall response, and the value of the enquiries the practice actually wants rather than total form submissions. Distinguishing elective from routine demand is what makes the numbers actionable. No digital system can increase chair capacity, shorten a treatment plan, or substitute for clinical reputation, and none of the pages built here assert success rates, comparative outcomes, or guarantees about treatment results.

## E-Commerce Development and Conversion Systems

URL: https://federicorao.com/read/ecommerce/

E-commerce architecture, product discovery, conversion measurement, operational automation, and maintainable integrations for online stores.

### Where online stores actually lose revenue

Most underperforming stores do not have a traffic problem. They have a catalogue that cannot be navigated, product pages that omit the information required to decide, a checkout that adds friction at the moment of commitment, and measurement that cannot attribute any of it. Diagnosis starts by separating the funnel into discovery, evaluation, and purchase, then locating which stage loses the most qualified sessions. Spending on acquisition before that is identified reliably increases cost per order without increasing orders.

### Catalogue structure and discovery

Category architecture should reflect how customers describe what they want, not how the business organises its inventory or how a supplier structures a feed. Faceted navigation needs deliberate rules about which combinations are crawlable, which are canonicalised, and which are blocked, because uncontrolled filters generate thousands of near-identical URLs that consume crawl budget and compete with the pages that matter. Search on site deserves the same attention as navigation: for large catalogues internal search is the highest-intent surface in the entire store.

### Product pages and the evidence to decide

A product page has to answer sizing, materials, compatibility, delivery time, total landed cost, and return conditions without requiring a support message. Images need consistent framing and scale references; variants must be selectable without breaking the URL; stock status has to be accurate rather than optimistic. Structured product data allows search systems to present price and availability correctly, and it should describe what is actually on the page rather than inventing ratings or reviews. Reviews, where used, must be genuine and their collection method disclosed.

### Checkout, payment, and post-purchase

Checkout should ask for the minimum required to fulfil and charge, present total cost including shipping and tax as early as possible, support the payment methods the target market actually uses, and never require account creation before purchase. Post-purchase is part of the same system: order confirmation, dispatch notification, tracking, returns initiation, and the reorder path all affect repeat rate more than any acquisition campaign. Abandoned-cart automation belongs here too, subject to consent and a frequency that does not train customers to ignore the sender.

### Operations, integrations, and data ownership

Commerce systems fail operationally more often than they fail visually. Inventory synchronisation, pricing rules, product feeds to shopping and marketplace channels, order export to fulfilment, and returns handling all need defined owners and failure behaviour. Where several systems hold product data, one has to be authoritative and the direction of synchronisation must be explicit. Customer and order data should remain exportable and owned by the business rather than locked inside a platform that controls both the storefront and the relationship.

### What to measure in a store

The metrics that matter are conversion rate by traffic source, average order value, contribution margin after acquisition and fulfilment cost, return rate by product, and repeat purchase rate. Sessions and impressions describe activity, not health. Consent-aware analytics will under-report, and modelling that gap honestly is better than configuring tracking that ignores consent. No technical improvement compensates for a product that is not competitive on price, availability, or delivery in its market, and that constraint is named rather than engineered around.

## Fitness and Wellness Digital Systems

URL: https://federicorao.com/read/fitness/

Membership, booking, retention, lead follow-up, reporting, and measurable conversion systems for fitness and wellness businesses.

### The decision window for a new member

Fitness demand is seasonal, emotional, and short-lived: someone decides to join within a narrow window and chooses whichever option removes friction fastest. The questions are consistent, namely what it costs, when the sessions run, whether there is a trial, how the first visit works, and whether the environment suits their level. Sites that hide pricing behind a contact form lose the majority of that window to a competitor who publishes it. Timetables and trial terms are the two pieces of information most often buried and most often decisive.

### Timetable, trial, and membership pages

The class schedule should be a real page, readable on a phone, current, and linked directly rather than distributed as an image or a downloadable file. Membership options need plain comparison: what is included, contract length, freeze and cancellation terms, joining fees, and what happens at renewal. The trial offer deserves its own page describing exactly what a first visit involves, what to bring, and what happens afterwards, because uncertainty about the first session stops more sign-ups than price does. Coach profiles carry qualifications and specialisms where they influence the choice.

### Booking, capacity, and the first session

Class booking and capacity management shape the member experience more than the website design. The system must show real availability, handle waitlists, allow cancellation within a stated window, and enforce no-show rules consistently. The first-session sequence is the highest-leverage automation in the sector: confirmation, practical preparation, a reminder timed to the session, and a follow-up afterwards. A trial that ends without a defined next step converts far below one where the offer, the deadline, and the person to speak to are all stated.

### Retention, churn, and lifecycle messaging

Acquisition cost is only recovered over months of membership, so the economically important system is retention, not lead generation. Useful signals are attendance frequency, gaps in visits, class type changes, and payment failures. Each supports a specific intervention: a re-engagement message after a lapse, a check-in after a failed payment, a progression suggestion after a plateau. Messaging has to be limited in frequency and genuinely relevant, because members treat unsolicited volume from a gym as a reason to cancel rather than a reason to return.

### Health claims, images, and consent

Advertising rules in most markets restrict claims about weight loss, body transformation, health outcomes, and timeframes, and transformation photographs are subject to consent and to representativeness requirements. Any medical or nutritional claim needs qualification appropriate to who is making it. Member photographs, class recordings, and social content require permission that covers the intended use. Health information collected at sign-up, including injuries and conditions, is sensitive data and must be handled with a defined basis, restricted access, and a retention period rather than sitting in a shared spreadsheet.

### What to measure in a membership business

Track trials booked, trial-to-member conversion, attendance frequency, monthly churn, and lifetime value against acquisition cost. Follower counts and page views do not describe the health of a membership business. Digital systems cannot compensate for equipment, class quality, coaching, or a facility at capacity, and where those are the constraint the honest recommendation is to address them first rather than to increase inbound volume.

## Real Estate Websites and Lead Systems

URL: https://federicorao.com/read/real-estate/

Local authority, clear property presentation, enquiry qualification, measurement, and maintainable lead follow-up systems for real-estate teams.

### Two audiences, two entirely different journeys

A real-estate site serves buyers and sellers, and confusing them is the most common structural error in the sector. Buyers want listings, filters, photographs, location detail, and a fast way to arrange a viewing. Sellers want evidence of local performance, an explanation of the process, fee transparency, and a low-commitment way to obtain a valuation. Those journeys need separate entry points, separate pages, and separate follow-up, because the seller enquiry is usually the commercially valuable one and it is routinely buried beneath property search.

### Listings, portals, and duplicate content

Portals dominate property search, and an agency site that reproduces portal descriptions verbatim competes with the portal using the same words and loses. Listing pages earn their place through material the portal does not carry: additional photography, floor plans, neighbourhood context, transport and school detail, viewing arrangements, and named agent contact. Listing lifecycle also needs a policy. Sold and withdrawn properties should not be silently deleted, because those URLs accumulate links and traffic; they should be retained with clear status, redirected to a relevant collection, or archived deliberately.

### Local authority and area pages

Location pages are where a local agency can outrank a national portal, provided each one is genuinely written about that area: price context with dated sources, property types, transport, amenities, and the kinds of buyers active there. The failure mode is a template repeated across dozens of neighbourhoods with only the place name changed, which produces near-duplicate pages that search systems recognise and ignore. Fewer, substantive area pages tied to actual instructions perform better than exhaustive coverage of every district the agency might theoretically serve.

### Valuation requests, viewings, and follow-up

The valuation request is the highest-value form on the site and should ask for as little as possible: address, contact, and preferred timing. Enrichment can happen afterwards. Viewing requests need real availability rather than a generic enquiry box, and both need acknowledgement within minutes, because property enquiries are usually sent to several agents at once and the first substantive reply sets the relationship. Longer-term nurture matters too, since a seller thinking about moving in six months is worth more than a form filled in today.

### Data, disclosure, and accuracy

Property particulars carry legal weight in most jurisdictions, and material information requirements increasingly extend to tenure, service charges, energy performance, and known defects. Descriptions must be accurate rather than aspirational, and any automation that generates listing copy needs review before publication for exactly this reason. Contact records include personal and financial circumstances, so retention, access control, and marketing consent have to be defined. Performance claims about average sale price or time on market require dated evidence from the agency’s own records.

### What to measure in an agency

The measures worth tracking are valuation requests, instructions won, viewing requests per listing, source of instruction, and cost per instruction. Listing page views are a weak proxy because portal traffic distorts them. No digital system creates supply in a constrained market, changes pricing reality, or shortens a transaction chain, and claims implying otherwise are excluded from the pages built here.

## Restaurant Websites and Direct Guest Systems

URL: https://federicorao.com/read/restaurants/

Fast local discovery, menus, reservations, ordering paths, analytics, and consent-aware guest relationship systems for restaurants.

### What a guest checks before booking

Restaurant decisions happen fast and on a phone, usually with a group waiting for an answer. The guest checks four things: the menu, whether a table is available at the time they want, where it is and how to get there, and whether it looks right for the occasion. Anything that delays those four answers loses the booking. Menus published as image files or downloadable documents, opening hours that contradict the map listing, and a reservation link that fails on mobile are the three most common and most expensive failures.

### Menu, hours, and practical information

The menu should be real text on a real page, current, and structured so that allergen and dietary information is findable, which in many jurisdictions is a legal requirement rather than a convenience. Prices belong on the page. Opening hours must be correct everywhere they appear, including holiday variations, and service periods should be distinguished when lunch, dinner, and bar service differ. Practical detail such as accessibility, parking, group capacity, and whether the venue takes walk-ins removes the phone calls that exist purely to resolve uncertainty.

### Reservations and direct ordering

Reservation and delivery marketplaces provide reach and take both a commission and the guest relationship. The objective is not to abandon them but to make the direct path at least as easy, so guests who already know the restaurant do not pay an intermediary to reach it. That means a booking flow that works in a few taps, an ordering system with accurate menu synchronisation and honest timing, and confirmation messages that come from the restaurant. Direct channels are also where a guest email or phone number can legitimately be collected.

### Local discovery and the map listing

For a single venue the map profile often outperforms the website in raw visibility, so the two must agree exactly on name, address, phone, hours, and category. Photographs, menu links, and attributes on the profile influence selection directly. Reviews are part of discovery rather than a separate reputation exercise: responding consistently, resolving issues privately, and never incentivising or fabricating feedback is both the platform rule and the only approach that survives scrutiny. Structured data describing the venue, its menu, and its hours supports accurate presentation in results.

### Guest data and repeat visits

The durable asset is a direct relationship with people who have already visited. Booking and ordering systems collect contact details, and with proper consent they support a small number of genuinely useful messages: a seasonal menu change, a holiday opening, an event, a reservation reminder. Frequency is the whole discipline here. Marketing consent must be separate from transactional confirmation, preferences must be honoured, and data collected through a marketplace generally cannot be repurposed for direct marketing without a valid basis of your own.

### What to measure in a venue

Track direct reservations against marketplace reservations, cover counts by service, no-show rate, direct order value net of commission, and repeat visit rate. Website sessions describe interest, not covers. No digital system fixes food, service, location, or capacity, and where those are the constraint the useful advice is operational rather than technical. Claims about awards, rankings, or recognition are published only where an independent source can be cited.

## Web3 and Crypto Digital Systems

URL: https://federicorao.com/read/crypto/

Trust-oriented web infrastructure, product communication, community journeys, measurement, and maintainable interfaces for Web3 projects.

### Credibility is established off-chain

Web3 products face a trust problem that ordinary software does not: the audience assumes a scam until shown otherwise, and the usual persuasion techniques increase suspicion rather than reducing it. What establishes credibility is unglamorous and verifiable, namely a named team where the project is not deliberately pseudonymous, published contract addresses, completed and readable audit reports, documented tokenomics with vesting, a working product rather than a roadmap, and a public record of what has shipped. Marketing language calibrated for consumer software actively damages this audience’s confidence.

### Documentation as the primary surface

For technical products documentation is the marketing site. Integration guides, API references, contract interfaces, error semantics, testnet instructions, and worked examples determine whether a developer proceeds or leaves. It should be versioned, searchable, indexable as text rather than trapped in a client-rendered application, and honest about limitations and known issues. A landing page that describes a protocol in abstractions while the documentation is incomplete converts nobody who matters, because the audience evaluating the product reads the reference before the pitch.

### Wallet flows and product interface

The interface has to make irreversible actions comprehensible: what is being signed, which network is active, what the transaction costs, what happens if it fails, and what state the user is in afterwards. Connection flows should degrade clearly when a wallet is missing, on the wrong chain, or rejecting a request. Pending, confirmed, and failed states need distinct handling rather than a spinner. Because on-chain mistakes cannot be reversed, confirmation design and plain-language transaction summaries do more for retention than any visual refinement.

### Community, disclosure, and manipulation

Distribution in this sector runs through communities, and the boundary between participation and manipulation is where most projects create legal and reputational exposure. Paid promotion must be disclosed, engagement must not be fabricated, and coordinated posting through purchased accounts is both platform-prohibited and increasingly treated as market manipulation. Sustainable presence comes from genuine technical contribution, published research, transparent post-mortems after incidents, and moderators who correct misinformation about the project rather than amplifying hype.

### Regulatory constraints on what you may say

Financial promotion rules in most jurisdictions restrict how digital assets may be advertised, frequently requiring risk warnings, prohibiting statements about returns, and in some markets limiting promotion to authorised firms or restricted audiences. Rules differ by country and change quickly, and the applicable set depends on where users are, not where the team is. Content is built to accommodate that: no yield or return claims, explicit risk language where required, geographic restrictions where they apply, and legal review before publication rather than after a complaint.

### What to measure in a protocol

Meaningful measures are wallets connected, transactions completed successfully against those attempted, documentation-to-integration conversion, developer retention, and failure rate by flow. Token price, follower counts, and community size describe sentiment rather than product health. Nothing in this work constitutes financial or investment advice, no engagement includes promotion of an asset as an investment, and projects whose primary requirement is manufactured hype are declined rather than scoped.

## Acquamarina Case Study

URL: https://federicorao.com/projects/acquamarina/

Explore a coastal beach-club platform combining commerce, administration workflows, visual content management, and a customer-facing digital experience.

### System shape

A storefront built around feature-scoped state stores for authentication, cart, and catalogue, with schema-validated forms and animation used for the cart drawer transition. Localisation covers English and Italian across the entire interface, including validation messages and formatted values, which is treated as an architectural decision rather than a translation pass added at the end. The visual system pairs serif and sans typography over a coastal palette.

### What can be inspected

This is an internal build rather than a client engagement, so it demonstrates interaction and internationalisation decisions rather than commercial results. Useful things to look at are variant selection, cart state persistence, form validation behaviour, and how the bilingual routing and content model avoid duplicating components per language.

## Lega Calcio a 8 Case Study

URL: https://federicorao.com/projects/lega-calcio-a8/

Explore a mobile platform for amateur 8-a-side football leagues with a live match center, standings, fantasy game, tournaments, and QR match check-in.

### System shape

A React Native application built with Expo on the Hermes engine, covering an amateur football league end to end: a live match center with provisional scores, event timelines, line-ups, and media; standings, calendars, and disciplinary records; published tournaments with brackets; and a league-scoped fantasy game with credit budgets, transfer orders, and submission deadlines. Roles and invite codes separate player, team, and organizer paths, and matches support QR check-in at the venue.

### What can be inspected

The record shows how a community sport product is structured: which surfaces stay public versus role-gated, how demo seasons are labelled to keep sample data honest, and how the fantasy layer reuses league data instead of duplicating it. Screenshots and showcase footage on this page come from the real application running on a device. No claim is made about league adoption or subscription numbers.

## Ostia in Love Case Study

URL: https://federicorao.com/projects/ostia-in-love/

Explore a seaside experience marketplace for the Roman coastline covering curated categories, zone maps, weekly events, and verified experience bookings.

### System shape

A React Native and Expo marketplace for seaside experiences on the Roman coastline. The catalog is organized by curated categories and coastal zones, with a weekly event calendar, experience sheets carrying real photography, per-person pricing, duration, group size, meeting points, and weather policies, and a map with zone filters and a QR scanner. Guest exploration is separated from authenticated booking, and reviews unlock only after a consumed experience.

### What can be inspected

The useful decisions are marketplace mechanics: how availability is driven by date and zone at the same time, how operator policies such as weather rescheduling are surfaced before payment rather than buried, and how verified-only reviews are enforced by the booking lifecycle. Visuals on this page are captured from the application itself. No claim is made about booking volume or operator revenue.

## Woowpark Case Study

URL: https://federicorao.com/projects/woowpark/

Explore a parking booking application for Rome combining live spot availability, map exploration, transparent tariffs, protected payments, and QR access.

### System shape

A React Native and Expo booking application for parking in Rome, built around destination-first search with arrival-time filters, live free-spot availability with explicit freshness states, and a Google Maps surface with availability-colored pins and full-screen exploration. Tariffs show hourly and daily rates with booking discounts computed per listing, checkout runs on Stripe with 3DS and SCA, operator cancellation conditions are saved before payment, and access at the barrier is by QR code.

### What can be inspected

The record demonstrates how a transactional mobility product communicates trust: freshness labels on availability data rather than implied certainty, prices decomposed into original, discount, and savings, and payment-protection copy carried into the interface. Visuals come from the real application on a device. No claim is made about parking volumes or operator conversion.

## Studio Legale Mazzoli Case Study

URL: https://federicorao.com/projects/studio-mazzoli/

Explore the multilingual website of a Rome law firm: eight languages, security hardening, performance work, HTTPS, and WordPress 7.1 on PHP 8.5.

Client: Studio Legale Mazzoli, Law firm based in Rome, Italy. Website: https://www.studiomazzoli.it/

### System shape

Studio Legale Mazzoli is a Rome law firm working in public, administrative, and constitutional law, gambling and betting regulation, ethics and disciplinary law, and legal advisory. The project covers the technical side of its website: a multilingual WordPress platform in eight languages (Italian, English, Estonian, French, German, Spanish, Russian, and Serbo-Croatian) with native URLs, translated menus and metadata, and hreflang; WordPress 7.1 on PHP 8.5 with Polylang Pro and Yoast SEO; HTTPS across the whole site; a mobile-first home hero written for every language; and ongoing technical management through a custom plugin that accepts only encrypted, signed requests with replay protection and automatic file backups.

### What can be inspected

The record is about running a professional website carefully: a must-use security plugin that closes user enumeration, XML-RPC, and dangerous unauthenticated actions, rate-limits login, and sends security headers; a verified backup before every change; platform releases rehearsed on a local staging copy with a tested rollback script; and legal terminology handled with care in each language. The figures on this page are technical measurements taken on the site itself, such as autoloaded options kept to about 43 KB and the home hero image optimised to 238 KB. No claim is made about traffic, enquiries, or revenue for the firm.

## Forecast Markets Case Study

URL: https://federicorao.com/projects/forecast-markets/

Explore a white-label prediction market platform with fixed-odds event markets, wallet, KYC onboarding, operator dashboards, and double-entry accounting.

### System shape

A white-label prediction market platform for regulated European operators, structured as a pnpm and turbo monorepo: a Next.js 15 frontend serving both the player experience and the backoffice, a Fastify 5 API, BullMQ workers for jobs, PostgreSQL through Drizzle, and hexagonal adapters isolating payment, KYC, and regulatory integrations per jurisdiction. Markets are fixed-odds against the house on non-sport events, each with declared resolution criteria, sources, close dates, and void rules; every ticket books balanced double-entry ledger records.

### What can be inspected

The interesting decisions are regulatory and financial: how an order ticket locks odds and previews payout before commitment, how wallet and portfolio surfaces present player money with bank-statement seriousness, how KYC onboarding is a first-class flow rather than an afterthought, and how the operator control plane exposes turnover, capital at risk, and activation funnels. The platform is in active development; screens shown are from the working product in demo mode, and no claim is made about live operators or wagering volume.

## Atelier Engine Case Study

URL: https://federicorao.com/projects/atelier-engine/

Explore AI-assisted fashion photography tooling designed around repeatable product-imagery generation, review, and creative production workflows.

### System shape

A batch pipeline for product imagery: source photographs are analysed by a vision model, prompts are constructed from that analysis, and multiple generation backends produce studio, close-up, and on-model variants per item. A Python API streams run progress to the interface over WebSocket, while the client provides prompt authoring, run management, and workflow configuration. Inventory arrives as CSV, runs can be paused and resumed, interrupted work recovers automatically, and every step is written to an append-only audit log.

### What can be inspected

The interesting decisions are operational rather than generative: how partial failures are handled across a long batch, how runs remain resumable, how model backends are abstracted so one can be swapped, and how output is reviewed before it enters a catalogue. A separate download page is kept for the desktop build itself and is deliberately excluded from search indexing.

## MTRM Case Study

URL: https://federicorao.com/projects/mtrm/

Explore a product interface for AI and machine-learning fashion-trend analysis, research workflows, structured signals, and decision support.

### System shape

A machine-learning pipeline for demand signals in fashion. Social text is classified by a custom model trained on a domain-specific demand lexicon, screenshots are processed by a fine-tuned object detector to identify garments, and the combined signals are scored and stored in a document database. Predictions are served through a Python API and the whole stack is containerised for deployment.

### What can be inspected

The record is about method: how the labelling scheme was defined, how detection and language signals are combined into a score, and how the pipeline is structured for repeated ingestion. Trend predictions are probabilistic and unvalidated against commercial outcomes here, so no claim is made about accuracy, lead time, or resulting revenue for any brand.

## Praxis Case Study

URL: https://federicorao.com/projects/praxis/

Explore a practice-management interface designed around structured professional matters, repeatable operational workflows, and clear information access.

### System shape

A deliberately narrow management interface covering one lifecycle: case creation, active handling with notes and attachments, review, archival, and reactivation. Two panels separate active from historical work, status is tracked explicitly, and file upload is scoped to what the practice actually exchanges. It was built as a minimum viable tool intended for immediate deployment rather than as a platform.

### What can be inspected

The value of this record is scope discipline. It shows what a system looks like when the requirement is to replace spreadsheets and paper for one practice rather than to serve a market, and it is a useful reference when deciding whether a bespoke tool or an existing product is the right answer.

## Third Eye Case Study

URL: https://federicorao.com/projects/third-eye/

Explore a mindfulness and ambient-audio mobile product combining onboarding, listening journeys, subscriptions, and maintainable product systems.

### System shape

A cross-platform mobile application for ambient sound mixing, layering selectable ambient tracks over master audio through a custom interface with GPU-accelerated, mood-reactive visuals. Client state is persisted locally for fast resume. The backend handles authentication, catalogue, mixer state, subscription entitlement, and object storage for audio, with an administrative surface for managing content.

### What can be inspected

This is the most complete mobile record on the site, covering onboarding, audio session handling, subscription state, and the administrative workflow required to keep a content catalogue current. It is an internal product, so no download, revenue, or retention figures are presented as evidence.