You've been listed in the professional register for twenty years, but when a client asks Perplexity who the reliable professionals in your field are, the AI cites colleagues with fewer credentials than you. It's not unfairness: the AI doesn't read your résumé, it reads structured signals — and your registration in the professional register, in the format it exists in today, is invisible to the models. Making it readable is a simple, targeted fix that turns your most solid credential into a visible advantage even in AI answers.
The question isn’t whether you’re listed in the professional register. It’s whether the AI knows it. And almost always, it doesn’t.
Let me explain what I mean. An engineering firm in Piacenza, duly registered with the provincial Order of Engineers, with three partners, twenty years of history, major structural projects in the area. Open Perplexity, ask “structural engineering firms in Piacenza with experience in construction supervision”. The firm doesn’t show up. Smaller, younger competitors with fewer credentials do.
The reason is almost always the same: the firm exists in the professional register, but it doesn’t exist as a verified entity in the graph the AI consults when it answers. It’s registered in a database the AI doesn’t read, or doesn’t link to the website, or doesn’t recognize as an official identity.
In the articles in this series I’ve told you how the AI builds its graph of local and vertical entities. Here I’ll explain the ground level: the professional registry as machine-readable proof of existence. Without this, everything else — schema, citations, reviews — starts uphill.
What it means to be an official-registry entity for an AI model
In the world of research on Entity Recognition applied to YMYL (Your Money Your Life — health, finance, legal, structural engineering) contexts, the documented principle is that AI models use credential verification sources before citing a professional in response to queries with safety or compliance implications. It’s not an editorial choice: it’s a risk-mitigation constraint built into the retrieval system.
It follows that for an engineering firm, the Companies Register of the Chamber of Commerce and the provincial Order of Engineers are the two sources the AI treats as “databases of truth” regarding your existence. The website is a claim. The professional register is proof.
Mine is a deduction from the principle of credential verification in the YMYL domain, applied to the Italian reality of professional registers. I’m telling you straight because it isn’t a documented fact: it’s a logical consequence of how AI systems handle high-risk queries.
The logic is simple. If the AI cites the wrong engineering firm in response to a question about a structural calculation or construction supervision, the consequences aren’t editorial: they’re legal, possibly a matter of public safety. The systems are calibrated to be conservative on these queries. They prefer to cite a less relevant candidate with a verified identity over a more relevant one with an ambiguous identity. It’s a precautionary bias, and it works against anyone who hasn’t sorted out their identity in the official registers.
Why this layer sits upstream of everything else
When the AI processes the question “structural engineering firms in Piacenza”, it does three things in sequence: it extracts the entities from the query (profession + location), retrieves candidates from its graph, and verifies that those candidates are recognizable entities.
Your website with an “About us” page and three photos of the engineers doesn’t make you a recognizable entity. It makes you a candidate. You become an entity when the system finds multiple independent sources confirming you: the Companies Register with your business name, the Order of Engineers with the partners’ registration numbers, your certified email (PEC) that matches, the registered office that matches.
I’ve already written about how author recognition works inside AI systems (you’ll find it in How AI recognizes who wrote a piece of content) and about how E-E-A-T weighs on retrieval (in E-E-A-T translated for AI). The professional registry is the layer below even that: it’s where the system verifies that you and your firm really exist, before it even evaluates whether you’re any good.
If the Chamber of Commerce says via Garibaldi and your website says via Cavour, the AI sees two different entities and doesn’t merge them.
The test you can run in ten minutes
Open your Piacenza Chamber of Commerce website and search for your business name. Check three things.
First: correct ATECO code. For a structural engineering firm the main code is 71.12.10. If there’s a generic or wrong code, the AI won’t associate you with the right sector when it retrieves candidates for professional queries.
Second: registered office and operating office up to date. If the Chamber of Commerce says via Garibaldi and your website says via Cavour, the AI sees two different entities and doesn’t merge them.
Third: certified email (PEC) and declared address. They must match those in the website footer, on the firm’s LinkedIn profile, and on the Google Business Profile.
Then go to the provincial Order of Engineers of Piacenza. Verify that each partner’s registration number is publicly visible and that on your website, on the professional’s page, that number is written in plain text (not just in the downloadable CV). “Eng. Mario Rossi, registered with the Order of Engineers of Piacenza no. 1234 since 2006”. It’s one line. It makes an enormous difference.
Finally, if you have technical skills or someone who helps you, add to the schema the `sameAs` field pointing to your firm’s page on the Chamber of Commerce register and the `identifier` field with the professional-register registration number. You can check whether the schema is well formed with Google’s Rich Results Test: paste in your homepage URL, look for the “Organization” block and check that the fields are recognized.
If you have no one to help you with the schema side, the first step — the one that matters most — is still to write the registration numbers in plain text on the professionals’ pages. The AI reads the text before the schema.
Write each professional’s registration number from the professional register in plain text on their dedicated page
The mistakes I see most often in professional firms
Five recurring patterns, valid for engineers but also for architects, accountants, notaries, lawyers.
First: registration number only in the PDF CV. The CV is downloadable but the AI rarely downloads and processes it. The number must be in the HTML text of the professional’s page.
Second: fluctuating business name. On the website “Studio Rossi Ingegneri”, at the Chamber of Commerce “Rossi e Associati Srl”, on LinkedIn “Studio Rossi & Partners”. Three entities for the AI, none with a strong signal.
Third: no explicit link to the Order. The “About us” page tells the firm’s history but never says “registered with the Order of Engineers of the Province of Piacenza”. It’s as if you didn’t want to declare it to the system.
Fourth: outdated Chamber of Commerce data. Registered office changed three years ago, updated on the website but not on the company registration record. The AI, when it compares, finds a mismatch and discards the candidate from the answer because it can’t merge the sources.
A fifth pattern I see less often but that matters when it’s there: the firm’s website without a dedicated page for the individual partners. The team page is a grid with photos and roles, but no unique URL for each engineer. The AI has nothing to grab onto to link “Eng. Mario Rossi” — an entity it might find in the professional register — to a canonical page on your site. The result is that the professional exists in the register but doesn’t exist as a web entity traceable to your firm. One page per person, with full name, title, registration number, areas of expertise and contacts, solves the problem.
What to do concretely, in order of impact
- Align the business name across the Chamber of Commerce, website, LinkedIn, Google Business Profile — the exact same string, commas included
- Verify and update the office, certified email (PEC), and ATECO code on the company registration record
- Write each professional’s registration number from the professional register in plain text on their dedicated page
- Add a line on the “About us”: “registered with the Order of Engineers of the Province of Piacenza”
- If you have technical support, add `sameAs` in the Organization schema pointing to your Chamber of Commerce profile, and `identifier` for the professional register
These are entry-level checks. They’re a first step: a full analysis of your positioning in AI answers requires professional tools, test queries across multiple engines, and longitudinal monitoring. But without the ground level of a certified identity, any subsequent work on schema, citations, and content starts one step lower.
Where this thread leads you
The professional registry is the proof of existence the AI consults before deciding whether you can appear in an answer to YMYL queries. Without it, you’re one website among many. With it, you’re a verified entity — and your visibility in AI answers changes accordingly.
In the next articles in this Entity & Knowledge Graph series I’ll take you further: I’ll explain how local geographic entities work for “near me” queries, how the AI builds the graph of industry entities, and why Wikidata as an anchoring node is the next step for those who have already sorted out the Chamber of Commerce and the professional register.
