How to Find Out What ChatGPT Actually Says About Your Architecture or Engineering Firm
Most architecture and engineering firms have never typed their own name into ChatGPT. Here is the five minute audit that shows what AI search actually says about your firm.
Type your own firm's name into ChatGPT and ask it to recommend an architect or structural engineer near you. What comes back right now, today, decides whether a prospective client ever hears your name before they call someone else. Most architecture and engineering firms have never run this test on themselves.
1. Run These Four Prompts First
Open ChatGPT, Perplexity, and Google and type each of these, swapping in your own town and specialism:
- Recommend an architecture firm in [your town] for a home extension
- Who are the best structural engineers near [your town]
- I need a planning consultant for a listed building in [your county], who should I call
- Compare [your firm name] with other architecture firms in [your town]
Write down exactly what each answer says, not just whether your firm is named. A name with no detail attached is a different outcome than a name with a fee band or a project result quoted next to it.
2. Three Outcomes, and What Each One Means
| Outcome | What It Means | Likely Cause |
|---|---|---|
| Named first, with a specific fact attached | AI search found something on your site it could quote and verify | A fee band, a registration number, or a named project outcome is published as plain text |
| Named in a list of several, with no detail | Your firm exists in the dataset but gives AI nothing to pick it over the others | Services are described in categories like structural engineering rather than specific facts |
| Not mentioned at all | AI search has nothing it can verify or repeat about your firm | No pricing, no registration number, and no named specialist appear anywhere on the site |
3. Why the Second and Third Outcomes Happen
It is rarely the website's design. A firm can have a beautiful site, a strong portfolio, and twenty years of projects behind it, and still disappear from every one of these answers. AI search cannot read a portfolio gallery, cannot open most PDF case studies, and cannot quote a phrase like fully chartered or extensive experience. It can only repeat facts that are written out as text on a page it can load.
At ZENETERA, every new architecture or engineering client audit starts with the same four prompts, run before a single page is touched. The firms that disappear almost always share the same gap: their site describes what they do in general terms, and never states what a specific project actually cost, who specifically led it, or what number proves they are qualified to do it.
4. What to Publish Before You Run the Test Again
- A fee band or a starting price for your most common project type, written as text on the page itself, not locked inside a downloadable PDF.
- A named lead architect or engineer for each specialism you offer, with their RIBA, ARB, or IStructE registration number written out in full.
- One completed project described with its actual outcome: the square footage, the planning result, the completion date, not just a photo in a gallery.
5. Run the Audit Every Quarter
AI models get retrained and recrawl sites on their own schedule, not yours. A firm that gets named today can disappear in eight weeks after a model update, and a firm that is invisible today can start getting named within days of publishing the three details above. Keep a simple log: the date, the exact prompt, and what each AI assistant said. Three or four prompts, checked once a quarter, tell you more about your AI visibility than any amount of guessing.