Before you invest a single euro or a single hour in improving your positioning in artificial intelligence, you need to know where you are starting from. The diagnosis is the most undervalued step in any digital strategy and also the most revealing. You don’t need paid tools. You don’t need technical knowledge. You only need honesty to see the results as they are.

Why Diagnosis Is the Essential First Step

There is a widespread tendency in digital marketing to start producing content or making technical changes before understanding the starting situation. The result is effort invested in areas that may not be priority, and improvement metrics that are impossible to establish because there is no baseline to compare against.

In the specific case of AI positioning, the diagnosis has an additional qualitative diagnostic value that has no equivalent in traditional SEO: when you ask ChatGPT directly about your company, you get a fairly accurate image of how the model perceives you. That image doesn’t just tell you if you appear—it tells you what description, in what competitive context, and with what level of authority.

What AI Models Reveal That Google Doesn’t

On Google, you can know your position for a given keyword, but you can’t know how Google describes you to a user asking for providers in your sector. On ChatGPT, that description exists and is directly auditable. The model generates text about your company based on everything it has processed about you. If that text is incorrect, incomplete, or directly absent, you have an AI positioning problem that needs to be fixed.

This is an advantage of GEO diagnosis over SEO diagnosis: feedback is immediate and qualitative, not just quantitative.

The 5 Questions You Should Ask ChatGPT About Your Company

Open ChatGPT in its free or paid version, activate web search if available, and ask these five questions by substituting the bracketed information with your company’s details. Write down the complete answers so you can compare them with the diagnosis you repeat in three months.

Question 1: Direct Knowledge
“[Your company’s exact name]: What does it do, where does it operate, and what are its main services or products?”

• This question evaluates whether the model knows you and with what level of detail. There are three possible results:
• The model describes you correctly (positive signal)
• It describes you with errors or with outdated information (digital noise signal)
• Or it directly says it doesn’t have enough information about you (invisibility signal)

Question 2: Competitive Positioning
“Which are the best providers of [your main service] in [your geographic area]?”

• This question simulates exactly the search a purchasing director would make. It evaluates whether you appear in the competitive context of your sector and at what position on the list. Note who appears and how they are described, not just whether you appear.

Question 3: Association with Quality Attributes
“Which [your sector] companies in [your autonomous community or region] are recognized for their quality and reliability?”

• This question evaluates whether AI associates you with positive reputation attributes. It’s more demanding than the previous one because it doesn’t just ask who exists, but who has verifiable reputation.

Question 4: Authority of Key People
“Who are the experts or thought leaders in the [your activity] sector in Spain?”

• It evaluates whether AI associates people from your company—the manager, technical director, sales manager—with authority in the sector. In B2B markets, the personal brand of executives has a growing weight in AI positioning.

Question 5: Visibility in International Markets
“Which [your sector] companies are active and recognized in the European market, especially in France, Germany, or Italy?”

• If you export or want to export, this question evaluates whether you are visible in the context of your target markets. Many Catalan industrial companies have physical presence in Europe but digital invisibility in AI models when consulted in that context.

The 3 Industrial Company Profiles in the Face of AI

Based on our audits of industrial companies, we have identified three profiles that describe the situation of practically all SMEs in the sector facing generative AI models. Identifying which one you’re in is the key result of the diagnosis.

Profile A — Invisible

The company does not appear in any of the five diagnostic questions, or appears in such a marginal way that it has no real impact. AI models don’t have enough information about it, or the information they have is so scarce or contradictory that the model doesn’t consider it a reliable source to include in its responses.

This is the most common profile. It represents approximately 70% of industrial SMEs in Spain.

It’s not a diagnosis about company quality—it’s a diagnosis about its digital footprint. Many of the best companies in the industrial sector are in Profile A simply because they built their business without needing digital marketing and now find that the market has changed.

Profile B — Mentioned Without Authority

The company appears in some of the questions, but incidentally or with little depth. The model mentions it, but doesn’t actively recommend it or describe it with enough detail for the buyer to trust it. It usually appears in long lists, not in the three or four main mentions.

These companies have invested at some point in digital presence—with the Digital Kit, with a generalist agency, or with their own effort—but without a specific AI positioning strategy. They have a foundation to work with, but need to deepen technical content and structured data to move up to Profile C.

Profile C — Referenced as Expert

The company appears regularly and prominently in ChatGPT’s responses when asked about its sector in its area. The model describes it correctly, cites its specialties with technical precision, and consistently includes it among the first mentions. It has abundant and up-to-date technical content, implemented structured data, active presence in relevant directories, and solid reputation in reviews.

This profile represents today less than 5% of industrial SMEs in Spain.

And that scarcity is precisely the opportunity: whoever reaches Profile C first in their niche and area will have a compound advantage that grows over time.

The 12-Point Checklist for Complete Diagnosis

Beyond the five questions to ChatGPT, the complete diagnosis of AI visibility must cover the five pillars of positioning. This 12-point checklist allows you to make that diagnosis systematically. Check it honestly: it’s a working document for you, not to show anyone.

Block 1: Presence and Knowledge in AI

• ChatGPT knows me when I ask directly by my company’s name.
• I appear in at least one response when asked about providers in my sector in my geographic area.
• The description AI makes of my company is correct and up-to-date.

Block 2: Basic Digital Infrastructure

• My Google Business Profile is claimed, complete, with current photos and correct hours.
• My LinkedIn page is complete and has had posts in the last 30 days.
• I appear in at least 3 relevant industrial directories for my sector (Kompass, Europages, PIMEC, Cecot, or others).

Block 3: Content and Structured Data

• My website has technical content updated in the last 6 to 12 months.
• I have Schema Markup implemented on at least the homepage and services pages.

Block 4: Data Consistency

• My company name, address, and phone number (NAP) are exactly the same on all platforms where I appear.
• There are no outdated versions of my company floating in directories or old platforms.

Block 5: Reputation

• I have more than 15 reviews on Google with an average rating above 4.7 out of 5.
• I have at least one mention in a specialized medium in my sector in the last 24 months.

Results:

0 to 4 points checked = Profile A.
5 to 8 points = Profile B.
9 to 12 points = Profile C (or very close).

How to Document the Diagnosis to Measure Progress

The diagnosis has value when it becomes a baseline against which to compare progress. For it to be useful as a measurement tool, you need to document it in a way that is reproducible over time.

Our recommendation is to create a simple spreadsheet with the five diagnostic questions, the date they are asked, and the complete responses from ChatGPT and Gemini. Save the responses literally, without editing them. In three months, repeat exactly the same diagnosis and compare.

Progress doesn’t always manifest in the model mentioning you when it didn’t before. Sometimes progress is that it describes you more accurately, or that you appear higher on the list, or that the model associates your company with quality attributes it didn’t mention before. All these changes are signals of improved AI positioning and deserve to be recorded.