Does AI recommend your competitors rather than your company?

A few years ago, just appearing on Google was enough. A well-optimised listing with sufficient reviews often ended up on the first page, and Google (which worked like a simple directory) would also display other results so users could pick and choose.

However, this approach no longer cuts it. Now, when a user searches for “the best accountant for a startup in London”, they don’t expect to work their way through a list of ten or so results. They ask an AI assistant a question and receive a personalized reply.

And that reply may be your competitor’s name. Partoo explains how it works!

Traditional search vs. AI query: What’s changed?

This isn’t so much about a change in how results are visually displayed: it’s about a complete overhaul in the way AI ranks businesses.

For years, Google evaluated companies on mechanical criteria: proximity, keywords, number of reviews, average rating…

AI, on the other hand, doesn’t classify: it interprets. It crawls thousands of reviews and mentions to extract a synthesised reputation, which it then uses to answer your prospects’ questions.

In traditional searches, coming in the top 4 was a good outcome. Conversational AI, however, only recommends one or two options, so you’ll be completely invisible if you rank any lower.

When choosing, AI won’t just look at average ratings, but what your customers have written about you. A 5-star rating without feedback doesn’t give these bots usable information. On the other hand, when a crawler repeatedly finds written feedback, such as “staff are incredibly patient with children”, it draws a reliable, trustworthy conclusion.

The content of the reviews, therefore, drives the recommendation: if your reviews lack content, AI won’t have anything to say about your brand… and will recommend a competitor with customer reviews that describe an actual experience. In short, if your competitors’ opinions are more descriptive than yours, AI has more material to highlight them.

Vague reviews equals invisible brand!

Is your brand ready for AI? Download the checklist

What are the 5 indicators that show your brand can be cited by AI? Download Partoo’s checklist to find out!

But what does AI actually analyse when scouring your company’s reviews?

AI doesn’t start by taking a look at your website. It first looks at what customers have written about you on your Google profile, TripAdvisor, Facebook reviews… And it then uses this raw material to form an opinion about your brand.

In practical terms, it relies on four signals to structure its analysis:

In concrete terms, if all your reviews boil down to variations of “good value for money” or “I recommend “, AI won’t find anything that sets you apart. It will choose a competitor whose customers have made the effort to describe a real experience that allows it to generate a trustworthy, personalized recommendation.

How do you conduct a competitive audit of your AI positioning?

The easiest way to see how AI perceives you is to directly ask it. Open your favorite LLM and ask:

“Give me 3 reasons why a customer should choose [Competitor Name] over [Your Brand Name] in [City].”

The answer you’ll get isn’t an opinion: it’s an accurate reflection of what AI has learned from your customers and those of your competitors. If the LLM says your competitor is “more responsive” or “more suitable for large companies”, it’s because that perception is rooted in the data it analysed. You’re seeing exactly how AI presents your brand to prospective customers, often for the very first time.

Self-assessment grid

To objectively measure your perception gap, use the following scoring grid:

The higher your score, the bigger the perception gap, and it means you need to act swiftly!

A high score doesn’t mean that your offer is worse than your competitor’s: it means that AI doesn’t have the necessary elements to choose you. This point is critical because it implies that the problem isn’t your product or service; it’s the “material” that your customers have left for AI to pick up and then use.