Skip to main content

Benchmark

When AI helps someone choose, who gets recommended?

Explore which companies AI understands and recommends when buyers ask for help choosing.

Select a dot to inspect a company.

All members in view with their method-specific indices or status
Company Understood Recommended Status
Beekdal Energie beekdal-energie.example 41 7 Valid result
Bergrug Energie bergrug-energie.example 56 25 Valid result
Dijkgraaf Energie dijkgraaf-energie.example 35 24 Valid result
Duinpan Energie duinpan-energie.example 25 1 Valid result
Eikenhout Energie eikenhout-energie.example 38 8 Valid result
Grachtenpand Energie grachtenpand-energie.example 37 3 Valid result
Hanzepoort Energie hanzepoort-energie.example 12 26 Valid result
Havenkom Energie havenkom-energie.example 61 28 Valid result
Heidezicht Energie heidezicht-energie.example 42 24 Valid result
Kadewerk Energie kadewerk-energie.example 26 3 Valid result
Kompasroos Energie kompasroos-energie.example 12 4 Valid result
Kustlijn Energie kustlijn-energie.example 28 12 Valid result
Lindelaan Energie lindelaan-energie.example 54 34 Valid result
Molenwiek Energie molenwiek-energie.example 40 1 Valid result
Noorderlicht Energie noorderlicht-energie.example 28 2 Valid result
Polderlicht Energie polderlicht-energie.example 46 5 Valid result
Rietveld Energie rietveld-energie.example 65 16 Valid result
Sluisweg Energie sluisweg-energie.example 54 4 Valid result
Stadspoort Energie stadspoort-energie.example 36 3 Valid result
Veenweide Energie veenweide-energie.example 36 23 Valid result
Vestingwal Energie vestingwal-energie.example 43 15 Valid result
Vuurtoren Energie vuurtoren-energie.example 45 16 Valid result
Waterlinie Energie waterlinie-energie.example 52 31 Valid result
Zandloper Energie zandloper-energie.example 30 46 Valid result
Zilvermeeuw Energie zilvermeeuw-energie.example 28 9 Valid result
Zuiderzee Energie zuiderzee-energie.example 14 1 Valid result

Energy and installation · 26 valid results · Illustrative dataset · 24–29 Aug 2026

Why this matters

More and more people ask AI where to buy.

It often starts with a simple question.

“I need payroll software for 80 people. We already use Exact. What should I look at?”

A few seconds later, AI gives a few options.

The buyer may never have searched for those companies. May never have visited their websites. May not even have known they existed.

Most companies are still optimising for being found. AI changes the question: when an AI system helps someone choose, which companies does it understand well enough to recommend?

Rough Diamonds makes that measurable.

What Rough Diamonds shows

AI first needs to understand what a company does.

Before AI can recommend a company, it needs to understand what that company offers. The better AI understands it, the better it can recommend that company to someone looking for what it offers.

Rough Diamonds therefore measures two things: AI Understanding and AI Recommendation.

This shows what a company can improve to be recommended more often than a competitor.

The measurement model

One prompt gives you an answer. Rough Diamonds gives you a measurement.

Rough Diamonds uses a purpose-built measurement model to test AI Understanding and AI Recommendation systematically.

See how the measurement works

  1. One measurement, multiple signals

    Multiple signals, one measurement.

  2. Semantic profile testing

    Identity, offer and audience.

  3. Buyer-situation testing

    Recommendation in real buying situations.

  4. Stability check

    The same picture across multiple answers.

  5. Fact consistency

    Important facts remain consistent.

  6. Pattern, not one answer

    Pattern over one isolated response.

Rough Diamonds Report

What about my company?

A Rough Diamonds Report shows how AI understands your company, when it recommends you, and where the picture can be clearer.

Any company. Any market. Any language.