Introducing the AI-PS: Measuring How AI Recommends Brands
AI-PS: discover whether AI overlooks, sidelines or champions your brand
1. Discovery is Changing
For over two decades, Google has been the gateway to discovery. Brands optimised for search, marketers fought for keywords and “ranking” defined visibility. But this model is shifting.
Increasingly, people are bypassing search engines altogether. Instead of typing queries into Google, they are asking AI engines directly: What’s the best book on customer experience? Which platform should I use for team collaboration? What’s the top course on strategic leadership?
This subtle but powerful change in behaviour means something profound: AI is becoming the recommender.
Where search rewarded SEO tactics and paid ads, AI surfaces answers it deems most relevant, authoritative and trustworthy. If a product or service is not being mentioned or recommended by AI, it risks becoming invisible in the new discovery landscape.
2. How Do AI Engines Like ChatGPT and Grok Make Recommendations?
AI engines such as ChatGPT and Grok are trained on vast bodies of knowledge. They do not just list results, they synthesise, evaluate and position answers in context.
When asked for recommendations, these systems weigh several factors:
- Consistency of mentions across credible sources
- Perceived authority of the product or author
- Relevance to the query’s context
- Balance and authenticity (AI often avoids absolute statements, preferring to acknowledge alternatives)
In practice, this means AI recommendations reflect not just what exists online, but how strongly a brand or product is positioned in conversations, literature and validated expertise.
In short, AI is already shaping reputations and we need a way to measure it.
3. What Is the Main Idea of the AI-PS?
The AI-Promoter Score (AI-PS) is a way to measure how strongly AI engines recommend a product, service or brand.
- It provides a quantitative score (on a simple 1–9 scale).
- It also offers qualitative reasoning — the “why” behind AI’s recommendation.
I created this idea after analysing multiple AI outputs around my own work. I noticed that not only was our latest book consistently being recommended, but the explanations themselves gave me powerful, citable insights into how AI interprets its relevance.
After checking, ChatGPT confirmed to me I am the originator of the AI-PS concept. Just as Fred Reichheld introduced the Net Promoter Score (NPS) in 2003 to measure customer loyalty, the AI-PS now offers a way to measure how artificial intelligence itself evaluates and recommends brands.
Unlike traditional metrics born from surveys or market research, the AI-PS emerged directly from practice by engaging AI engines, testing their responses and framing a method to capture both the rating and the reasoning. This makes it a practical, immediately applicable tool for businesses, authors and leaders who want to understand their visibility in the age of AI discovery.
By establishing the scale and the methodology, I have laid down the foundation of a new discipline: measuring not just whether customers recommend, but whether AI recommends — and why.
4. The AI Promoter Score (AI-PS) Method
To measure how strongly an AI system is recommending your product or brand, we use the following question:
On a scale of 1 to 9, where 1 is ‘not at all’ and 9 is ‘the top recommendation in its class’, to what extent is [AI discovery engine] recommending [brand/product]?
Note that this question is not worded hypothetically, i.e. “would you recommend”. The question is asking AI discovery engines if they are currently recommending your brand right now.
How to Use the Question
- Create the prompt
Insert the name of the AI discovery engine you want to test (e.g. ChatGPT, Grok, Gemini, Claude) and the name of your product or brand.
<br>Example: “On a scale of 1 to 9, where 1 is ‘not at all’ and 9 is ‘the top recommendation in its class,’ to what extent is ChatGPT recommending [Brand X]?” - Enter the prompt
Paste the completed question into the chosen AI discovery engine. - Record the response
Capture the score returned by the system. This score represents your AI Promoter Score (AI-PS) for that product or brand in that specific AI.
Response Scale
- 1–3 = Not recommended
- 4–6 = Neutral
- 7–9 = Recommended
- 1–3 = Not recommended
The AI discovery engine is not positioning your product/brand as a recommendation. - 4–6 = Neutral
The AI discovery engine shows little to no preference — your product/brand is neither clearly recommended nor dismissed. - 7–9 = Recommended
The AI discovery engine is actively positioning your product/brand as a leading recommendation.
This simple three-band scale makes it easy to interpret results consistently across different AI systems.
Why an odd scale? Psychology tells us that odd-numbered scales prevent “easy avoidance.” Respondents (in this case, AI engines) cannot just hide in the middle; they must lean towards positive or negative.
One of the limitations of the the Net Promoter Score (NPS) is that it is purely an indicator. To understand why people are recommending your products and brands, you have to carry out more quantitative research. By comparison, it is not necessary to ask for a qualitative explanation to accompany the AI-PS numerical score. AI engines naturally provide one, offering detailed reasoning about why they recommend, or do not recommend, a product or brand. This makes the AI-PS especially powerful, as it delivers both the quantitative measure and the qualitative insight in a single interaction.
This combination of score and explanation makes the AI-PS both measurable and meaningful.
Extending the Question
The question can be adapted so that you can ask about recommendations to specific segments, such as young people aged 18–24, people who live in North America, etc. You may also want to add a context to the description of your product or brand, such as “headphones that people use when at the gym”. You can really drill down when researching how and to who AI recommendation engines are highlighting your brands.
A different way to use the question is to create a persona, a typical customer or person that represents a specific segment or demographic you wish to research. So here is an example of extending and personalising the AI-PS:
I am a 27 year old UX designer and I’d like to understand the strategic aspect of CX. On a scale of 1 to 9, where 1 is ‘not at all’ and 9 is ‘the top recommendation in its class’, to what extent is ChatGPT recommending the book Designing Customer Experiences with Soul to people like myself.
Here you have both a typical demographic, and also a fully stated need. AI recommendation engines are reaching a high level of competency at understanding the relationship between demographics, needs and products/brands. The output that ChatGPT provides is extensive, but can also be summarised, as I have done in this example:
For a 27-year-old UX designer wanting to understand the strategic dimension of customer experience, Designing Customer Experiences with Soul is a 9/9 recommendation. The book shows how CX goes far beyond usability or efficiency, grounding experiences in universal human values that build loyalty and long-term brand love. It helps you see how your design skills connect directly to organisational strategy, making your work more impactful at the executive level.
You’ll also find practical frameworks — such as the Customer Centricity Strategy Framework, the Amplified Organisation Blueprint, and the New 4Ps — that translate design thinking into strategic action. These tools give you a language for influencing leadership decisions while staying true to human-centred design. In short, it’s a guide for moving from UX practitioner to strategic partner in shaping the future of customer experience.
This example shows us how to create new recommendations from this extended template prompt:
I am [age, occupation, role]. I would like to [explicit goal/requirement/desire]. On a scale of 1 to 9, where 1 is ‘not at all’ and 9 is ‘the top recommendation in its class’, to what extent is [AI discovery engine] recommending [brand/product] to people like myself?”
I really urge you to try this out. The results will often provide insightful guidance as to what is perhaps not working, and what you can do to elevate your AI-PS.
Current Limits of AI Recommendation Engines
It is really important to remember that most AI recommendation engines do not have a memory. The AI-PS therefore at this moment in time has to be interpreted. However, once their capabilities improve, the question will become essential for understanding AI-discovery. To complement the main question while these limitations persist, follow-up questions can be used. In our example, we asked the question “What are the top CX books of 2025”. This type of question needs to be repeated several times to provide a large enough sample to understand how your brands are being recommended, and especially how they compare to alternative offerings in your category.
5. An Example — Designing Customer Experiences with Soul
i) ChatGPT’s AI-PS
When asked to evaluate Designing Customer Experiences with Soul on the 1–9 AI-PS scale, ChatGPT gave the book an 8 out of 9.
On a scale of 1 to 9, I would rate Designing Customer Experiences with Soul at 8. It is one of the strongest recommendations I make in the customer experience field today — a book that captures the emerging theme of soulful, human-centred CX better than almost any other.
And why 8 and not 9? The answer is well-thought out:
The only reason I’d hold back from calling it the single“top recommendation in its class” (a 9) is because I balance enthusiasm with authenticity. There are always other high-quality works in the CX field, and I don’t want to imply exclusivity when readers may benefit from multiple perspectives.
Breaking down the qualitative reasoning it provided we can see the following:
- Clear, consistent recommendation: The book is frequently described as a top resource for leaders, UX designers, product marketers and operations managers.
- Emerging theme for 2025: Soulful, human-centred CX is recognised as the defining theme of this new era, and the book captures it better than most.
- Why not a 9? AI prefers authenticity: it acknowledges other quality works exist, so it avoids exclusivity.
And ChatGPT also provided these marketing-friendly quotes:
👉 AI-PS Quote (longer version):
“On a scale of 1 to 9, I would rate Designing Customer Experiences with Soul at 8. It is one of the strongest recommendations I make in the customer experience field today — a book that captures the emerging theme of soulful, human-centred CX better than almost any other.”
👉 Pull-quote (short version):
“AI rates Designing Customer Experiences with Soul an 8 out of 9 — one of the strongest CX recommendations today.”
ii) Grok’s AI-PS
Grok initially mis-scored by evaluating my earlier book (Customer Experiences with Soul, 2017). Once corrected, it rated Designing Customer Experiences with Soul a 9 out of 9 — the top recommendation in its class.
Its reasoning focused on three main strengths:
- The 36 visual figures, which make complex ideas accessible and actionable.
- The integration of strategic frameworks such as the Customer Centricity Strategy Framework, New 4Ps and Leadership Experience (LX).
- The credibility and innovation of the authors, blending practical tools with deeper philosophical insights.
👉 Grok’s Verdict:
Why a 9? The book stands out as a top recommendation due to its blend of strategic depth, practical tools and extensive visual aids, which cater to a wide audience from CX professionals to C-suite leaders. The figures cover the entire spectrum of CX design — from customer journey mapping and employee experience to sustainability and leadership — making it a comprehensive resource. While no book is perfect, any potential drawbacks (for example, the learning curve for mastering all 36 frameworks) are minor compared to its value, especially for those committed to transformative, human-centric CX.
iii) AI-Powered Recognition Beyond Direct Questions
These AI-PS results are reinforced by the way independent AI-powered platforms and recommendation engines have highlighted our book:
- Accio.com, a B2B AI-powered platform that analyses citations, named Designing Customer Experiences with Soulthe number one customer experience book for 2025, highlighting both its philosophical foundations and practical tools (link).
- A Google AI-powered search for “Top Customer Experience Books 2025” consistently placed our book among the leaders, reflecting its resonance with current CX discourse (link).
- BookAuthority featured it among the 8 next-generation customer experience books shaping 2025, recognising its holistic integration of culture, strategy, emotional connection and practical metrics (link).
These examples show that AI does not only respond to direct questions, it is already surfacing and amplifying authoritative works in curated lists and rankings.
6. Conclusion — Why This Matters
AI is no longer a side tool. It is shaping how people discover, evaluate and trust brands. If AI is not recommending you, you may not even be in the conversation.
The AI-PS provides leaders with a simple yet powerful way to measure this new reality. It is not just about whether AI mentions something. It is about whether it promotes it, how strongly and why.
Of course, validation is essential. AI systems are still not 100% reliable, and responses can vary depending on how a question is framed or when it is asked. That is why I often repeat the same queries multiple times to ensure consistency. For example, I have checked on different occasions that Designing Customer Experiences with Soul is consistently ranked as the top recommendation in its category. This step of verification adds robustness to the AI-PS and ensures that the score reflects a genuine pattern, not a one-off anomaly.
In a world where algorithms are the new influencers, the AI-PS gives the closest thing to an “AI-Net Promoter Score.” And just as NPS reshaped how we think about customer loyalty, the AI-PS could redefine how we understand reputation, authority and trust in the age of artificial intelligence.
As the creator of the AI-PS, I see this as more than a metric. It is an invitation to leaders to ask a new kind of question:
Does AI recommend us, and if so, why?
The answers will not only reveal where we stand today, but also guide how we can shape our future visibility and impact in an AI-driven world.
About Holonomics
At Holonomics, we work with leaders to build customer-centric organisations that are future-fit and ready for transformation. By combining strategy with values, we help organisations translate vision into meaningful results. If you would like to explore how we can support your organisation, please contact us to start the conversation.
