An AI answer is shaped far more by who is asking, and under what conditions, than by the keyword in the question. In a simulation Recro built for a coffee supply brand, the brand was mentioned in 8 of 8 runs for a tourism group’s F&B director, and never for the procurement directors of hotel, café and restaurant chains. Within the same category, 236 runs produced 50 different “dominant competitors”. A keyword shows how your site performs; a key person shows whether you’re present where the decision is made.
For a coffee supplier’s marketing team, the target keywords are obvious: “corporate coffee supplier”, “HORECA coffee”, “wholesale coffee”. A site that ranks well for them looks successful in reports. But who is typing them?
A hotel chain’s F&B director thinks about guest experience, brand perception and barista training. A café chain’s procurement director looks at minimum orders, price stability and delivery frequency. A retail chain’s category director asks about private-label production capacity. All three are searching the same category. Each asks AI a different question, gets a different answer and sees a different shortlist.
How did the same brand change by person in our simulations?
In a simulation Recro built for a coffee supply brand, questions asked on behalf of different decision-makers produced this picture:
| Key person | Priority conditions | Brand mentions | Type of competitor that stood out |
|---|---|---|---|
| Tourism and restaurant group F&B director | Experience, service, equipment support | 8 of 8 runs | Corporate coffee service brands |
| National retail chain category director | Private-label production, scale | 8 of 8 runs | Industrial roasters |
| Premium hotel chain F&B and supply director | International standards, consistency | 0 | Global coffee brands |
| Café chain procurement director | Specialty coffee, flexible minimum order | 0 | Local specialty roasters |
| Restaurant chain procurement and operations director | Price stability, delivery, service | 0 | Global professional service brands |
Same brand, same category, same period. Results range from 100% to 0%, and each person meets a different type of competitor. Across the simulation, 236 runs produced 50 different “dominant competitor” names.
The pattern repeated in other sectors. In a marine equipment simulation, the brand appeared in 25% of answers for some personas and 0% for others; the dominant competitor was a domestic manufacturer for one persona and a global player for another. In an enterprise software category, 112 runs produced 28 different dominant competitors.
What does a keyword show, and what does a key person show?
A keyword is a phrase typed into a search engine. It measures how well a site answers that phrase, and it still matters for SEO.
A key person is the decision-maker with buying intent: their role, sector, company size, the situation that triggers the question and their acceptance threshold. In the AI era, buyers don’t type keywords; they describe their situation: “We’re a café chain with a hundred branches, we want protection from price swings; which specialty roasters offer flexible minimum orders?” The model weighs each condition in that sentence and builds the shortlist for that person.
Same category, two key people
Hotel chain F&B director: “Which corporate coffee suppliers offer barista training and equipment support that would raise the guest experience?”
Café chain procurement director: “Which suppliers offer annual price locking against green-bean price swings and flexible minimum orders?”
Both questions sit behind the keyword “corporate coffee supplier”. In the first, proof of experience and service wins; in the second, proof of pricing policy and flexibility.
Why isn’t a single “target audience” enough?
In B2B, decisions aren’t made by one person. According to 6sense, the average buying group has 11 people. According to Gartner, buyers spend only 17% of their buying time meeting suppliers. Each member of the group asks their own question, and AI offers each of them a different shortlist.
A definition like “target audience: hotels, restaurants and cafés” erases this difference. The insight that a brand is strong with F&B directors and invisible to procurement directors only surfaces when each key person is measured separately. We explained how this spread is analyzed in Persona GAP analysis.
How do you build a key-person map?
- List the deals you won and lost in the last two years. Who was in each decision?
- Write four things for each role. The situation that triggers the question, the condition they look for in the answer, their acceptance threshold and their veto point.
- Write each role’s question in their own words. Without your brand’s category names.
- Measure each role separately. Brand mentions, the dominant competitor and the rationale should be read for each person.
- Build content per person. Proof of experience for the F&B director, pricing policy for procurement, risk management for finance.
We covered how decision questions are mapped in what is a decision question.
How does Recro make this gap visible?
The Recro Insight Model starts measurement with the key person with buying intent. Which people the brand is strong with, which it’s invisible to and which competitor each person meets appear in the same report.
One simulation · 4 key pillars · 90+ steps
What sets Recro apart is four pillars built from scratch for each brand. Together they form a single simulation of 90+ steps, with each pillar producing the input for the next.
- Persona builder: Decision-maker profiles that represent the brand’s real buyers, specific to its sector and sales structure.
- Question builder: The brand-specific decision questions these people ask AI along their buying journey.
- Report builder: A report that reads the answers against the brand’s goals and competitors, together with the sources behind them.
- Action recommendation builder: A prioritized list of actions that closes the sector-specific signal gaps.
In the key-person approach, each pillar produces the following:
- The persona builder defines the brand’s real buyer structure by role, sector, scale and triggering situation. Single persona–many questions, many personas–single question and many personas–many questions setups cover the whole buying group.
- The question builder produces each person’s question with their conditions, in a natural buying-journey flow.
- The report builder reads brand mentions, the dominant competitor and the rationale per person, and assesses KPIs set against the brand’s own goals per person.
- The action recommendation builder ranks which proof and content matter most for the key people where the brand is weak.
From a Recro simulation · anonymized
- Brand: Coffee supply brand · HORECA and retail · Turkish
- Strong with: Tourism group F&B director and retail category director: 8/8.
- Invisible to: Hotel, café and restaurant chain procurement directors: 0.
- Pattern: 50 different dominant competitors across 236 runs; the competitor type changes by person.
- Class: Persona GAP
- Action: Produce proof content on the pricing policy, minimum order and delivery topics procurement directors ask about.
Executive summary
- The AI answer is shaped far more by the asker’s role and conditions than by the keyword in the question.
- In a Recro simulation, the same brand was mentioned 8/8 for two decision-makers and 0 for three; 236 runs produced 50 different dominant competitors.
- Measurement and content should be built around key people with buying intent: each role’s question, conditions and competitor read separately.
Frequently asked questions
Is keyword work unnecessary now?
No; it shows how a site performs in search engines. But it doesn’t show which decision-maker an AI answer recommends the brand to. Use both.
Is a key person the same as a persona?
A persona is the modeled form of a key person. The key person is the real decision-maker with buying intent; the persona makes their role, conditions and question measurable.
How many key people should we define?
At least one for every critical role in the buying group. In our simulations, when the role changed, both visibility and the dominant competitor changed.
Doesn’t such a volatile dominant competitor make measurement meaningless?
It does if you look at a single answer. With repeated measurement, which competitor type stands out for which person appears as a stable pattern.
What data are these findings based on?
Recro simulations for 6 brands between April and September 2026, with 460 question-answer runs. The data is anonymized and shows observed patterns.
Which key people is your brand strong with, and which is it invisible to? Recro builds a sample simulation for each role in your buying group and shares the result in a demo report.
Request a demo insight report →
Sources
- Recro simulation data · April–September 2026 · 6 brands, 460 question-answer runs, 73 reports (internal, anonymized; run on a single AI model)
- 6sense · The B2B Buyer Experience Report for 2024
- Gartner · The B2B Buying Journey
This article was prepared with AI assistance and published under the review of Mehmet Semih İpek.
