A corporate marketing director choosing an agency asks AI about risk: the cost of a wrong choice, sector-specific cases and return on investment, hidden costs, data security, pilot projects and how to defend the decision internally. In a simulation Recro built for a marketing agency, the agency wasn’t mentioned in any of 32 answers. The competitors that stood out were recommended through editorial agency lists, case studies with numbers and a transparent engagement model. There’s a clear distance between the language agencies use about themselves and the language buyers ask in.
Picture the marketing director of an IT and systems integration company. Organic growth has slowed, competitors are moving faster in B2B marketing. They want to work with a new agency, but a past agency switch went badly. The sales director is wary of outside agencies, and the finance team is strict about budget.
This person was one of the key people in a simulation Recro built for a marketing agency. None of the eight questions they asked AI was “Which is the most creative agency?”
Which questions did the marketing director ask?
When we summarized the questions of the simulation’s two marketing-leader personas by decision stage, this picture emerged:
| Stage | Core of the question | The concern behind it |
|---|---|---|
| Problem | What does standing still cost in market share? | Justifying change internally |
| Discovery | Are there neutral guides comparing agencies focused on our sector? | Not trusting sales-driven showcases |
| Proof | What concrete results and ROI do these agencies share with similar companies? | A sector case to convince the sales team |
| Business case | What are sector benchmarks for lead conversion and payback period? | Bringing numbers to finance |
| Risk | Which hidden costs and out-of-scope items come up beyond the monthly fee? | Not repeating past surprises |
| Compliance | How does an agency fit our CRM and data protection rules, and who is liable? | Legal and IT approval |
| Validation | Where can I get unfiltered views on agencies from peers? | Seeing behind the showcase |
| Decision | What’s the ideal length and success measure for a pilot? How do I answer internal objections? | Testing without committing the full budget, defending the decision |
Almost every question is about risk, proof and internal defense. Creative capability is taken as a precondition; what decides is how the agency reduces risk.
What did the simulation of the never-mentioned agency show?
The agency we researched wasn’t mentioned in any of 32 runs. The model recognized it neither by name nor by description. In the same period, four different agencies stood out as the dominant competitor on different days and for different personas. The reports’ shared diagnosis:
- Editorial lists decided discovery. Competitors appeared in lists prepared by industry magazines and agency directories. The source links included marketing publications, an international agency directory and an executive community site. Ahrefs’ research on ChatGPT sources shows the same pattern: “best X” lists make up 43.8% of cited pages, and 80.9% of those lists are third-party publications.
- Case studies with numbers won comparison. Competitors’ concrete success stories were quoted as proof of their “expertise” claims.
- Transparency won the decision. Agencies that clearly described their reporting model and cost structure pulled ahead in “transparent reporting” and “profitability model” questions.
The distance between agency language and buyer language
Agency websites usually speak through portfolio, creative approach and service lists. The questions a marketing director asks AI use a different vocabulary: payback period, out-of-scope work, data liability, pilot project, service-level commitment, internal objection.
This is one of the clearest examples of the difference between a keyword and a key person. An agency ranking well for “B2B digital agency” may answer none of the eight questions of the marketing director behind that keyword. AI, meanwhile, surfaces the sources that can answer those eight questions.
How can agencies answer decision questions?
- Publish sector-specific cases. Without the client’s name if needed: sector, company size, starting point, work done, measurable outcome and timeline.
- Describe the engagement and cost model. State clearly what the retainer covers, what’s out of scope and how extra costs arise.
- Offer a pilot option. Define its length, scope and success measure.
- Write out your data and compliance process. Explain your role under data protection law, your CRM integration approach and your service-level commitments.
- Get into editorial lists and directories. Map which lists are cited in discovery.
- Give the marketing director material for internal defense. An in-house vs. agency comparison, a first-90-days plan, the impact on the sales team.
We covered how price and cost questions are answered in general in what AI says when buyers ask about price, and the shortlist that changes by person in from keywords to key people.
How does Recro make this gap visible?
The Recro Insight Model measures an agency’s visibility through the decision journey of the marketing leader who will choose it. It shows at which stage and in which question the agency drops, and which proof puts the competitor ahead.
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.
For agency selection, each pillar produces the following:
- The persona builder sets up the person who will choose the agency, with their past experience, internal pressures and sector.
- The question builder produces this person’s questions from problem to decision, in their language and natural flow.
- The report builder reads the agency’s standing at each stage, the dominant competitor and the links that source the conversation.
- The action recommendation builder sets out which equivalent content to produce against standout competitor content and which lists to enter.
From a Recro simulation · anonymized
- Persona: IT company marketing director and corporate deputy marketing GM · Turkish
- Question: An eight-stage agency-selection journey from problem to decision
- Inclusion: Not mentioned in any of 32 runs; four different competitors dominated.
- Competitors’ proof: Editorial agency lists, case studies with numbers, transparent reporting model.
- Class: Authority GAP + Content GAP
- Action: Sector-specific anonymized cases; cost and scope model; pilot definition; entry into agency lists.
Executive summary
- A marketing director choosing an agency asks AI about risk: the cost of a wrong choice, cases, payback, hidden costs, data compliance, pilots and internal defense.
- In a Recro simulation, an agency wasn’t mentioned in any of 32 answers; competitors were recommended through editorial lists, cases with numbers and a transparent model.
- Until the gap between agencies’ portfolio language and buyers’ risk language closes, ranking for a keyword doesn’t mean appearing in the decision question.
Frequently asked questions
Is the creative portfolio no longer important?
It matters, as a precondition. In our simulations, decision questions centered on risk, proof and transparency more than portfolio.
How does an agency that can’t name clients share cases?
Anonymized cases describing sector, company size, starting point, work done and measurable outcome are strong proof even without the client’s name.
Which agency lists matter?
The ones cited in discovery questions in your sector. In our simulation, marketing publications’ lists and international agency directories stood out; measurement sets the order.
Who asks questions besides the marketing director?
The sales director, finance and legal or IT. In the simulation, the marketing director’s questions already anticipated these people’s objections.
What data are these findings based on?
A Recro simulation built for a marketing agency, within 460 question-answer runs between April and September 2026. The data is anonymized.
When the marketing director who will choose you asks AI, is your agency in the answer? Recro builds a sample simulation with your buyer’s decision journey 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)
- Ahrefs · Do Self-Promotional “Best” Lists Boost ChatGPT Visibility?
This article was prepared with AI assistance and published under the review of Mehmet Semih İpek.
