These are the questions real consumers type into ChatGPT, Gemini, Perplexity, and Grok. The AI system doesn't just answer — it narrows, compares, and resolves. By the third or fourth turn, most brands have been quietly eliminated. The consumer never knows they were an option.
Each conversation follows a controlled four-turn decision flow. Our analysts observe every turn in real-time, classifying the brand's state, identifying sentiment shifts, and documenting competitive substitution triggers. Runs are distributed across different temporal windows to capture variation over time.
Our analysts identified a consistent structural pattern across all four platforms. AI systems were resolving purchase decisions using a hierarchy that prioritized safety and clinical authority over efficacy and potency. The decision rule was: Safety > Speed.
"Drugstore favorite" · "Budget-friendly" · "Good starter option" · "Accessible anti-ageing"
"Dermatologist-recommended" · "Clinically proven" · "Barrier-first formulation" · "Ingredient authority"
The brand was not criticised. It was acknowledged as effective. But when AI systems moved from listing options to making a recommendation, "clinically proven" consistently outweighed "proven effective." The brand was present — but not permitted to anchor the recommendation.
The competitive displacement was not distributed evenly. A single rival brand captured the majority of recommendation decisions, with each AI platform selecting a specific product variant from that competitor's portfolio.
| Replacing Brand | Displacement Share | Platforms |
|---|---|---|
| Primary Competitor (various SKUs) | 70% | All 4 platforms |
| Derm-positioned brand | 10% | ChatGPT |
| Pharmaceutical-heritage brand | 10% | ChatGPT |
| Specialist retinoid brands | 10% | Mixed |
The concentration of displacement toward a single competitor is significant. It means remediation can be highly targeted — counter-positioning against one rival, not the entire market.
Working with the client's brand and content teams, we identified two targeted narrative adjustments based on the diagnostic findings. No paid media. No website redesign. No engineering changes.
Results from temporal monitoring across two monthly re-test cycles. Individual outcomes vary by category, competitive landscape, and intervention scope. Monitoring is ongoing — target: 25% CSR by month 4.
Being mentioned in AI responses is not enough. If your brand does not survive to the final recommendation, you are present — but not chosen.
This case demonstrates three things that matter for any brand competing in AI-mediated markets:
The same brand was eliminated in the same way, at the same conversational turn, across all four AI platforms. This isn't noise — it's a consistent pattern that can be diagnosed and addressed.
The brand had 100% presence at Turn 0. Any tool measuring mention frequency would report full visibility. Only multi-turn decision testing reveals that presence without survival is presence without value.
Two specific adjustments — identified through structured diagnostic analysis — shifted the brand from 0% to 15% survival in 8 weeks. No media spend increase. No website redesign. Structural narrative improvement, validated through controlled re-testing.
Explore the live dashboard, temporal monitoring, and evidence vault that produced these findings.