The AI engines rarely agree about the same brand
What this rests on. Early read: 5 audits of 4 brands. Each engine answered 5 to 9 buyer questions per audit, none of them naming the brand.
What we looked at
When a buyer asks an AI engine for recommendations in a category, does it name your brand? We counted, for each engine, in how many of those answers the audited brand appeared at all, then compared the engine that named the brand most with the one that named it least.
What we saw
The gap was large every time. In one audit a single engine named the brand in 5 of 8 answers while another never did. In another, one engine named it in all 7 answers and a second in only 2.
It was not always the same engine on top. The engine that was most generous to one brand was among the least generous to another.
What it suggests
A single combined score hides the story. Which engine your buyers actually use matters, and a brand can look healthy on one engine and invisible on another. Looking at each engine separately shows where the work is.
| Audit | Most generous engine | Least generous engine |
|---|---|---|
| A | 5 of 8 answers | 0 of 8 answers |
| B | 3 of 7 answers | 0 of 8 answers |
| C | 7 of 7 answers | 2 of 7 answers |
| D | 2 of 8 answers | 0 of 8 answers |
| E | 4 of 9 answers | 0 of 5 answers |
What this can't tell you
- The numbers are small: five audits, four brands, a handful of answers per engine. Treat this as a first look, not a finding.
- The audits used different questions, categories and markets, so the audits are not directly comparable with each other.
- Audits B and A are the same brand, run on different days.
What you can try
- Ask a few of your own buyers which assistant they use before you decide which engine to prioritise.
- Check each engine on its own. Look at the gap between engines, not just the average.
Get the next note by email
And tell us what to investigate next.