Your AI has an answer for everything. That’s not always good.

Retail has committed to AI. 82% of retail and CPG executives are raising their AI spend in the next 12 months. 16.5% can point to a return on what they have already spent. The distance between those two numbers is not a model problem. An agent does not arrive with its own understanding of your consumers. It borrows yours, and it can only act on what it is able to read. Most of what is being sold right now cannot read enough to justify the line item.

Issue 02 goes after that gap. A marketing team who asked their AI a straight question about conversions and got back something they were not looking for. What the numbers say about where agentic adoption actually sits in retail, every figure linked to its source so you can check the lot. And a set of prompts you can run against whatever AI you already have, with a way to grade what comes back. None of it needs our software to be useful.

 

What’s in Issue 02

• The interview: the question the Society of London Theatre marketing team put to their AI, and the gap in their own data it found on the way to answering it

• The evidence: where agentic adoption actually sits in retail and CPG, with every figure linked to its primary source

• The anatomy: what an agent can read when your data sits in one place, and what it misses when it does not

• Generative or agentic: the difference most vendors are currently blurring

• The prompt card: prompts that ask your AI for decisions instead of summaries, four tells that it is guessing, and five questions that say whether your stack can answer them