Zac Nash, our Data Engineering Director, shares what the team learned building Inspectre, Fresh Egg’s AI analytics agent. He covers why they deliberately made it less intelligent, why its context goes out of date faster than you’d think, and why the most trusted part of the tool doesn’t use AI at all.
Pointing AI at your analytics data is easy. Trusting the answers isn't.
Replay available
Watch the 45-minute replay on what it really takes to get answers you can trust from AI analytics.
Want something like Inspectre working on your own data? Talk to Zac and the team
What you'll take away
Why choosing the AI model is rarely the hard part, and where the real work is
Which analytics questions AI is good for, and which it should never answer
Why the fix that made Inspectre work was knowledge about the data, not more architecture
How to keep your AI’s context current, and who should own doing it
Why telling an AI to “be careful” fails, and what to put in place instead
Two things you can do this week: turn on the GA4 BigQuery export, and write down your metric definitions and who owns them
Who should watch
Useful if you're responsible for analytics data or for the decisions made from it:
Head of Data / Head of Analytics
Head of Digital / Digital Director
Marketing Director / CMO
Head of Marketing
Your session expert
Zac Nash, Data Engineering Director, Fresh Egg
Zac leads data engineering at Fresh Egg, where he's spent the past three years building the agency's technical foundations, Inspectre included. Before that, he worked as a data science consultant, building customer segmentation and marketing mix models for brands across sectors.
