27/08/2026
Plenty of retailers we talk to have solid reporting and dashboards they trust, so it surprises them when their first AI projects fall behind expectations. The reason is, more often than not, the data.
Reporting and AI need very different things from it, yet most data today is set up for reporting.
You see, reporting looks backwards. It tells you what happened, so the data gets summarised into the totals a dashboard needs.
But AI looks forwards, and to guess what a customer does next, it needs just the detail those summaries drop. It needs explicit input like who bought something, what they bought with it, and what usually follows for people like them.
Nowadays, a lot of the retail platforms we see are built on the Medallion architecture (the Bronze, Silver, and Gold layers). That works well for reporting, but AI needs another step on top, where the data holds context rather than just a structure.
We covered the whole operational strategy in just 10 minutes with Google Cloud, now viewable on demand. The link is below. Watch it at your own pace, whenever you're ready.