Skip to main content
All cases

AI Operating System

The numbers of the business on one screen, reconciled with the books.

Client: Agency delivering AI systems to ecommerce brands and multi-location businesses

The result

8 / 8
data sources connected
7 / 7
modules live for the first client
3×
counting error in ad purchases, detected and fixed
Oct 30, 2026
first module for the second client signed off

Figures as of October 30, 2026.

We hold no figures for hours saved or for the effect on the end clients' sales or bookings.

The problem

The first end client, an online store, needed its margin and ad return to be exact, not estimates stitched together by hand from several tools.

The second, a chain of three locations with about 20 users, saw bookings falling and needed to know which customers to reach again, and when.

What we built

  • For the online store: 8 connected sources (store, email, accounting, ads, analytics, search and social) and 7 modules, including profitability, ad performance, site monitoring and a report every morning.
  • The system's margin reconciled against the client's accounting for the test quarter.
  • One error found and fixed along the way: purchases attributed to an ad platform had been counted three times across the whole history.
  • For the multi-location chain: the first module, with 7 components and role-based permissions, centered on customer retention.
  • Deployed in the cloud, with scheduled data refreshes.

Tools

PythonFlaskSQLiteRailwayClaudeShopifyXeroMeta AdsGoogle Ads

Want a system like this in your company?

Let's talk for 15 minutes about your company.

You tell me what you want solved and I tell you what AI can do today, how long it takes and whether it makes sense for you. If it is not a fit, I say so on the same call.