Field note · Sip Champagnes
Most consultancy case studies start with a story and end with a vague claim. This one starts with the numbers, because the numbers are unusual enough to lead with.
Sip Champagnes is a UK wholesale distributor. It is also where ODEG’s working method was tested — by being applied, not described. We redesigned how the business operated and built the AI workflows, automations and internal tools to support the new shape. The figures below are taken from Sip’s own audited management accounts, comparing the nine months after switch-on to the nine months before.
On a slightly smaller top line. Margin doing the work, not volume.
Year on year, against the same period a year earlier.
Cleaner purchasing, fewer order-entry errors, tighter supplier terms.
From a sizeable loss to near break-even.
Same team, more throughput. AI took the manual work, not the people.
The whole P&L moved. Not a one-off, not a single line item.
What changed first.
Before any AI was switched on, the operating model was redesigned. Where work began, where it ended, who owned each handover, how exceptions were treated. Without that, AI just speeds up the wrong process.
The biggest unlocks were the most ordinary. Order processing moved from inboxes and re-typing to a single review surface. Stock and forecasting moved from a Friday spreadsheet to a continuously-updated picture. Invoice chasing moved from awkward phone calls to polite, automatic, on-time reminders. None of those are glamorous — they are simply where the friction sat.
“We didn’t fire anyone. We just stopped asking them to do work a machine could do. They started closing deals and managing customers instead of fighting spreadsheets.”
What held into the next quarter.
It is easy to move a P&L for a single quarter. The harder test is whether the trajectory holds. Through Q1 2026 it did:
- Gross margin sustained above 35%, nine months after switch-on.
- Operating run-rate at break-even, on track for post-tax profit this year.
- Order-processing automation running above 95% straight-through on priced lines.
The point of this journal entry is not to claim the same numbers are possible everywhere. They will not be. The point is to be specific about what is possible when strategy and implementation sit in the same room — and to be specific in audited numbers, not vague language.
More from the journal
Essay · Operating model
Why most automation projects do not move the P&L.
After enough engagements, a pattern: businesses spend on AI and automation, see operational improvement, and then watch the bottom line refuse to move. Here is why.
Read entryProcess · The ODEG method
From founder-led ops to a real operating model.
The hardest growth transition for a founder-led business is the one nobody puts on a slide: going from ‘the founder is the operating model’ to having an operating model the business runs on.
Read entry