Stopping a programme heading for four times its budget.

I joined with the programme already in flight: an agency on time and materials was re-architecting the core data model with no product involvement. The technical debt was real, but nobody had asked users or clinicians what the new platform needed to do.

The programme was personally sponsored to the leadership team by its senior technical owner. It had around 100 unmeasurable KPIs, no architecture documentation and no user discovery, and it was about to redesign clinical scheduling and the pathway that manages controlled medications.

• Raised my concerns formally and in writing, without waiting for consensus.
• Insisted on clinical sign-off before any further build.
• Proposed a two-week pause and an independent audit, which was agreed.
• Co-led the audit with the Director of Delivery, who did most of the technical work.

I'd make the challenge again; the risks were too serious not to. But it cost me my relationship with the sponsor. Alongside escalating, I wish I'd invested more one-to-one time with them, helping them find a way to change course that they could own, so that we could have gone into the leadership discussion as a team rather than on opposite sides. They weren't open to challenge on this programme, so it may not have worked, but I should have tried harder.

The audit found no intellectual property, warranty or termination provisions in the contract, clinical safety standards (DCB0160 and DCB0129) unmet, inadequate testing and no rollback plan. I recommended termination, and it was accepted by the leadership team and the board.

It also left something lasting: a clinical-safety gate for any future platform change. Afterwards there was still pressure to present the abandoned work as a reusable asset, which is its own lesson in how hard it is to stop things.