Paused
Could we build one important, concrete user need (promo code creation) on top of the Phase 3 architecture, using the existing POC stack augmented with agentic AI capabilities, so that clients could log in to the client portal and create promo codes themselves?
This was a deliberately narrow POC. Rather than building on the full Harbour and Crane infrastructure from Phase 3, the scope was kept minimal: GCP as the hosting environment and the ESco Myriad API as the integration point.
The implementation drew on two complementary threads:
The goal was a working demo of end-to-end promo code creation: a client logs into the portal, goes through an AI-assisted flow, and a promo code is created in the system.
The feature was built and reached demo-ready state by April 2026. The technical work was completed: the client portal flow and the agentic AI layer functioned as intended.
However, when the feature was demoed to Anthem a client the team was pitching for and ultimately did not win the response was ambivalent. There was no clear excitement or enthusiasm, and the demo did not generate the commercial pull needed to justify continued investment.
Louise decided to pause the programme in April 2026 following the Anthem demo. The feature worked technically, but the client response did not provide sufficient confidence to continue at the same level of investment. The pause was not a rejection of the underlying architecture; it was a recognition that the commercial signal needed to justify the next step had not materialised.
In hindsight, the decision by the senior leadership team to pursue promo code creation as the Phase 4 focus was probably the wrong call. The choice was influenced by feedback from a client the team was pitching for and ultimately did not win which means the strategic direction of the phase was shaped by a commercial outcome that never arrived.
The Current State page (May 2026 onwards) reflects this pause: the programme is on hold while the business considers its options.