Building the bank's AI-enabled lending platform.

I picked up a key strategic outcome on Lloyds' Commercial Lending Platform, leading around 80 people across product, engineering, design, credit and data. The brief was demanding enough on its own: build the bank's first automated SME decisioning across secured and unsecured lending, plus a GenAI financial spreading product, both visible at COO level.

The harder problem was that the platform teams were not yet working as one. Dates came top down, priorities moved week to week, and trust between the delivery groups was thin. Fixing that was not a side quest, it was the precondition for shipping anything.

I had two problems to solve at once, delivery and people. On delivery, I chose to build the SME decisioning capability incrementally rather than as a big-bang launch, so real value reached customers early and we could learn while we built. On the GenAI product, I insisted on a robust human-in-the-loop workflow and a proper evaluation harness before we scaled anything, so the output could be trusted.

Underneath both, I worked to turn the hostility between platform teams into a group that actually delivered together.

The bank's first automated SME decisioning workflow, shipped in increments so lending teams saw value from month one. Alongside it, a GenAI financial-spreading product with human-in-the-loop review and evaluations built in from the start, not bolted on later.

Alongside both, I ran the discovery shaping the bank's next SME lending workflow, challenging existing policy and process to define an AI-first, data-driven end-to-end journey.

The automated decisioning workflow generated £5.8M in its first month live, and automated acceptance went from nothing to half of all applications. It shipped incrementally across payment lines, charge cards and credit cards, then extended to overdrafts and term lending, so lending teams saw value from the first release rather than the last.

The GenAI financial spreading product is now used across the entire 200-person trading relationship manager population, having gone from pilot to full rollout on the strength of the feedback loop rather than a launch date.

Delivery was recognised at COO level. And the teams that started out pulling in different directions now work together properly, which I am as proud of as the numbers.

In a top-down, politically charged environment, your delivery strategy is also a leadership tool. Shipping incrementally bought trust that dictated dates never could. And with GenAI, the discipline of evals and human-in-the-loop before scaling is not a brake on delivery, it is the thing that makes the value safe to claim.