We sat with their team, walked every existing flow, and wrote the brief on day three.
From a 90-day windowto a body of work.
Designed, evaluated, and integrated an LLM copilot into the ops console. Reduced time-per-decision by 42% across 11 enterprise customers in the first quarter post-launch. Below is the long version — the brief, the approach, and the numbers we put on the wall after launch.
NovaTech's enterprise customers spent hours triaging shipment exceptions by hand. An LLM demo looked magical and shipped nothing useful — every answer was a screenshot.
We redesigned the copilot as a typed action surface with an eval harness built into CI. Every suggestion came with a confidence band, an undo, and a one-click export to the customer's playbook.
42% faster decisions across 11 enterprise customers in the first quarter, an NPS jump of 23 points, and a copilot the ops team actually trusts.
