Friction lives in the operation, not in the committee
Real frictions are known by the people who execute: the analyst who reprocesses a request, the supervisor who reconciles by hand, the advisor who loses a sale to a response time.
The first job is to gather them with those people, in their own language, without asking them to propose the solution.
Prioritize on two axes
Not every friction deserves a pilot. We order them on an impact versus feasibility matrix and pick the ones that move a known indicator and can be tested without touching the core of the operation.
The output of that prioritization is a challenge: a short statement with scope, an internal owner, success criteria and the indicator that will move.
A well written challenge attracts better solutions
When the challenge is well defined, matching with startups stops being a search by category and becomes a comparison by real delivery capacity.
And when the pilot ends, the decision is simple, because the criterion was agreed before starting and not after seeing the results.
