Small enough to evaluate. Useful enough to matter.
A pilot.
Not a leap of faith.
Choose one workflow. Agree what good looks like. Learn from the
actual work.
A first conversation is for understanding fit, not promising an
automatic transformation. Feasibility, access requirements,
scope, timing and commercial terms need to be agreed for each
project.
Talk through a possible pilot
↗
Useful questions to bring
Where does the work slow down?
Bring one concrete example: the trigger, the steps, the people
involved and the output you need. A rough sketch is enough to
begin.
What should stay in human hands?
Identify decisions that need approval, sensitive inputs, access
boundaries and the consequences of an incorrect result.
How would we know it is working?
Describe the result you would accept and the examples you would
use to evaluate it. We can discuss a pilot against those
criteria.