Starter
One or two queues, activated one at a time.
Two assistants and a monthly action allowance sized for a single team.
AI assistants for repeat work, with clear boundaries and your team in control of important decisions.
Catalyst needs one repetitive workflow, a named owner and source data someone already trusts. Without those three the pilot has nothing to measure, and a pilot that cannot be measured is not worth running.
One or two queues, activated one at a time.
Two assistants and a monthly action allowance sized for a single team.
A department with several recurring queues.
Five assistants and a larger action allowance across connected workflows.
Multiple departments running in parallel.
Ten assistants and the largest action allowance, with usage limits agreed in advance.
Measure the current volume, the time spent and the error pattern.
Connect approved inputs and write down the authority boundaries.
Review sample outputs and exceptions before anything runs live.
Monitor quality, approval delays and usage. Tune before expanding.
The next step
We’ll determine whether there is a system worth building — and tell you if there isn’t.