Skip to content
MirrorFirm

How narrow should a first customer change pilot be?

A useful first pilot follows one customer instruction through a complete dependency path, with explicit boundaries around the decisions and systems it represents.

Start with a change whose beginning and completion can be stated clearly. In utilities, that could be changing the recipient of future invoices across an account and its billing arrangement. The pilot can cover request interpretation, the controlling record and the dependent delivery setting.

Choose the boundary according to the agent capability you need to inspect. If dependency discovery is the target, provide clear authority and straightforward policy. Adding identity disputes, payment renegotiation and service relocation at the same time makes failure harder to diagnose.

Ask for a scope statement identifying included decisions, excluded systems, assumed facts and the evidence that establishes completion. A useful first sample contains enough linked state to execute the whole bounded task, rather than disconnected records that merely resemble the domain.

Expansion can follow a specific finding from the pilot: a missing relationship, a misunderstood rule or an unsupported tool operation. This keeps later additions tied to demonstrated development needs. It also makes the limits of the initial evaluation clear when results are discussed with other teams.

Scope a pilot around a measurable business outcome rather than a broad collection of disconnected change requests.

Audience:
Agent developers
Content type:
Explanation
Buyer question:
Scope and getting started

Content release: r-a220eb6e099b243b

More on Customer change handling