Working Theory · Rachel McBride

← Working frames · Rowan Advisory ↗

Where the tool stops

Five things building Mappa taught me about keeping people in command.

Mappa is a living map of a firm: what it is working on, who it knows, where it sits in its market and how the parts connect. The record is held separately from the AI that reads it.

I audited the places where the instrument stops being reliable on its own.

1Mark every guess

Mappa can suggest that a relationship exists or that someone may be able to help. The suggestion is shown differently from a confirmed fact. A person checks it before it enters the record.

2Let named people decide what is true

Only a few people can confirm the firm’s record. That limits speed and protects quality. Adding more information without a clear owner would make the map larger and less trustworthy.

3Say what the map cannot see

Mappa shows the relationships around a firm. It does not show everything that makes the firm work. A manufacturer still has production, stock and logistics. The page for each map states its blind spots so nobody mistakes the instrument for the whole business.

4Keep the method separate from the product

The map itself will not suit every organisation. The method travels: keep a checkable record, let AI work over it, and keep a person responsible for the input and the action.

5Prove that the record is current

A sourced fact can still become stale. Each item needs a source, a date and a way to check whether it remains safe to act on. This is the part of the work that is still developing.

Mappa has been useful because its limits are visible. The test for any AI system is similar: can you see where its answer came from, where its confidence ends and who remains accountable?

ip.rowanadvisory.co.nz/where-the-tool-stops/