Too many ideas, not enough clarity
Most organisations now have a long list of AI ideas — from leadership, from vendors, from enthusiastic staff — and little clarity on which are worth doing. Some ideas would save thousands of hours; others are solutions looking for a problem. Some are blocked by data that doesn’t exist; others could be live in a month. Without a structured view, budget goes to the loudest idea rather than the best one.
What the audit looks at
We look at value (how much time, cost or revenue is at stake and how confident we are), feasibility (whether current AI can reliably do the task with your data), readiness (whether the data, systems access, process clarity and owners exist) and risk (privacy, regulation, customer impact and what happens when the AI is wrong).
We talk to the people who do the work, not just the people who manage it — the best opportunities are usually hiding in daily routines. And we look at real data and systems, because a use case that looks perfect on a whiteboard can fall apart when you discover the data lives in scanned PDFs or nobody can grant API access.
Evidence, not opinions
For the one or two most promising ideas, we run quick technical spikes on real data. A few days of prototyping often answers the questions a month of discussion can’t: is accuracy good enough, what will it cost to run, what’s the hard part?
A roadmap you can execute
The output is a prioritised roadmap: quick wins to build momentum, foundational work such as data clean-up or access setup, and larger initiatives sequenced by dependency. Each item has an effort range, an owner, a success metric and a decision on build, buy or configure.