Nobody runs eighteen skills at once. Pick the question your board is asking, deploy the organisation that will own the answer, then follow the sequence.
Where are we on data?→ Data & AI Maturity Assessment
Why don't we trust this number?→ Data Lineage & Audit Evidence
What is bad data costing us?→ Cost of Bad Data
What are we actually buying?→ IT & Data Due Diligence
Who owns this decision?→ Data Operating Model & Agent Org
What did the AI spend produce?→ AI Delivery Value
Every skill is run by a role. If the roles don't exist, the output has nowhere to land, so step zero sets decision rights and ownership before the first assessment.
Every output below needs a named owner. Decision rights, RACI and agent roles come first.
Two diagnostics, one sitting each. A scored picture of the data and the organisation, and a red-flag list.
A defensible dollar figure on what the gaps cost, computed from your own tickets and logs. This is what funds everything after it.
Route the defects to owners. Fix the definitions. The top three red flags become workstreams with names against them. For an agentic AI use case, Day Zero turns the baseline into a hundred-day plan.
Re-run and trend: Patch Ratio, exposure, maturity. The second run is where the value is, because movement is what a board acts on.
The annual baseline. Run it, act on it, re-run it next year and show the movement.
For a platform decision or a major re-architecture.
For deal teams, and anyone buying, selling or separating a business.
For proving what the AI spend is actually producing, and that it behaves.
Baseline, then re-run to show movement
Monitoring: trend the numbers, not the anecdotes
A deal, a platform decision, a migration
Tell Harveer the question and he'll point you to the first skill.