Is your due-diligence workflow ready for AI?
Test the source hierarchy, missing-evidence policy, human decision boundary, and production controls before choosing a model or automating a business outcome.
Free membership. After you confirm your email, Ghost opens the download page automatically. Already a member? Sign in and open the download page.
Five practical worksheets
Why this exists
The expensive mistakes happen before the model call.
A polished AI summary can still sit on top of the wrong company, an unavailable source, an unverified sanctions match, or a business decision nobody authorised the system to make.
Source sprawl
The team has a list of websites and APIs but no approved hierarchy for official facts, screening signals, public leads, freshness, or substitution.
False clean results
A timeout, weak entity match, or empty response is later read as though the check completed and found nothing adverse.
Authority drift
A research assistant quietly becomes a quote, refer, decline, or compliance-clearance engine without approved rules or accountable human ownership.
What is inside
Turn the vague readiness debate into five visible artifacts.
The workbook is deliberately practical. Each section should leave your team with an agreed map, policy, owner, or next action.
Research map
Define the business questions and mandatory checks before selecting sources.
Source authority
Name the primary source, backup, freshness rule, failure path, and owner for each fact.
Missing evidence
Separate complete, partial, missing, unavailable, contradictory, and adverse states.
Human authority
Show what AI recommends, what code enforces, and what an authorised person decides.
Readiness score
Score 32 controls, apply seven critical gates, then choose build, narrow, or stop.
Critical gates
A good total cannot rescue a missing control.
The workbook prevents teams from averaging away the failures that matter most. Any critical gate scored zero blocks production consideration.
Use the score to prioritize work, not to certify the workflow.
A readiness score is useful only when the evidence behind it remains visible. The critical-gate overlay stops a strong architecture score from hiding an undefined business authority or missing sanctions policy.
Who should use it
This is a team workshop, not an engineering self-assessment.
The strongest answers need business policy, specialist judgement, governance, operating ownership, and technical evidence in the same room.
Business and product
Underwriting leaders, product owners, business analysts, and process owners defining scope, appetite, and success.
Risk and governance
Compliance, financial-crime, legal, model-risk, audit, and governance teams defining verification and accountability.
Engineering and operations
Architects, AI engineers, data teams, QA, and service owners implementing controls and monitoring the workflow.
The decision
Finish with one of three honest outcomes.
The workshop should not end with a vague action list. Decide whether the next move is to build a bounded capability, narrow the control surface, or stop until a critical dependency is resolved.
The business policy and evidence controls are defined well enough for the next bounded capability.
Reduce products, decisions, dimensions, users, or source coverage until the workflow can be controlled.
A critical source, authority rule, legal basis, or operating owner remains unresolved.
Get the free checklist.
Join AI Tool Stack as a free member. After email confirmation, Ghost will take you directly to the members-only download page. You will also receive practical project chapters about evidence, evaluation, workflows, review boundaries, and production hardening.
The checklist and the linked application are research and pre-screening frameworks. They do not provide a binding underwriting, legal, compliance, sanctions, credit, or investment decision.