GOVERNATLAS / THE RESEARCH LIBRARY
Independent thinking for consequential technology decisions. Field guides, practical methods and working tools to take into your next review.
START WITH THE DECISION
What it does. When you need it. How to evaluate it. Independent buying methods connected to source-backed product dossiers.
From an inventory of AI to decisions you can explain.
Explore the category ↗FIELD NOTES / 02Know what the data means. Know who can act on it.
Explore the category ↗FIELD NOTES / 03Find the failure. Understand the impact. Close the loop.
Explore the category ↗FIELD NOTES / 04Know the service. Challenge the answer. Own the decision.
Explore the category ↗THE EDITOR’S STARTING POINT
01 / FIELD GUIDECHOOSING WITH CONFIDENCE
AI governance tools can promise similar outcomes and support very different work. Start with the decisions your team needs to make, then ask for evidence you can actually inspect.
Explore the buyer’s field guideSix demonstrations to structure a vendor conversation. A working evaluation record to carry the evidence forward.
FIELD GUIDE + DOWNLOADABLE WORKSHEET
FROM QUESTION TO WORKING METHOD
Choose the problem in front of you.
Leave with a method you can use.
Choose between catalog, quality, access and stewardship capabilities. Run a concrete buying trial, challenge connector coverage and retain a defensible decision—with a private access-testing matrix.
Explore the playbookAn operating guide to feature attributions, reference populations, counterfactual feasibility and explanation stability. Includes a private, downloadable explanation review record.
Explore the playbookAn operating playbook for testing evidence exports, replacement workflows, access revocation and retained copies before renewal. Includes a private, downloadable exit rehearsal record.
Explore the playbookA practical pilot for data access governance: map enforcement paths, test human and service identities, inspect masking and export boundaries, verify revocation and preserve an operation-level decision record. Includes a private access-test matrix you can download.
Explore the playbookA practical playbook for reviewing vendor security questionnaires: define service scope, challenge unsupported answers, reconcile evidence dates, control AI-assisted reuse and record a defensible follow-up decision. Includes a private evidence register with review-date checks and a portable download.
Explore the playbookA buyer’s operating playbook for AI gateways: compare SDK and proxy boundaries, test retries and partial streams, inspect key controls, reconcile provider spend and document an exit path. Includes a private failure-testing worksheet and cost reconciliation calculator.
Explore the playbookAn implementation playbook for evaluating a consent management platform: map the tags, test six consent journeys, investigate conflicting signals and keep a versioned decision record. Includes a private, downloadable testing worksheet.
Explore the playbookA practical evaluation method for data quality and observability: define the failure, replay controlled incidents, measure investigation effort and warehouse overhead, and retain a decision record. Includes a synthetic fixture and an empty results template.
Explore the playbookAn implementation playbook for connecting AI use cases, models, suppliers, owners and decisions, with an editable inventory template and a practical reconciliation method.
Explore the playbookA practical method for constructing datasets, choosing metrics, calibrating model judges and investigating regressions without hiding critical failures inside an average score.
Explore the playbookA threat-model-first playbook for authorized AI security trials, with paired benign cases, failure accounting and replay boundaries. Includes a private, downloadable trial record.
Explore the playbookA practical distinction between finding data, tracing its dependencies and enforcing its use, with a cross-system demonstration that reveals gaps hidden by feature lists.
Explore the playbookA service-specific diligence playbook for product scope, data handling, assurance documents, operational controls, material changes and an executable exit plan.
Explore the playbook