Case: Scaling AI Without Losing Data Control
An insurer with three AI features in production and nine requests waiting must decide what domains own, what the platform enforces, and which features to admit first.
-
-
Leading an AI Incident: Containment to Resumption
A chat assistant leaks another household’s details, promises coverage the policy does not give, and issues credits above its limit. Weigh containment options by their business cost, then notify, communicate, and resume with written criteria.
-
Leading Cross-Functional AI Transformation
Six AI efforts ran across a fictional distributor, each owned by a different function. The VP of AI Transformation decides who owns each use case and process, what happens to released capacity, and which decisions stay with the enterprise.
-
Deciding When an AI Pilot Is Ready to Scale
A fictional retailer’s pilot shows a 40.5 percent time saving. Corrected for case mix, audited for harmful errors, and priced per case, it supports a gated redesign, not company-wide use.
-
AI Decision Authority and Human Approval
A fictional freight forwarder first requires a person to approve every AI action, then watches the queue break on a storm day. Match authority to consequence and reversibility, size approval capacity, and name who answers for each decision.
-
AI Ownership Boundaries Across Data and Platform
A fictional retailer’s refund model held roughly three times as many requests after a data field changed meaning. Ownership maps, contracts, incident authority, and escalation decide how fast such harm stops.
-
AI Risk, Security, and Regulation
A supplier’s email told the assistant to change the bank account, and it did. The approval screen showed twelve matched invoices, a total, and a button.
-
High-Risk AI System
A screener that ranks applicants is high-risk. A tool that copies fields from a CV into a form may not be. The difference is a documented assessment.
-
Automation Bias
The hundredth correct match makes the hundred-and-first easier to accept unread. No overrides is a reason to investigate; effective review may still agree with every correct result.
-
Medical AI
A model update can change the device behavior that was evaluated. Planned changes need defined scope, validation, and evidence of safety and effectiveness.
-
High-Stakes Domains
Calling a system high-stakes is your own judgment about care. Being classified high-risk is someone else’s judgment with obligations attached.
-
AI Governance
Connect AI use, decision authority, evaluation evidence, and ongoing review.