Visa VAS

The workflows lived in people's heads and inboxes.

The corporate-strategy team ran on repeatable work that was never written down as a process. Before anything could be automated, someone had to map what actually happened between a request and a finished deliverable.

Stakeholder interviews into a repeatable operating model.

I mapped the team's workflows through stakeholder interviews and redesigned them around the AI use cases with the highest value. The output was an operating model the team could reuse, plus pilot blueprints for each use case.

Leave instructions at end of day. Find the work done in the morning.

I built traceable agent pilots on Visa's existing stack. A team member leaves instructions at the end of the day; the agent works overnight and the result is ready for review in the morning, with every step traceable back to the instruction that caused it.

Monitoring, troubleshooting, and a rollback for every pilot.

A pilot that only I could operate would have died the day the engagement ended. I delivered monitoring, troubleshooting, and rollback runbooks alongside the pilots. That is what turned them into AI-native workflows the team adopted.

• Mapped Visa VAS corporate-strategy workflows through stakeholder interviews and redesigned them around high-value AI use cases.
• A repeatable operating model with pilot blueprints for the team.
• Traceable agent pilots on Visa's existing stack that turn end-of-day instructions into morning-ready work.
• Monitoring, troubleshooting, and rollback runbooks driving team adoption of AI-native workflows.

Discovery

Stakeholder interviewsWorkflows mapped and redesigned.

Handoff

Pilot blueprints + runbooksMonitoring, troubleshooting, rollback.

Result

Team adoptionAI-native workflows on the existing stack.