Where a regional bank should start with AI

Many regional banks get the most traction by starting AI with one specific workflow and governing the handful of definitions that workflow depends on. A full data cleanup first can take a long time before anyone sees results. Jumping straight into AI tools without agreed definitions risks answers that don't match official reports. Picking one workflow and its key figures can deliver value sooner and lays a foundation for the next one.

Three common starting points

1. Data cleanup first. Thorough, but it can take a long time before anyone sees AI value.
2. AI tools first. Fast to launch, but answers may conflict with official reports, and trust can drop quickly when they do.
3. One workflow plus its definitions. Narrow enough to finish, grounded enough to check.

Good first workflows

1. Commercial deposit and loan growth reporting.
2. Recurring management reporting, such as a monthly performance package.
3. Answering ad hoc executive questions about a small set of key figures.

How to start

1. Pick a workflow with a clear owner and a measurable time cost.
2. List the figures it depends on.
3. Have the right business owner approve one definition for each.
4. Confirm the AI use fits your bank's AI policy and vendor review process.
5. Connect an AI assistant to those definitions. Check its answers against the official report before relying on them.
6. Measure the result: hours saved, accuracy against the official figures, and how often people use it.
7. Repeat with the next workflow.

How Preql helps

Preql helps teams capture the business logic behind a workflow's figures, so owners can review and approve it. That logic becomes a reusable layer the next workflow can build on.

FAQ

Do we need a Chief AI Officer first?

Not necessarily. You need an owner for the workflow and its definitions, and the AI use should fit within your bank's existing governance and risk processes.

What about AI for customer service or fraud?

Those are valid starting points, and many banks begin there, often through existing vendors. Finance and reporting workflows are worth considering because their figures can be checked against official reports, which makes accuracy easy to measure.