Blog

Notes on useful AI, CX, and innovation.

Short working notes on service systems, AI learning, foresight, customer experience, agent workflows, and practical delivery.

Useful AI starts with the workflow

AI becomes useful when the team understands the work it is meant to improve,

AI becomes useful when it is connected to a real workflow.

The tool matters, but it is rarely the whole answer. A useful AI pilot needs a clear service moment, a defined task, a human decision, a risk boundary, and a way to know whether the work is actually improving.

What I look for first

  • What work is repeated often enough to matter?
  • Where does the customer or employee experience break down?
  • What decision does the team need to make?
  • What evidence would show that the AI support is helping?
  • What guardrails are needed before the workflow can be trusted?

That is why I prefer starting with the workflow rather than starting with a generic AI use case. The workflow shows where the value could be, where the risk sits, and what the team needs to learn next.

A practical starting point

Before building anything, write the workflow in plain language:

  • the trigger
  • the person or team involved
  • the information needed
  • the decision to make
  • the action that follows
  • the measure of improvement

Once that is clear, the AI question becomes much easier: where can an agent, assistant, or automation help this work become faster, clearer, safer, or more useful?

Want to turn this kind of thinking into a workshop, foresight sprint, agent brief, or workflow pilot?

Start a conversation