Redesign the workflow before choosing the technology

What the sources contribute

The Deloitte–ServiceNow workflow outlook links technology transformation with process transformation, governance and business value. Deloitte’s agentic-transformation research examines process and workforce readiness among participating US organisations already at least piloting agents. Its findings should retain that sample context.

Microsoft’s Power Automate guidance also treats monitoring as ongoing operational work. This is product-specific guidance, not proof that a particular platform is right for every workflow.

Execution Clarity’s interpretation

Map what the workflow is trying to achieve and where people wait, reconcile information or handle exceptions. Establish the baseline before proposing a benefit. Some work may be eliminated; other work may need clearer policy, fewer handoffs or better data before technology helps.

Then compare the mechanisms. Stable rules may suit deterministic automation. Variable inputs may justify AI assistance. A bounded multi-step task may justify an agent if its authority and recovery path are clear. Human judgment may remain the right answer.

Define ownership after launch alongside the pilot design. Specify how failures are detected, who intervenes and how the workflow returns to a safe operating state. Scaling should depend on evidence of net value and acceptable control, not a successful demonstration alone.

A question for leadership

If this workflow ran faster tomorrow, would it remove the underlying friction—or move the same problem through the organisation more quickly?