Putting It into Practice
2026Introduce and Adopt
Redesign roles, routines and incentives so that consultants and client teams can work confidently with agents.
The main adoption barrier is unlikely to be prompt writing. It is the redistribution of work, influence and accountability.
Explain the change honestly
People need a concrete answer to three questions: What work will the agent do? What decisions remain mine? How will my performance and role change?
If the message is limited to productivity, experienced consultants may hear cost reduction and junior colleagues may see the disappearance of their learning path. A credible program addresses both.
Redesign work, not just tasks
When an agent prepares a fit-to-standard analysis, the consultant’s role shifts toward framing, challenge, exception handling and decision facilitation. That requires different allocation, review practices and commercial models.
The client role changes too. Business owners must provide timely decisions, validate context and take visible accountability for policy choices.
Build agent literacy by role
Everyone needs a basic understanding of limitations, evidence and safe use. Beyond that, learning should be role-specific:
- leaders need to set decision rights and interpret assurance.
- workstream leads need to supervise agent workflows and exceptions.
- specialists need to curate skills, rules and evaluation cases.
- platform teams need to operate secure, observable services.
- assurance teams need to test both outcomes and controls.
Preserve professional development
Consulting has traditionally developed judgment through preparation, analysis and repeated exposure to detail. If agents perform more of that work, organizations need deliberate alternatives: simulations, supervised exception handling, rotations, trace reviews and earlier participation in decisions.
Measure adoption as behavior
Logins and generated documents are weak measures. Look for reduced rework, earlier risk discovery, use of traceable evidence, appropriate escalation and growing confidence among accountable users.
Adoption is successful when the mesh becomes part of how the program works—and people know when not to rely on it.