The Agentic Workforce, Part 2: Put the Agents on the Scorecard
If agents are doing work, leaders need a scorecard for ownership, quality, risk, escalation, and value creation.
Source basis: McKinsey, The Future of Work Is Agentic
Once agents enter the workflow, the next mistake is treating them as invisible.
If a human does work, we know how to ask management questions. Who owns it? How is quality reviewed? What happens when something goes wrong? How do we measure performance?
Agents need the same operating questions, translated for non-human capacity.
I would put agents on a scorecard with five categories:
- Ownership: named human owner, business purpose, and approved use case.
- Quality: accuracy, rework, override rate, and exception patterns.
- Risk: data access, confidentiality, bias, compliance, and client impact.
- Escalation: when the agent stops, flags, or hands work to a person.
- Value: cycle time, capacity created, cost avoided, or quality improved.
This does not mean pretending agents are employees. It means refusing to let work happen outside the management system.
The executive question is not, “How many agents do we have?”
The better question is, “Which agents are creating trusted operating leverage, and how do we know?”