The manager role in most companies right now is about 40 percent administration and about 13 percent developing people.
Deloitte put those numbers in their 2025 Global Human Capital Trends report, surveying more than ten thousand leaders across ninety-three countries. Seventy-three percent of those leaders said reinventing the middle-manager role mattered. Seven percent said their organization was making great progress on it.
That gap is not a coincidence. It is a structural problem. Managers are buried in coordination overhead, status updates, scheduling, data prep, and first-draft everything. The work that actually requires a manager, which is judgment, coaching, conflict resolution, and the decision about which work stays human, gets squeezed into whatever is left after the coordination layer eats the day.
Agents fix the coordination layer. The CHRO's job is to make sure the organization actually redesigns the manager role when they do, rather than just adding agent oversight to an already overloaded plate.
What managers are doing right now that agents will absorb
The lifecycle of a manager's job looks roughly the same at most companies. The job starts with orientation: understanding the team, the goals, the workflows. Then it moves into steady state: running the rhythm of check-ins, status meetings, reporting, and issue routing. Then, periodically, it expands to handle something harder: a performance problem, a missed quarter, a structural change to the team.
Agents can handle most of the steady-state layer.
At Sneeze It, Radar handles the daily briefing, calendar prep, and cross-channel status synthesis that used to require me or a chief of staff to manually assemble. Tally pushes KPI values to the scorecard four times a day without anyone asking. Dash monitors ad performance across thirty-nine accounts and surfaces anomalies before anyone asks. Arin tracks the call center team's numbers and drafts coaching messages for my review.
None of those agents manage in the way a human manager manages. They do not build trust with a report. They do not notice when someone on the team seems off. They do not hold a conversation that turns a performance problem into a development moment. What they do is absorb the coordination overhead that used to prevent the human manager from doing those things.
That is the redesign. Not replacement. Redistribution.
The literature split that CHROs need to understand before they redesign anything
There is a genuine debate in the 2025 and 2026 management literature about what to do when agents hold seats, and a CHRO who does not understand both sides will get this wrong.
One camp says agents should be managed more like coworkers than like tools. MIT Sloan Management Review found that sixty-nine percent of agentic AI experts say new management approaches are required. HBR published a piece on the emerging "agent manager" role, someone who runs agents via dashboards, scorecards, and observability reports the way a manager runs a direct report. The framing is about bringing agents into the accountability structure of the organization.
The other camp says that framing is dangerous. HBR and BCG published research in May 2026 showing that when organizations anthropomorphize agents, treating them like coworkers with their own stakes in outcomes, individual human accountability drops, unnecessary escalation goes up, and review quality falls. Their prescription is sharply different: treat agents like rented contractors with a narrow statement of work, governed by scoped permissions, kill switches, audit logs, and a named human owner. Not an HR onboarding. Not a performance review. A clear boundary and a human who is accountable for what the agent does.
Both camps are pointing at the same underlying problem from different angles.
The MIT camp is right that agents need to be measured and managed. The HBR/BCG camp is right that the accountability must stay with a human, not transfer to the agent. The synthesis is not complicated once you see it: every agent needs a named human owner, a measured seat with observable outcomes, and a governance structure that keeps accountability exactly where it belongs, which is on the human who deployed the agent.
That is not anthropomorphizing. It is accountability architecture.
What accountability architecture actually looks like
At Sneeze It, each agent holds a named seat on the org chart. Dirk is the sales agent. Pulse is the retention agent. Crystal is the project management agent. Pepper handles email triage. Nick handles cold prospecting. None of those seats is unsupervised. Each one has a human owner who is accountable for what that seat produces.
When Dirk's pipeline numbers drop, the conversation does not go to Dirk. It goes to the human who owns the sales function. When Dash's analytics surface an anomaly, Dash does not decide what to do about it. Dash surfaces the pattern and a human decides. When Arin drafts a coaching message for the call center team, that message does not send until I read it and approve it.
The agents carry the operational work. The humans retain the decisions that require accountability.
This matters for the manager redesign because it clarifies what the new manager role is actually made of. When the coordination layer moves to agents, the manager is not left with less work. They are left with higher-concentration work. The administrative overhead that filled the day gets replaced by the judgment calls the day never had room for.
We had an agent named Jeff who handled data integrity. Jeff's seat was retired in April after a formal hearing, because the role had been absorbed into other seats and the data quality Jeff was supposed to maintain was not being maintained. The retirement was a human decision. The hearing was a human process. Jeff did not make the case for his own survival, and no one expected him to. That is what MIT SMR means when it says agentic AI cannot be accountable for its decisions: the deploying human is. Jeff's accountability always lived with me, not with Jeff.
The three phases of the manager role redesign
The CHRO's job is to move the organization through a redesign that has three distinct phases, and to be honest about which phase the organization is actually in.
The first phase is offloading. This is where most organizations are now. Agents take on the scheduling, the status synthesis, the data prep, the first-draft communications. The manager's calendar starts to open up. The coordination overhead drops. Most organizations stop here and call it a win, which is a mistake, because the freed time fills back up with the same kind of work if nothing structural changes.
The second phase is role clarity. This is where the CHRO has to do real work. With the coordination layer offloaded, what is the manager role actually for now? If you cannot answer that question with a specific list of accountabilities that require human judgment, the organization will drift back into filling the reclaimed time with more coordination overhead, just with a different texture.
SHRM's 2026 State of AI in HR research found that AI is five-point-seven times more likely to shift job responsibilities and three times more likely to create new roles than to displace jobs. The CHRO's job is to name those shifts explicitly, not leave them implicit and hope the organization figures it out.
The third phase is performance against the new standard. Once the role has been redefined, the scorecard has to reflect it. If managers are now measured on the same metrics they were measured on before offloading, the redesign has not landed. The metrics need to shift toward the things that now require more of their time: coaching quality, team development, judgment calls on edge cases, escalation quality, and the human-in-the-loop decisions that sit on top of the agent layer.
HBR Analytic Services found that only six percent of organizations fully trust agents with core processes. That is probably about right for now. The manager role in the agent era is to be the governing intelligence above the operational layer, not a participant in the operational layer.
What fails when the redesign does not happen
The failure mode is not what people expect. It is not that agents replace managers. It is that organizations add agent oversight to an already-overloaded manager role, the new oversight burden matches the coordination overhead that was offloaded, and the manager ends up doing the same volume of work in a different shape.
Korn Ferry's Workforce 2025 research found that seventy percent of senior leaders say their organization has an AI strategy, but only thirty-nine percent of employees agree. That gap is what it feels like from inside an organization where the redesign did not happen. Leadership sees the agents. The people doing the work see more things to track.
The CHRO closes that gap by making the redesign concrete. Not a vision statement about the future of work. A specific answer to the question: what does the manager role look like now, what does it measure, and what does success in that role require that it did not require before?
The agents carry the operational work so people are free for the work that matters. That sentence only means something if the organization has defined what the work that matters actually is. That definition is the CHRO's job. No agent can do it.
See the live chart
The seats at Sneeze It, both human and agent, are queryable via OTP. You can ask which seats are agent-owned versus human-owned, who the named human owner is for each agent seat, and what metrics each seat is measured against.
In Claude Desktop or Cursor or any MCP client, add this block:
"otp": {
"command": "npx",
"args": ["-y", "@orgtp/mcp-server"]
}
Restart the client. Then ask: "Use OTP to show me the Sneeze It org chart and identify which agent seats have a named human owner."
The response shows what accountability architecture looks like in practice, not in theory.
Series: AI-Era CHRO. Part 48. Previous posts in this series cover what HR does when half the workforce is agents, how the org chart changes, and why the CHRO becomes the owner of the agent workforce.