Week 31: AI and the Death of the Reporting Manager
AI and the Death of the Reporting Manager
I need to say something that a lot of people in middle management don't want to hear.
If your primary contribution to your organisation is aggregating data, compiling reports, and presenting information upward — your role is being eliminated. Not threatened. Not disrupted. Eliminated.
AI does what you do. Faster, cheaper, and more accurately.
This isn't a future prediction. This is happening now. Every organisation I talk to is having some version of this conversation. And the people most at risk are the ones who've built entire careers on being the information layer between the data and the decision-maker.
I've spent twenty years in continuous improvement — thirteen of those at Shell, plus leadership roles at Johnson Controls and KCA Deutag. I've watched technology reshape operations teams before. ERP systems. Business intelligence tools. Automation. Each wave displaced people whose primary function was data handling.
AI is the final wave. And it's bigger than the others combined.
Let me tell you about two managers at the same company. I won't name the organisation, but I'll tell you the story precisely because it illustrates the difference between what survives and what doesn't.
Both were operations managers. Same level. Same company. Same type of operation. Similar team sizes. Both were considered strong performers.
Manager A was a master of the report. His weekly operations review was a work of art. Fourteen slides. Every KPI tracked, trended, and colour-coded. Variance analysis. Pareto charts. Commentary on every deviation. He spent roughly eight hours per week — a full working day — preparing this review. His director relied on it. It was the primary mechanism through which performance information flowed from the operations floor to the leadership level.
Manager A was valuable because of what he knew and how he presented it. He was the person who could answer any question about operational performance because he'd spent hours immersed in the data, synthesising it, making it digestible.
Manager B was different. Her weekly review was simple. Six slides. Key metrics only. No elaborate commentary. She spent about ninety minutes preparing it. Her director sometimes wished it was more detailed.
But Manager B spent the time she didn't spend on slides doing something else. She was on the floor. Every day. Walking the lines. Talking to team leads. Observing changeovers. Sitting in on problem-solving sessions. Coaching her supervisors on how to run their daily stand-ups. Following up on improvement actions.
When a performance issue appeared in her data, she usually already knew about it — because she'd seen it on the floor before it showed up in the dashboard. When her director asked why Line 3 had dropped, she didn't just have the data. She had the story. She'd been there. She knew what had happened, who was working on it, and what the plan was.
Manager A aggregated information. Manager B created understanding.
Then AI arrived. Not as a dramatic overnight shift, but as a gradual integration of automated reporting, AI-generated analysis, and intelligent dashboards.
The company implemented an AI-powered operations dashboard. It pulled data from every production system in real time. It generated automated variance analysis. It produced Pareto charts, trend analysis, and exception reports. It identified anomalies before humans noticed them. It generated commentary on deviations using pattern recognition against historical data.
The dashboard produced, in real time, something very close to Manager A's fourteen-slide review. Not identical. But close enough that the director no longer needed a human to aggregate and present the information.
Manager A's eight hours of weekly preparation became unnecessary. His primary contribution — the synthesis and presentation of operational data — had been automated.
He still had things to do, of course. He still had a team. He still had operational responsibilities. But the thing that had made him valuable — the thing that had earned him his reputation, his seat at the table, his weekly audience with the director — was gone.
Manager B barely noticed the change. The AI dashboard gave her better data, faster. She used it. It saved her thirty minutes of the ninety she'd previously spent on her review. She redirected that thirty minutes to the floor. Her contribution — coaching, problem-solving, leading improvement, creating understanding — was unchanged. If anything, the better data made her more effective.
Six months after the AI implementation, the company restructured. Manager A's role was redesigned. His direct reports were redistributed. He was offered a lateral move into a process improvement role — which, to be frank, he wasn't equipped for, because he'd spent his career managing information rather than managing improvement.
Manager B was promoted. She was given a larger scope, more teams, more responsibility. Because the thing she did — lead, coach, enable, improve — was exactly what the organisation needed more of.
I've been watching this pattern develop for years, and AI has accelerated it dramatically.
At Shell, when we implemented real-time dashboards on the shopfloor, the supervisors who thrived were the ones who used the data as a starting point for conversation, not as the conversation itself. They'd look at the board, see a deviation, walk to the process, and have a conversation with the operator about what was happening. The data was the trigger. The leadership was the response.
The supervisors who struggled were the ones who had been managing through information control. They knew things. They reported things. They were the conduit. When the data became transparent, their conduit role evaporated, and what remained was — in some cases — an empty space where leadership should have been.
AI does this at a larger scale and a higher level. It doesn't just make shopfloor data transparent. It makes every layer of information aggregation transparent. The monthly report that took three days to compile. The quarterly business review that required a week of preparation. The market analysis that occupied an entire team. All of it is being automated.
What's left after the automation is the human work. The judgment. The coaching. The relationship-building. The difficult conversation. The Gemba walk. The problem-solving session. The decision that requires courage, not data.
Here's the uncomfortable truth for middle management.
A significant percentage of middle management activity — and I'm drawing on observations from more than forty organisations when I say this — is information logistics. Moving data from one level of the organisation to another. Reformatting it. Adding context. Presenting it.
AI handles information logistics. Completely.
The managers who survive the transition will be the ones whose contribution extends beyond information handling. The ones who coach their teams. The ones who go to the floor. The ones who facilitate problem-solving. The ones who make decisions with courage and conviction. The ones who build capability in their people.
The managers who don't survive will be the ones whose primary skill was knowing things that other people didn't have access to. Because AI gives everyone access.
I'm not saying this with satisfaction. Some of the most dedicated, hardest-working managers I've met over twenty years were information managers. They worked long hours. They took pride in their analysis. They provided genuine value to their organisations by being the human bridge between raw data and executive understanding.
But the bridge has been automated. And pretending it hasn't is a disservice to the people who need to hear this.
The question for every manager is not "will AI affect my role?" — it will. The question is "what do I do that AI cannot?"
If the answer involves aggregating, compiling, analysing, presenting, or summarising information — that's the AI's territory now.
If the answer involves coaching, deciding, leading, enabling, developing people, going to the Gemba, having difficult conversations, and building a team that can solve problems — that's human territory. And it's more valuable than ever.
At Johnson Controls, I worked with managers across the EMEA region who had very different leadership styles. The ones who were most future-proof were always the ones who spent more time on the floor than in front of their screens. They weren't better with data. They were better with people. And they used data as a tool, not as an identity.
At KCA Deutag, in the demanding environment of drilling operations, the distinction was even sharper. The best operational leaders were the ones who could walk a rig, read the situation, make a call, and keep their crew safe and productive. Their value was judgment under pressure. Data supported them. It didn't define them.
The enabling manager — the manager who develops capability, coaches performance, and creates the conditions for improvement — is the future. Not because it's a nice leadership philosophy. Because it's the only management model that can't be automated.
If your job disappears when AI generates the reports, it means you were the reports. And that was never management. That was data processing with a management title.
The managers who lean into the human work — who invest in their coaching skills, their Gemba practice, their facilitation capability, their ability to have honest conversations and develop people — will thrive. They'll have better data than ever, delivered faster than ever, and they'll use it to lead more effectively than ever.
The managers who cling to the information-broker role will find that the broker is no longer needed when the data speaks for itself.
So here's my question.
If an AI dashboard could generate your weekly report tonight, what would you do with your time tomorrow?
If the answer is "go to the floor, coach my team, and solve problems" — you're ready.
If the answer is "I'm not sure" — the clock is ticking. And the best time to figure it out is now.
If you want to develop the performance management and leadership capabilities that AI can't replicate — the coaching, the cadence, the Gemba discipline, the people development — explore the Performance and Process Management Mastery programme at stormholt.org/products/performance-and-process-management-mastery.
It's built for the manager who wants to lead, not just report.
The Stormholt operator stack — AI-native.
If your primary contribution is no longer report compilation, the Stormholt AI tools take the lower layer off your plate so you can focus on the higher one. Free, all of them.
- Custom GPT — Stormholt CI Coach in ChatGPT. Walks one real problem end to end. Install
- Claude Skill — same coach inside Claude.ai. Install
- MCP Server — eight CI tools (start_a3, walk_vsm, audit_kpi_tree, design_cadence, etc.) inside any MCP-compatible editor. Install
- AI Workshop Generator — €49, full CI workshop in under 90 seconds. Try it
If you want one operator-grade Stormholt note a week to keep the discipline up, join the Operator Network: stormholt.org/op
A number you don't react to is not a KPI — it's wallpaper. — Stormholt