Editorial
White Paper
The Future of Leadership Development: AI-Driven Manager Enablement
“You should use AI more.” “We need more AI to improve our work.” The pressure is everywhere: from workplaces, leaders and vendors. The message I took from Satya Nadella at the recent Microsoft AI Tour was hard to miss: use Copilot, or risk being left behind. He conveniently forgot to mention the upcoming pricing changes.
There’s logic to this. The technology has moved so fast in the last couple of years, moving from assistance to doing. We hear stories of AI transforming some processes, such as how Ikea’s use of customer agents changed the role of contact center staff — rather than replacing them — from handling queries to being design consultants and increasing revenue as a result. However, stories like this appear to be the exception. Where we’re not leading with AI, it starts to feel like AI is being done to us. And that’s where the problems begin.
When AI Starts as the Answer
The pace of change is impossible to keep up with. When combined with the pressure to use AI, it can feel as though we are always reacting to events outside our control. We jump onto things because we feel we should, without a clear view of what we’re trying to achieve.
A different mindset is required. Conor Grennan, author and former Chief AI Architect at NYU Stern School of Business, recently highlighted a key issue: AI isn’t replacing a tool in the way traditional technology rollouts do. It is replacing a way of thinking.
Prompt libraries are still a tool-first approach. They can add to the pressure of AI being done to us. We must step away from thinking about AI as another tool to adopt and, in some cases, from talking about AI itself. We need to talk about what we do. About work.
Control comes through leading, not following. We lead by getting clear on what we’re doing, where we’re going and why. That clarity often comes from tough conversations: sense-making about what is really happening, understanding the needs, ideas and fears of our colleagues, and surfacing the things we do not always see but can use to give our work purpose.
The pressure to use AI, though, is typical of technology adoption without a clear need. We put the solution ahead of the problem, like the days when a new intranet platform was rolled out because the head of IT enjoyed the vendor’s client management skills. We then became stuck with Yet Another Tool: one without an obvious purpose, or one launched alongside existing tools that already did much the same job.
Prompting in Pursuit of Purpose
Prompts of the day can help people get started, but they won’t transform anything on their own. They also promote a largely individual use of AI: employees prompting AI in one-off use cases that can be time-consuming and relatively low reward. We’re using AI because it’s there, not because we’ve found a clear use for it.
This means we spend time looking for ways to use AI, trying a prompt, trying again, sometimes giving up, sometimes persisting. Adopting AI in this way is hit and miss. Even when it works, the time saved can be marginal once we count how long it took to get the output we wanted.
An AI-first approach is leading us into what the Work AI Institute has termed “botsitting”: the time-consuming work it takes to make AI usable, which turns out is more time than we spend using AI to produce the work in the first place. Botsitting, however is preferable to what happens if we don’t bother with the clean up — where it turns into “botshitting.” The shipping of AI-generated work that hasn’t been verified, isn’t fully understood or authors can’t stand behind. AI is now being done to us more explicitly, because we have to deal with the consequences of someone else’s lack of judgement.
When the mental load of monitoring AI outweighs its productivity benefits, it leads to what Harvard Business Review and Boston Consulting Group have called “AI brain fry.” The mental impact caused by heavily overseeing multiple AI tools. Instead of managing the task, we’re managing the outputs of agents, another form of botsitting.
Work-First, AI Later
We have our priorities the wrong way around. We need to start with work, not the technology. We can fall into the trap of creating a grand AI strategy deck full of beautiful dashboards (AI generated, naturally). However, AI should support business strategies — not the other way around. In day-to-day work, teams can take some simple steps to bring more clarity to how they use AI. And, just as importantly, when they choose not to.
Agree on Simple Guardrails
As a team, agree which uses of AI require review or approval, and what those steps look like. For example, what sort of AI use needs sign-off, what can be handled informally and who should be involved? This helps clarify how much oversight is required before work is shared more widely.
Don’t Talk About AI. Talk About Work
Prompts, use cases and scenarios are helpful for understanding what AI can do. But we need to stop starting with AI and start with the work. Back when meetings were convened to solve a problem, rather than to fill our calendars with important updates, we looked to the participants to help figure out the way forward. We should treat AI the same way: as something we bring into the work only when it helps us think, explore or decide.
For example, instead of asking “how can we use AI in customer reporting?” ask “why does customer reporting take so long, where do errors creep in, and which parts require judgement?” AI may still help, but now it is serving the work rather than driving it.
Make it Social
The use of chat interfaces has led to a very individual relationship with AI. We typically work alone, treating AI as a personal resource, advisor or tool. Working on team workflows, or bringing others into our conversations with AI, shifts it from a personal productivity hack to a shared way of improving how work gets done. The AI tools themselves don’t help: the starting point is always one-to-one. Bringing others into an AI chat is usually possible, but not obvious. Otherwise we end up copying and pasting outputs — and who’d have thought that good old CTRL+V would be the future of work, still?
The Human Role
“Human in the loop” implies that AI is in charge, with people somewhere lower down the pecking order. Instead, roles should make human leadership of AI outcomes clear. Who is responsible for approving higher-risk AI activities? Who leads team AI brainstorming? And, crucially, when do we decide not to use AI? We need to be clear that we do not have to force AI into everything we do. The weight of expectations can lead to overuse rather than thoughtful use.
Agree on Some Norms
Without shared norms, AI use can quickly become either too individual or quietly avoided. Teams benefit from simple agreements about what’s OK, what’s not, when AI should or shouldn’t be used, how work is reviewed, and how concerns are raised. The key is not to impose these rules from above, but to co-create them with the team so everyone has a clear, shared understanding of how AI fits into their work.
AI Is Not a Solo Sport
Writing emails, looking for information, summarizing data, producing visuals. The individual focus of AI often leads to lower-value outputs. Low risk, but low reward. Add the time it takes to botsit and the impact of botshitting, and sometimes it is barely worth the effort.
Engaged, But Not Convinced: What's Eroding Beneath Your Strongest Engagement Numbers
Watch Now
Retention Is a Communication Problem: New Data from 6,000+ Employees
Watch Now
But when we work on shared outcomes, such as team workflows, the value increases. That value comes from clarity about the problem, the need or the opportunity. It is where we take ownership by shifting the narrative from an AI thing to a work thing. We combine our collective experience and skills to develop a clear reason and better outcomes. There is shared judgement, application, knowledge and, crucially, human-to-human collaboration driving everything. The real opportunity is not getting everyone to use AI more. It is helping teams understand their work better, make smarter choices together and use AI only where it genuinely helps. That is how we stop AI being done to us and start making it work for us.
Editor's Note: How else are we rethinking how we work with AI?
- The Human Advantage: MAPI - What AI Disrupts and How to Restore It — Why the age of AI needs more human judgment, not less. Part three in a four-part series. Today: how AI undermines motivation and how leaders can restore it.
- The Cognitive Economics of AI — AI promised to free up our time, but it's making us busier than ever. Here's why protecting time to think is a leadership imperative.
- The Teammate With No Manager: Who's Accountable for AI? — Why cross-functional leadership, accountability and employee experience design are central to AI’s success at work.
Learn how you can join our contributor community.