Management has always been hard. In distributed organisations, it became harder. Add AI into the mix, and the entire premise of what a manager is for starts to shift in ways that most organisations are only beginning to reckon with.
At Running Remote 2026, leaders from Yelp, Zapier, and Time Doctor sat down for one of the conference’s sharpest conversations: what does the middle manager look like now, and where does this role go next?
Clarity breaks down first
The panel opened with a point that resonates for anyone who has tried to run a distributed team through a period of rapid change. Clarity breaks down first. In office environments, mediocre communication can be papered over. In distributed work, it becomes immediately visible — and the pace of AI-driven change is compressing the already-difficult task of providing strategic direction.
Yelp’s annual engagement survey revealed a significant gap: employees understood their day-to-day tasks well but struggled to connect that work to shifting strategy and the reasoning behind major decisions. Time Doctor, fully remote for 15 years, described rethinking their entire leadership playbook, not because remote work failed them, but because the AI moment requires a different kind of leader entirely.
The goalpost for talent development moves every 90 days now. For managers and L&D teams trying to build frameworks in that environment, it is genuinely difficult.
From executors to system architects
The central shift the panel kept returning to: leaders can no longer operate purely as executors. With AI handling increasing volumes of operational work, the manager’s role is less about doing and more about designing — building the systems, workflows, and conditions through which people and AI work together effectively.
Three possible futures for managers were outlined: managing more people as AI compresses the work of smaller teams; becoming player-coaches with genuine IC accountability; or focusing deeply on talent development. Zapier’s preference was a combination of the second and third — reintroducing real individual contributor work while simultaneously investing more meaningfully in the development of people. Less task management, more architecture. Less oversight, more enablement.
New career paths for a flatter world
One of the most concrete things the panel discussed was how Zapier and Time Doctor are handling the challenge of career growth when vertical advancement opportunities have shrunk. Retention has more than doubled at some companies even as the org chart flattens. That sounds like a success story, and it is — but it comes with a challenge. When 8 in 10 employees say they feel they have development opportunities but don’t see forward progression, development requests are often really promotion requests in disguise.
The response at Zapier was structural and creative. They created the AI Automation Engineer role — their single biggest new role design since their AI Code Red moment. Filled entirely from internal candidates, the role pays well, is viewed as future-facing, and gives people a genuine sense of career movement without requiring a people management track. Three of ten executive roles at Zapier currently involve no people management at all, including senior positions handling M&A strategy.
The deeper principle: organisations can choose not to follow market compensation structures that pay managers more than ICs at the same level. That’s a deliberate design choice, and companies that make it intentionally tend to attract people who want to work that way.
The AI Automation Engineer role requires three capabilities: domain knowledge, the ability to build, and coaching and facilitation skills — so the person teaches rather than becoming an IT bottleneck. It is, in many ways, a template for the kind of hybrid role that AI makes possible.
Co-creating the future
The panel was clear that this moment is unusual in the same way COVID was unusual: there is no playbook. Success comes from working with employees at every level, not delivering change to them. That means proactively asking what they think, understanding how tools are changing their workflows, and building psychological safety for honest answers.
Time Doctor’s People Check initiative — running almost four years — involves HR proactively reaching out to employees with three questions: What’s happening with you? What roadblocks do you have? Where do you want to go from here? The goal is not evaluation. It is genuine connection and early intelligence about what is and isn’t working.
Performance management, one-on-ones, and the treatment of AI agents
For remote cultures where the one-on-one is foundational, the panel was firm: if you are a people leader, you do one-on-ones without exception. Full stop. If a team gets too large for meaningful one-on-ones to happen, culture and work product both break down.
On AI agents, the panel had a practical and grounding take. Agents work for us, not with us. They should not be given human names, they should not receive performance reviews. The humans who design and oversee agents are reviewed on agent output. It is a clean framing that avoids the confusion — and the liability — of treating AI as something other than a tool.
Early career talent entering AI-native environments will have a significant advantage: they will have grown up building with these tools, with no pre-AI habits to unlearn. They will also be able to onboard far more quickly through AI-enabled personalised learning and simulated work environments. The companies thinking about this now are building something that will matter a great deal in the next three to five years.
