AIs Collaboration Problem

AI has made work dramatically faster. What it hasn’t done is make teams any better.

Right now, “peak productivity” looks like one person running a wall of prompt-based chats and a handful of agents on their laptop. One person, a fleet of tools, and a constant stream of output. It’s an impressive model of individual efficiency. It’s also built on a flawed assumption: that important work can be done alone.

It can’t. Businesses aren’t built by individuals scaling themselves. They’re built by teams agreeing on what matters, what to prioritise and what to ignore. And this is exactly where AI falls short. The tools we’re using are fundamentally single-player as they increase the volume of output but do nothing to improve shared understanding. In fact, they often make it worse.

More output doesn’t resolve misalignment it amplifies it. Instead of better decisions, you get more work that nobody fully agreed on in the first place. That’s because the real bottleneck has shifted, it’s no longer about getting the work done, it’s about deciding what work is worth doing at all.

When production becomes cheap, choice becomes expensive because every prompt is a decision and every output carries an opportunity cost. Do the wrong thing quickly and all you’ve done is waste time faster.

This isn’t an entirely new problem. Alignment has always been difficult but before AI the pace of work forced collaboration. Teams had to plan, discuss and sense-check along the way. By the time something was delivered, most people had at least seen it and helped shape it.

Now that constraint has disappeared, work can be generated instantly, so planning gets skipped. Context stays trapped in individual heads, agents run in isolation and feedback arrives at the worst possible moment. After the work is already done.

The consequences are predictable. Duplicate efforts, conflicting directions, late-stage rewrites and endless conversations that start with “why are we doing this?” The tools have made production faster, but the organisation itself has slowed down.

If the problem is shared thinking, the solution has to be shared space.

Teams need somewhere to see the same context, ask the same questions and challenge the same assumptions before the work begins, not after it’s been produced. Because remember, the real input to AI isn’t the prompts, it’s context.

And context doesn’t live neatly in a document or a system it lives in people. It’s the accumulated understanding of what’s been tried, what’s worked, what hasn’t, what matters and what the team is actually trying to achieve. If that stays fragmented, AI doesn’t solve the problem, it simply scales the fragmentation.

That’s why we use Miro. Not as a digital whiteboard, but as a shared thinking environment. A place where ideas are visible, decisions are connected and context is captured in a way the whole team can see and build on. In that environment, AI stops being a personal productivity tool and starts becoming something teams can use together.

AI has given teams time back. The question is whether that time gets spent producing more or thinking better. Because quality still takes effort, it still requires focus and it still depends on people aligning around what actually matters.

The mission for Jointly is to create environments where teams think rigorously together about hard problems. To do higher quality work, get aligned faster and do a few exceptional things rather than a thousand average ones.

AI won’t replace teams. But teams that behave like individuals will lose to those that don’t.

Work jointly. Or don’t bother.