Canva announced last week that it’s cutting or redeploying about 22% of its people, roughly 1,271 roles, in what Melanie Perkins’ memo called its hardest year yet (The Next Web has a good summary). Layoffs are never something to be glib about, and to Canva’s credit the package sounds decent: six months’ pay, people keep their laptops, and there are grants for those who want to start their own ventures. That’s more generous than most.

But it’s the other half of the announcement that I think engineering leaders should read twice. Canva is merging its Product, Growth and Platform teams into a single Product organisation, reducing narrow specialisations and removing management layers. To me that reads as an org design statement far more than a cost-cutting footnote, and I suspect we’ll see a lot more of it.

The Specialist Org Chart Was Built for a Different Era

Here’s the thing: the classic scale-up structure (a platform team, a growth team, several product teams, each with its own managers, roadmaps and rituals) made perfect sense when the expensive part of software was writing the code. You split the work so people could go deep, then you added managers and meetings to stitch the pieces back together. The stitching was the price of specialisation, and everybody paid it without much complaint.

AI changes that maths. When an engineer with a decent agent can work across the frontend, the backend and the data pipeline in the same afternoon, the walls between teams stop protecting depth and start creating queues. The ticket goes from Product to Platform, waits for a sprint, comes back with questions, goes to Growth for the experiment framework… you get the picture. Every hand-off is a place where work sits still. As such, a structure with fewer seams moves faster, almost regardless of how talented the people inside it are.

I see this a lot with the companies I advise: the teams moving fastest with AI aren’t the ones with the best tools, they’re the ones with the fewest hand-offs. A small group that owns a problem end to end, with an agent or three doing the heavy lifting, will run rings around three specialist teams coordinating through a programme manager.

Fewer Layers, More Decisions

The removal of management layers is the bit that makes people nervous, and I understand why. But I’d argue it’s the logical consequence. A lot of middle management in tech exists to move information around: status up, priorities down, dependencies sideways. Agents and decent tooling now do much of that moving for free. What doesn’t go away is the need for judgement, and the managers who stay need to be the ones who make calls rather than schedule them.

I believe AI enables velocity, and companies that still decide things by committee, sitting on a decision for days while everyone gets their say, will be swallowed by smaller, agentic competitors. Canva is hardly small, but this restructure reads like an attempt to behave like a smaller company before a smaller company eats its lunch. (Somebody at Canva has probably already designed the new org chart in Canva. I just hope it fits on one slide now 🙂)

What I’d Take From It

Sure, you can argue this is just a company with slowing growth dressing up cost cuts in strategy language, and some of that may be true. But even if the motive is money, the shape is worth copying deliberately, rather than waiting for a bad year to force it on you. If I were looking at my own org this week, I’d ask three questions:

• How many teams does a typical feature touch before it ships, and how long does it wait at each border?
• Which of my specialisations exist because the work needs them, and which exist because the org chart grew that way?
• Which managers spend most of their week deciding, and which spend it relaying?

None of those answers require a single redundancy. They do require honesty, which is often harder to find.

Bottom line: the real work in restructuring around AI is removing the seams that AI exposed, and you’d much rather redraw them yourself than have the market do it for you.