1300 166 138 Tell us what you need
News / Canva

Canva paused work for a week to teach 5,300 staff AI

By Monika Tantau • Published 6 months ago
Canva runs an AI Discovery Week for 5,300 staff at its Sydney headquarters

Key takeaways

• Canva halted normal operations for five days across its 5,300-strong workforce for an AI Discovery Week.
• More than 60 talks and workshops featured speakers from OpenAI, Anthropic and Google, with optional attendance.
• Internal automation has redeployed routine roles like the Vibe and Thrive HR benefit, freeing staff for higher-value work.
• The approach contrasts with recent layoff-led pivots at firms such as Cloudflare.

Most Read

Canva paused normal operations across its global workforce for five days last week, dedicating its 5,300-strong team to an AI Discovery Week of talks, hackathons and hands-on workshops. The company brought in speakers from OpenAI, Anthropic and Google, ran more than 60 sessions, and reportedly filled its Sydney office to the point staff struggled to find desks, as reported by The Age.

It is a noticeable contrast to the playbook some of Canva’s overseas peers have favoured. Cloudflare confirmed this week it was cutting 1,100 roles to refocus on agentic AI; Canva’s Australian-headquartered approach has been to keep the headcount and redirect a week of it into learning.

What Canva actually did

Attendance was optional, but uptake appears to have been strong. Staff travelled into Canva’s Sydney HQ in numbers large enough to overwhelm desk capacity, by the company’s own internal accounts. The 60+ sessions covered everything from prompt engineering to building custom AI assistants, with engineers, finance staff, operations teams and even Canva’s chefs participating side by side.

The week culminated in a company-wide hackathon framed around supercharging existing tools and inventing new ones. Internal AI use was already part of Canva’s operations: automation has handled a chunk of routine HR ticketing for the company’s Vibe and Thrive benefit programme, freeing the team to work on higher-impact projects.

Why pause instead of cut

The framing matters. Canva chose to slow the business deliberately, treating AI literacy as a capability investment, comparable to a finance module rollout or a security audit. That sits in sharp contrast with companies that have absorbed the AI shift via redundancy: Cloudflare’s recent restructure trimmed 1,100 roles in the same week, with management citing AI productivity as the rationale.

Canva executives have pointed to staff demand for unstructured time to experiment as a key driver. The bet is that broad, hands-on exposure across teams, rather than concentrating AI capability in a small cluster of specialists, produces faster integration into daily work.

Canva's About Me feature began as an internal hackathon project, illustrating how AI Discovery Week ideas reach product

From hackathon to product

Canva already has a track record of pulling hackathon outputs into its commercial product. The About Me personalisation feature, now visible across the platform, started life as an internal hack project. That precedent makes the company-wide event more than goodwill optics: it doubles as a product-discovery pipeline, low cost relative to formal R&D and with a wide enough net to surface ideas the senior team would not otherwise see. The question for any organisation copying the format is whether it has the operational maturity to ship the best ideas afterwards.

What this means for Australian businesses

Most Australian operators do not have the cash or scale to pause a 5,300-person team for a working week. But the underlying principle of protected time for AI experimentation, structured around real business problems, translates down. Mid-market firms can run two-day workshops, walk finance and operations leads through hands-on sessions with Claude or ChatGPT, and capture the experiments their staff are already doing privately.

For sectors where data sensitivity is the blocker, the format also flushes out where on-shore deployment or a controlled vendor list is necessary before scaling tools. That is closer to the work an AI discovery workshop typically covers: a first audit of where AI fits, where it does not, and what governance has to come first.

A different bet on the AI transition

The wider tech sector is sorting itself into camps. Some firms are treating AI primarily as a way to reduce headcount; Canva is treating it as a capability the existing workforce should grow into. Both bets carry risk. The cuts approach is fast but loses institutional memory.

The upskill approach is slower and more expensive, but keeps the people who already understand the business in the room when the tools are introduced. For Australian companies still drafting their AI plans, the more useful question is probably which of those trade-offs they can live with.

Scroll to Top