Canva Lets Staff Choose AI Tools Instead of Mandating Copilot

Canva is telling staff not to wait for a corporate decree on which AI assistant to use. Co-founder and chief product officer Cameron Adams said employees can choose the tools that fit their work rather than being required to use a single product such as ChatGPT, Claude, or Gemini, according to Business Insider’s interview coverage from July 15.
Adams’ argument is simple: mandating one tool encourages reluctant compliance rather than experimentation. Canva instead gives employees budgets to test services and build workflows around the problems they actually need to solve.

A managed freedom problem​

That approach will sound familiar to Windows administrators who have spent the past two years dealing with shadow AI. Users will use AI tools whether IT has standardized on one or not; the difference is whether their use happens within a governed framework.
Canva’s model is not a free-for-all. The company has described a core set of supported AI services, including enterprise offerings from Anthropic, Google, and OpenAI, while giving staff dedicated time to explore practical uses. Its second AI Discovery Week paused ordinary work across the company so employees could attend sessions, test tools, and participate in a hackathon. Canva says more than 5,300 employees took part.
For enterprise IT, the useful distinction is between prescribing a single interface and setting guardrails. A company may permit multiple approved models while still controlling identity, data classification, retention, logging, procurement, and access to sensitive connectors.

What Windows shops should take from it​

For organizations already deploying Microsoft 365 Copilot, the lesson is not that Copilot should be abandoned in favor of a sprawling collection of consumer AI accounts. It is that adoption metrics alone do not prove usefulness, and a single approved assistant may not fit every workload.
A sensible policy can allow teams to evaluate alternatives while keeping the controls that matter:

• Require enterprise accounts and prohibit sensitive data in personal AI services.
• Define approved use cases, data boundaries, and review requirements.
• Fund limited, time-boxed pilots rather than unmanaged subscriptions.
• Measure task outcomes and quality, not merely prompt counts or token use.
• Keep Microsoft 365, endpoint, identity, and data-loss-prevention controls in the loop.

Canva’s experiment also highlights a less technical barrier: staff need protected time to learn where AI is useful. The company’s own 2026 account says it stopped normal work for a week rather than asking employees to learn new tools between meetings and deadlines.
The practical next step for IT leaders is to pair any preferred AI platform with a small, governed evaluation program for real team workflows.

---

References​

1. Primary source: Business Insider
   Published: 2026-07-15T04:03:13.154000+00:00
   
   

   Canva Cofounder on Why Company Won't Force Staff to Use One AI Tool - Business Insider

   Employees will use the tool "begrudgingly," Canva cofounder Cameron Adams said, which in turn hinders experimentation.

   www.businessinsider.com
2. Independent coverage: Business Insider Africa
   Published: 2026-07-15T04:03:13+00:00
   
   

   Canva's cofounder explains why the company isn't forcing employees to use a specific AI tool

   Canva cofounder Cameron Adams said theres one way to get your employees to stop caring about AI: forcing them to use a specific AI tool

   africa.businessinsider.com
3. Related coverage: hrgrapevine.com
   
   

   Canva gave 5,000 employees a week off for AI training. What did they learn? | Learning | HR Grapevine | Feature

   Following Canva’s inaugural AI Discovery Week, HR Grapevine reveals the impact of giving a global workforce five unfiltered days of AI training...

   www.hrgrapevine.com