Recruiting · Production

Canva redesigns technical interviews to require AI tool use by engineering candidates

The problem

Canva's technical interviews asked candidates to solve coding problems without AI tools, failing to reflect real-world engineering work where AI is used daily; traditional CS Fundamentals questions were trivially solvable by AI, giving no meaningful signal about how candidates would perform on the job.

First attempt

The Computer Science Fundamentals interview format pre-dated the rise of AI tools; when Canva tested those questions with AI, the tools produced correct, well-documented solutions in seconds without any follow-up prompts, rendering the format an ineffective evaluation signal.

Workflow diagram · grounded in source
1
Candidate informed to use AI
trigger
“we now inform candidates ahead of time that they'll be expected to use AI tools”
2
Complex challenge presented
trigger
“These problems can't be solved with a single prompt; they require iterative thinking, requirement clarification, and good decision-making”
3
Candidate uses AI tools
ai_action
“candidates are expected to use their preferred AI tools to solve realistic product challenges”
4
Candidate validates AI output
validation
“Critically reviewed and improved AI-generated code”
5
Interviewer evaluates AI collaboration
human_review
“We're evaluating candidates on skills that matter more than ever: Do they understand when and how to leverage AI effectively? How well do they break down complex, ambiguous requirements? Can they make sound technical decisions while usin…”
6
Predictive performance signal produced
output
“they're providing strong predictive signals about candidate performance”
Reported outcome

The piloted AI-Assisted Coding interview replaced the CS Fundamentals screen and delivered more engaging interviews with stronger predictive signals about candidate performance.

Reported metrics
engineers as daily AI coding tool usersalmost half
Interview engagement qualitymore engaging for both candidates and interviewers
Reported stack
CopilotCursorClaude
Source
https://www.canva.dev/blog/engineering/yes-you-can-use-ai-in-our-interviews/
Read source ↗

Frequently asked questions

What did this team achieve with this AI workflow?

The piloted AI-Assisted Coding interview replaced the CS Fundamentals screen and delivered more engaging interviews with stronger predictive signals about candidate performance.

What tools did this team use?

Copilot, Cursor, Claude.

What results were reported?

engineers as daily AI coding tool users: almost half; Interview engagement quality: more engaging for both candidates and interviewers (source-reported, not independently verified).

What failed first in this deployment?

The Computer Science Fundamentals interview format pre-dated the rise of AI tools; when Canva tested those questions with AI, the tools produced correct, well-documented solutions in seconds without any follow-up prom…

How is this recruiting AI workflow structured?

Candidate informed to use AI → Complex challenge presented → Candidate uses AI tools → Candidate validates AI output → Interviewer evaluates AI collaboration → Predictive performance signal produced.

Related recruiting cases