We spoke with Rob Giglio (Chief Customer Officer), Jess Chiew (GTM Strategy & Operations), and Michael Denari (Global Head of IT) about how AI and agents are being used across Canva’s go-to-market teams and internal systems.

Rather than treating AI as a separate initiative, their focus has been on redesigning how work gets done — with clear principles, close coordination between GTM and IT, and an emphasis on adoption over experimentation.

1. Start with a simple philosophy

The approach to AI flows directly from Canva’s mission to make complex things simple. Three ideas consistently guide decisions.

• AI is first used to remove friction, collapsing multi-step tasks so people can focus on higher-value work.
  
  
• Use cases are expected to tie clearly to business or customer value, not novelty.
  
  
• AI is positioned as supporting people — freeing time for customer conversations, coaching, and creative work rather than replacing them.
  
  

These principles come first. Tools follow.

2. Map the journey before introducing AI

From a GTM perspective, Rob’s team hasn’t changed what it is responsible for — it has changed how the work happens.

They start with a full view of the customer lifecycle, spanning awareness, engagement, evaluation and decision, deployment, and long-term success. From there, the team breaks down the jobs each role performs at every stage across marketing, sales, customer success, and support.

Only once this work is clear do AI opportunities get introduced — typically where tasks are manual, time-sensitive, or difficult to personalise at scale.

3. Business problems drive the roadmap

Jess is clear that starting with tools leads to the wrong outcomes.

The GTM AI roadmap was shaped using two complementary lenses:

From the business side, the focus was on performance gaps, the input metrics that actually drive revenue and customer value, and where cycles were too slow or overly manual.

From the frontline, time was spent with AEs, BDRs, and CSMs to understand what felt repetitive or low leverage, and where additional support would make a difference.

AI use cases were only prioritised where these views aligned, and each one needed a clear, near-term payoff — whether that was time saved, faster deal cycles, quicker onboarding, or better prioritisation.

4. Run quick wins and bigger bets in parallel

The team deliberately runs two tracks at the same time.

Quick wins focus on lightweight automation:

• Call preparation
  
  
• Research summaries
  
  
• Follow-up drafts
  
  

These save small amounts of time each day and help prove value quickly.

In parallel, the team invests in more substantial changes, such as rethinking enrichment and CRM intelligence around how their GTM model actually works, and building more nuanced account scoring. These efforts are treated like product launches, with proper planning and change management.

Quick wins build momentum; bigger bets create lasting advantage.

5. Fewer tools, deeper automation

Jess points out that reps already operate across many tools, and adding another interface often creates friction rather than value.

Instead of introducing more point solutions, the focus has been on embedding agents inside existing workflows. Agents surface in Slack for collaboration and live directly in Salesforce for account and opportunity work. Complexity is handled behind the scenes, while the experience for reps stays simple.

6. A shared GTM–IT operating model

None of this works without IT involvement. Denari's team owns the non-product tech stack, internal AI and automation platforms, and governance.

Because GTM and IT already co-owned systems like Salesforce, Netsuite, and Customer Billing, there was a strong foundation of trust. That made it easier to extend into agents and AI.

Security and governance were built in from the start, including:

• Role-based access control
  
  
• Controlled access to integrations and API keys
  
  
• Monitoring, observability, and audit logs
  
  
• Clear decisions about where data can be used
  
  

Business teams build on governed platforms; IT sets the guardrails.

7. Treat AI like product

AI is not treated as a side project. It is run like a product capability, with ownership, cadence, and accountability.

Company-wide enablement included initiatives like AI Week, which combined training with a hackathon. At the team level, groups run ongoing sessions that fit their work, such as IT’s weekly “Hour of Power.”

Clear ownership matters. The creation of a dedicated GTM AI Lead role helped move initiatives from experimentation into durable workflows.

8. Be clear about what stays human

Across Rob, Jess, and Denari, there is strong alignment on boundaries.

AI is used for preparation, summarisation, enrichment, scoring, and repetitive outreach. The moments that matter most — coaching, high-trust customer conversations, negotiation, and post-purchase value creation — remain human-led.

AI compresses work around these moments rather than replacing them.

9. Production requires a different mindset

From an IT perspective, prototypes are easy; production is harder.

The team starts in sandbox environments, uses human-in-the-loop review, and expects multiple iterations to tune prompts, context, and orchestration. Workflows only move into production once governance, monitoring, and user experience are in place, at which point ownership shifts to the business team.

Speed matters, but not at the cost of security or trust.

10. What others can take from this

At its core, the approach comes down to a few repeatable moves:

• Anchor AI to mission and customer value
  
  
• Map journeys and jobs before choosing tools
  
  
• Balance quick wins with deeper changes
  
  
• Embed agents into existing workflows
  
  
• Formalise GTM–IT partnership with clear governance
  
  
• Protect the human core of customer-facing work
  
  

This is how AI becomes part of day-to-day operations rather than a separate experiment.

‍