Canva's customer success operation has inverted the reactive support model by deploying predictive AI interventions that identify and resolve issues before customers escalate them, achieving net retention above 140% in the process. Rather than optimising for deflection—the industry standard of routing customers toward self-service to reduce support costs—Rob Giglio, Canva's customer chief, has positioned resolution as the primary metric. This philosophical shift means support teams are equipped to solve problems directly rather than funnel users toward knowledge bases or chatbots, fundamentally reframing what "efficiency" means in customer success. The implication is stark: teams currently measuring success by ticket deflection rates or first-contact resolution through automation may be optimising for the wrong outcome. If Canva's net retention figures are genuine, the business case for proactive, human-centred resolution over deflection becomes difficult to ignore.
The tension this creates across the CX stack is significant. Most platforms—Zendesk, Freshdesk, Salesforce Service Cloud—have been architected around efficiency metrics that reward automation and self-service adoption. Canva's playbook suggests these metrics may be inversely correlated with retention, particularly for products where customer success directly influences expansion revenue. The question becomes whether this model scales beyond high-growth SaaS companies with mature customer bases, or whether it requires the operational maturity and margin profile that Canva possesses. For support leaders at mid-market organisations already stretched thin, the philosophical appeal of "resolve, don't deflect" may collide with resource constraints that make predictive intervention feel like a luxury rather than a standard operating procedure.
What distinguishes Canva's approach is the timing and intent of intervention. Predictive systems that surface issues before customers notice them represent a fundamentally different use of AI than chatbots trained to handle common queries. This requires investment in data infrastructure, customer health scoring, and proactive outreach workflows—capabilities that sit upstream of traditional support platforms. For teams evaluating their AI roadmap, the question is whether to continue layering automation onto existing reactive workflows, or to redirect investment toward the predictive layer that Canva has clearly prioritised. The net retention metric suggests the latter approach has material business impact, even if it demands a different operational model than most CX teams currently run.
Zigging while others zag: Canva flips AI service playbook as predictive, proactive intervention push net retention above 140%; Resolve, don't deflect, says customer chief Rob Giglio Mi-3.com.au