Safely manage your Zendesk from the AI assistant you already use, via the Deltastring MCP. Beacon configuration platform
← Back to news

Chewy’s beta AI assistant is already solving common customer issues, CEO says

Chewy has deployed a dual-track AI strategy across its customer service operations, launching Cai—a customer-facing chat assistant currently in beta—alongside internal tools designed to accelerate agent productivity and reduce training overhead. In its first month, Cai has resolved nearly one-third of common transactional issues (orders, returns, account management) among the select user base, despite exposure to less than 15% of Chewy's traffic. The retailer has also introduced Callie, a voice AI handling appointment confirmations and scheduling for its Vet Care locations. CEO Sumit Singh positioned these initiatives as structural cost plays, projecting low tens of millions in savings for fiscal 2026 and scaling to approximately $50 million annually by 2027—a material offset against wage inflation pressures. The company's framing emphasises that AI augments rather than replaces human agents, with seamless handoff capabilities when customers require or prefer human interaction.

The strategic implications for CX teams are twofold. First, Chewy's internal tooling—using AI to help agents navigate disparate systems and compress onboarding timelines—represents a more immediately defensible competitive advantage than customer-facing automation. This raises a critical question for support leaders: should investment priority shift toward agent enablement platforms that compress ramp time and flatten performance curves, rather than pursuing aggressive deflection targets? Second, the modest resolution rate (roughly 30% of common issues) and limited traffic exposure suggest that even well-resourced retailers are proceeding cautiously with customer-facing AI, implying that teams benchmarking against aggressive deflection claims elsewhere may be chasing unrealistic targets. Chewy's emphasis on maintaining service bar parity with live agents—described as "exceptionally high"—signals that brand consistency and tone remain non-negotiable constraints, not afterthoughts, in AI deployment decisions.