Customers are three times more likely to reach for third-party generative AI tools than brand-owned chatbots when facing service issues, according to Gartner research covering over 3,500 B2B and B2C customers. The data reveals a stark divergence: whilst third-party AI tool adoption has doubled in the past year, company-provided chatbot usage has remained flat since 2022. This gap reflects both familiarity and capability. Customers increasingly trust tools like Claude and ChatGPT because they use them daily across multiple contexts—work, personal tasks, research—and have developed confidence in their responses. Brand chatbots, by contrast, remain episodic touchpoints that customers visit infrequently, making them less intuitive when a service problem arises. The implication is uncomfortable: deploying AI into an existing chatbot that customers already avoid will not reverse that avoidance. Simply bolting generative AI onto a platform with poor baseline adoption is a capital allocation problem masquerading as a technology problem.
The strategic failure lies in two dimensions where brand chatbots fundamentally misalign with customer behaviour. First, most brand chatbots answer questions but cannot transact—they respond to queries and then redirect users elsewhere to complete actions, fragmenting the experience. Customers, however, increasingly expect to both ask and act within the same interface; 58% of consumers and nearly three-quarters of B2B customers have used generative AI specifically to complete tasks. Second, the presentation model—the pop-up widget in the bottom right corner—feels dated against the conversational interfaces customers now expect. Leading organizations are reimagining their entire digital experience as a single intelligent front door rather than a standalone chatbot, replacing navigational structures with conversational ones. For teams already invested in Zendesk, Salesforce Service Cloud, or similar platforms, this raises a critical question: are your implementations positioned to handle transactional resolution, or are they still configured as question-answering layers that funnel users back to traditional self-service? The ROI shortfall many teams experience stems not from AI quality but from architectural choices that preserve legacy workflows rather than reimagine them around customer expectations.
In customer service, third-party generative AI tools are beating brand chatbots CX Dive