AI Market Splits by Vertical
Sesame AI
The market is likely to split by job, not collapse into one all purpose companion. A voice bot for customer support needs to answer from a company knowledge base and hand off cleanly to a human, while an education bot is judged on pedagogy and classroom controls, and a health oriented bot faces privacy and safety requirements that shape what it can say, store, and automate. Sesame’s broad conversational quality helps, but each vertical still rewards purpose built product design.
-
Customer service is a workflow product, not just a chat product. Intercom trains its AI agent on procedures, policies, and help content, then routes unresolved issues to agents inside the same helpdesk. That favors vendors built around ticketing, knowledge bases, and support team metrics, not a general companion shell.
-
Education tools win on teaching structure and school controls. Khanmigo is packaged as a tutor for learners and assistant for teachers, grounded in curriculum content, district deployments, and assessment modes that can limit AI help during testing. That is a very different conversation pattern from open ended companionship.
-
Health and emotional support raise the bar further. Federal health guidance flags conversational agents as security and privacy sensitive, and the FTC has opened an inquiry into companion chatbots, including harms to children and handling of health data. Replika’s privacy fine in Italy shows how quickly this becomes a compliance issue, not just a UX issue.
The next wave will produce a shared model layer underneath, with the winning products above it becoming more specialized. Sesame can still supply the natural voice and emotional responsiveness, but the biggest businesses in education, care, and service will likely be built by teams that own the workflow, trust controls, and distribution inside each domain.