Full Answer

The typing preference is deeply ingrained. Smartphones made typing second nature — billions of people compose messages, search queries, and social posts by thumb every day. When the same people encounter an AI on a website, they default to what they know. Voice requires a quiet environment, a willingness to speak aloud, and trust that the system will understand. Typing requires none of those.

The data backs this up across multiple studies. Tidio found that 62 percent of consumers who prefer chatbots over human agents do so specifically because they get quick text-based answers (Blogging Wizard). This is not a rejection of AI — it is a preference for speed and control. Typing lets visitors scan, edit, and re-read before sending. Voice commits them to a linear exchange where corrections feel awkward and backtracking means repeating themselves.

This does not mean voice has no role. Accessibility, hands-free contexts, and complex explanations all benefit from voice input. The design principle is optionality: a conversation layer that defaults to text but lets visitors switch to voice mid-interaction without losing context (Seresa). When every voice interface sounds the same, personality becomes the differentiator — and personality works in both typed and spoken channels (Seresa).