Workflow Owners Capture AI Avatar Value
Chris Savage, CEO of Wistia, on the economics of AI avatars
The hard part is that infrastructure and software win by opposite rules. Infrastructure buyers treat avatar generation like electricity, they want the best quality, uptime, and price, then pressure vendors as the tech becomes interchangeable. Application buyers pay for a workflow that saves a team time, like writing a training script, generating a presenter, hosting the video, embedding it on a site, and tracking who watched which seconds.
-
Wistia has already lived this split. Its core business shifted away from selling raw video infrastructure and toward marketing software, where hosting, editing, lead capture, webinar tools, and viewer analytics are bundled into one product for marketers, not developers.
-
The reason a startup struggles to do both is that each side demands a different moat. Infrastructure needs scale, cost efficiency, redundancy, and trust. The app layer needs easy UX and tight fit to one job, like HR training at Synthesia or marketing video at Wistia.
-
AI avatars are already following this pattern. Specialized generators like Synthesia and HeyGen grew fast by making video creation cheap and fast, while incumbent platforms like Wistia, Canva, Vimeo, and Veed can plug avatar, dubbing, and editing APIs into broader workflows customers already use.
The market is heading toward a stack where a few trusted model and API providers supply cheap avatar capability underneath, while the biggest value accrues to workflow owners that package it for specific teams. That favors platforms like Wistia in marketing, and pushes startups to choose either being the engine or being the destination.