AI shifts video value to workflows

Diving deeper into

AI and the future of video

Document
if someone comes out with a better thing that is faster, better, faster, cheaper, we switch.
Analyzed 5 sources

This is why video infrastructure tends to behave like a commodity until a vendor creates a clear product edge above raw compute. Wistia is willing to swap underlying providers because speed, reliability, and unit cost flow straight into its own gross margin and pricing. The durable value is usually not the encoder or model itself, but the workflow it unlocks for the end customer, like easier editing, analytics, lead capture, and publishing inside one product.

  • The comparison Chris makes to CDNs is important. In both CDN and video infrastructure, buyers keep testing alternatives, and vendors have to keep lowering price while improving performance or they get replaced. That is why infrastructure winners need real scale advantages, not just a good relationship.
  • Wistia learned this early by moving up the stack. Instead of selling storage and bandwidth, it built marketer specific features like email capture inside the video player and integrations into marketing automation tools. Customers then bought leads and workflow outcomes, not price per gigabyte.
  • AI makes this switching dynamic even harsher. Wistia explicitly chose to stay at the application layer and plug in outside AI models, while Tavus is betting the opposite, that specialized avatar models can be the infrastructure layer for many apps. The market is splitting between workflow owners and model providers.

Going forward, the easiest money in video will keep moving away from raw infrastructure and toward products that package multiple steps into one simple workflow. The providers that win underneath will be the ones that keep getting cheaper and better fast enough to stay invisible. The providers that win on top will own the user, the workflow, and the business outcome.