Cutting feedback loops down on a six-engineer AI team
Viable, an AI company that turns customer feedback into product insights, has processed 3.8 million data points for businesses including Latch, Uber, and AngelList — all with a team of just six engineers. Since its founding in 2020, the company has built its entire AI product with Next.js and deployed it on Vercel, relying on the platform to keep pace with rising customer demand.
Growth, though, brought a familiar bottleneck: slow feedback cycles. With feedback scattered across Slack DMs, emails, and Zoom calls, the team found it harder to move quickly. The fix came from adopting comments for feedback directly in Vercel Preview Deployments, which cut back-and-forth by 80%.
One place for review
Using Vercel's pre-production environment means Viable doesn't have to build or maintain a custom staging setup. The entire UI review process now happens inside Preview Deployments, where comments can be attached directly to the design. That has eliminated much of the follow-up questioning about where a change is located, since the context is preserved visually on the page itself.
"Once we starting using them during their beta period, it was such a clear value prop for us to start using comments for feedback across our entire product."
Dan Erickson, CEO
Scaling with a faster loop
With the feedback process streamlined in Vercel, the team now spends less time chasing down context and more time shipping. For Viable, that means delivering customer insights faster and scaling its analysis capabilities as more businesses adopt its platform. The result is a workflow where a small engineering team can maintain speed without outgrowing its infrastructure.
