Unwrap Reads Every Angry Review So PMs Don't Have To

Every product team says it listens to customers. Then a hundred thousand pieces of feedback arrive in a single week - support tickets, one-star app reviews, half-formed complaints on social - and the listening quietly stops. Nobody has read all of it. Nobody can. Unwrap is the company built on the bet that a machine should read it instead.

Unwrap is an AI customer-intelligence platform. In plain terms: you point it at the places your customers already talk - Zendesk, app stores, survey tools, Slack, Salesforce - and it pulls the feedback together, then uses natural language processing to surface the patterns worth acting on. Not a bigger inbox. A shorter list of what is actually going wrong and what people actually want.

The company calls its approach "zero-shot" insights, which is a technical way of saying you do not have to set anything up. You do not tag categories, you do not train a taxonomy, you do not tell it what to look for. It reads the pile and tells you the themes. That distinction - the AI doing its work before you look, rather than after - is the whole product.

OriginThe problem they lived before they solved it

Ryan Millner and Ashwin Singhania met the problem at Amazon, where both worked as product managers on Alexa. Alexa generated an enormous volume of feedback, and the people responsible for the product were expected to make sense of it. They mostly could not. There was too much, arriving too fast, across too many channels.

So the decisions got made the way product decisions usually get made - on intuition, on the loudest voice in the meeting, on whatever complaint happened to land most recently. Millner has been blunt about the flaw in that. "We relied on intuition," he has said of the Alexa years. "The problem is, intuition is riddled with bias."

We relied on intuition. The problem is, intuition is riddled with bias. Ryan Millner, co-founder and CEO

The insight was that the same NLP technology powering Alexa could be turned on the feedback about Alexa. The data was already there. What was missing was something willing to read all of it, without the recency bias, the preference bias, or the politics of whoever sat closest to the decision-maker.

How it worksFrom noise to a ranked list

The mechanics are easier to see as a pipeline. Feedback comes in from many sources, gets read and clustered by the model, and comes out the other side as prioritized themes a product manager can actually take to a planning meeting.

Around that core sit four products: Customer Intelligence for the aggregate view, Surveys for asking directly, SupportIQ for the support queue, and VOC (voice-of-customer) Insights for the wider picture. The pitch to a product leader is simple - understand what customers need without spending your week reading to get there.

TractionWho is paying, and why it matters

In January 2025 Unwrap raised a $12 million Series A led by Scale Venture Partners, with Atlassian Ventures joining. The angel list reads like a who's-who of people who have felt this pain from the operator's chair: David Singleton, former CTO of Stripe; Johnny Ho, co-founder of Perplexity; Karen Ng, an SVP at HubSpot; and Tyler Schleich, an SVP at Oura. Rory O'Driscoll, a veteran of Box, Bill.com and DocuSign, joined the board.

The customer roster is the more telling signal. Microsoft, Lyft, Perplexity, Oura, JetBlue, GitHub and Sunrun are named users. These are companies that do not lack for data or engineers. They could build feedback tooling themselves. That they are buying suggests the reading problem is harder, and more universal, than it looks.

• 2018
  Millner and Singhania start building; several years of iteration follow.
• Sept 2021
  The Unwrap idea crystallizes; the company forms around NLP-driven feedback analysis.
• 2022
  Launches with early adopters; raises roughly $4M in seed funding.
• Jan 2025
  Closes $12M Series A led by Scale Venture Partners; adds board member Rory O'Driscoll.

The readWhat Unwrap is really selling

Strip away the category language and Unwrap is selling relief from a specific, familiar dread: the knowledge that the answer to your product question is already sitting in your systems, and you will never have time to find it. Voice-of-customer programs have always had this gap. The data was there. The reading was not.

The company's stated mission is to "fill the world with products people love," which is the kind of line every startup writes. The actual job is smaller and more honest - make sure someone, or something, finally reads the feedback, so the roadmap comes from evidence instead of the person who complained loudest last Tuesday. If that reading holds up in practice, four hours a week back per user is not a marketing number. It is a full afternoon returned to people whose afternoons are the product.