Before Enterpret had a website or even a logo, Varun Sharma was replying to an old job application he had once sent to Notion. The application had gone nowhere, but when he and his brother, Arnav began building a system to make sense of customer conversations, he dug up the email thread and wrote back. Could he speak with the company’s head of customer experience.
She agreed. The problem he described resonated immediately. Teams at Notion were drowning in user feedback scattered across tickets and social threads, with no way to interpret any of it. All he asked for was thirty days.
Enterpret had no marketing plan or sales playbook at that point. It did have a rough model and a willingness to move quickly. Over the next month, the founders scraped Notion’s public feedback from Reddit, Twitter and app store reviews, then fed everything into their early pipeline. By the time they returned with an analysis, they had surfaced insights the company had never seen compiled in one place. When asked if the same could be done for Zendesk tickets, they agreed, with one condition. Notion would have to sign. That became Enterpret’s first major customer.
The idea came years earlier. As one of the earliest employees at Amplitude, Sharma watched the analytics market grow more precise. Teams could track every tap, click and funnel drop-off across mobile and web. The sophistication improved each year, but the underlying questions remained basic. Retention curves moved and no one could explain why. Experiments launched without understanding the friction behind them. “People tracked everything, but the why was missing,” he said.
Meanwhile, real feedback was accumulating everywhere. Support tickets reached the tens of thousands. Sales calls captured long explanations of bugs and workarounds. Reddit threads became informal focus groups. “Customers were telling companies exactly what they liked and didn’t like,” he said. “But that data was sitting in silos.”
The missing link arrived through his younger brother, an NLP researcher who studied bilingual speakers. He kept pointing out major leaps in language modelling around 2019. “The research coming out in 2019 was incredible,” Sharma said. “The models were dramatically improving.” The technology had matured enough to reliably read unstructured customer conversations at scale.
Enterpret was founded in 2020 on that premise. The company raised twenty five million dollars from investors including Kleiner Perkins, Canaan Partners, Wing Venture Capital, Peak XV and Recall Capital. It now operates at a seven figure annual recurring revenue run rate.
From the beginning, the challenge was not collecting data. It was interpreting wildly different formats. A forty minute sales call that meanders before getting to the point. A support ticket involving multiple agents. A Twitter thread with nested replies. A survey answer buried inside a long prompt. “Getting the raw data in is easy,” he said. “The hard part is transforming messy, unstructured data into a unified structured format with real insight.”
This led to two architectural decisions that now define Enterpret’s system. One is a Customer Knowledge Graph that connects feedback to users, accounts, product areas and revenue impact. The other is an Adaptive Taxonomy, a five level classification system that tags feedback across every channel without manual configuration. These choices allow teams to analyse conversations with consistency rather than relying on scattered summaries. “Anyone can connect to Zendesk or Gong. Anyone can connect to the GPT five API,” he said. “The moat is the data model in the middle.”
As the product matured, Enterpret discovered that engineers were becoming its most active users. They used the system to read raw tickets and transcripts rather than filtered summaries from customer success. Designers examined workflow friction. Marketing teams observed what customers valued but the company never highlighted. Finance teams reviewed billing confusion. “Every business relies on customer lifetime value,” he said. “To improve that, you need continuous learning from your customers.”
The company also shifted from being a standalone dashboard to being a tool used directly inside development environments. Enterpret recently launched as an MCP server, allowing engineers to bring customer feedback directly into AI powered tools such as Cursor or Claude powered coding workflows.
The product and the market were evolving at the same time. Inside the company, the debate over category naming continued. Legacy firms like Qualtrics and Medallia defined the traditional “voice of customer” segment, which still carries the most search volume. But a broader shift was underway, and the team increasingly leaned toward the term “customer intelligence.” As Sharma noted, both labels describe overlapping needs, but customer intelligence now reflects the way companies consume feedback across multiple channels.
After the Series A, Enterpret chose not to immediately expand sales. Instead, the team paused to re-evaluate the product. What should the platform look like if rebuilt using everything the company had learned from customers such as Notion, Figma and Canva. That reset led to Enterpret 2.0, a redesigned system that captures all conversations in real time, updates the knowledge graph continuously, supports natural language questions and deploys AI agents to detect issues and quantify impact.
The platform meets SOC 2 standards, and the company removes personal information at the edge unless customers request otherwise. All new capabilities were added to existing subscriptions without additional charges. “We’re trying to maximize the value we can deliver to as many customers as possible,” he said.
Qualtrics and Medallia remain two of the largest and most established companies in the space, with long standing billion dollar revenue businesses that helped shape the original voice of the customer category. Newer entrants across the United States and the United Kingdom are also expanding the field with more modern, AI driven approaches, creating a competitive environment that continues to push the category forward. “One company alone cannot create a category,” he said. “Competition forces us to provide the best product and fight hard for customers.”
Teams want one place where all customer conversations are interpreted in a structured, reliable way. “We want to help companies understand and act on customer feedback with clarity and speed,” he said. Even as Enterpret expands into new workflows and channels, its direction remains anchored in the idea that customer conversations are one of the most underused sources of insight inside companies. Sharma described the long term aim in straightforward terms. “Our vision is to become the operating system for customer centricity,” he said during the interview. As Sharma put it, the company intends to keep building around the same principle which is helping organisations understand what their customers are saying and using that understanding to support better product and service decisions.
