Enterprise AI and software company DevRev has announced the launch of Computer, a conversational AI product designed to serve as an expansive interface for enterprise employees that gives them access to all their core data and schedules, and can take actions on their behalf.
The product is currently in beta for existing DevRev customers, with a broader public release planned later in 2025. Interested users can join the waitlist through DevRev's website.
The company positions Computer as more than a chatbot or search interface. Instead of simply reading and summarizing information, Computer is built to create, update, and act on tasks across enterprise systems. It integrates with customer relationship management platforms, product backlogs, documents, and communications tools, enabling it to maintain synchronized records while giving teams contextualized insights.
Dheeraj Pandey, co-founder and chief executive officer of DevRev, said the guiding principle for Computer is that AI needs to go beyond passive search.
"AI shouldn't be read-only -- agents must update things, run workflows, and change data, not just retrieve it," he said in a video call interview with VentureBeat recently.
That design goal underpins the product's focus on unifying data and actively supporting enterprise workflows rather than just reporting on them.
DevRev plans to offer Computer on a consumption-based pricing model, designed to be accessible for startups through freemium-style entry while scaling up for large enterprises. The architecture relies on edge computing and micro-tenancy to keep costs efficient, contrasting with AI vendors that focus only on bespoke, high-cost enterprise deals.
What is DevRev?
DevRev emerged in late 2020, co-founded by Dheeraj Pandey and Manoj Agarwal -- both formerly leaders at Nutanix (a U.S.-based enterprise cloud computing and virtualization company), with Pandey as co-founder and CEO and Agarwal as SVP of Engineering.
Headquartered in Palo Alto, California, DevRev operates with a global footprint, including offices in India and other regions. The company has positioned itself as an AI-native, developer-centric platform that bridges customer support, product management, revenue operations, and engineering through a unified "enterprise knowledge graph" architected to break down data silos and enable conversational, contextual AI workflows.
In August 2024, DevRev raised $100.8 million in its Series A led by Khosla Ventures, Mayfield Fund, and Pandey's own Param Hansa Values -- propelling it into unicorn status at a $1.15 billion valuation.
Its flagship platform, AgentOS, serves as the backbone: ingesting real-time data from legacy systems to power AI-driven tools across customer service, product, and engineering workflows. DevRev pitches itself as a viable replacement or complement to traditional CRM and developer tools like Zendesk, Salesforce, Intercom, and Jira, offering a more connected, gen-AI-powered alternative.
Breaking down siloed enterprise data
DevRev's new Computer release aims to eliminate the structural fragmentation that has developed across enterprise software.
Over the past two decades, specialized tools have been built for departments rather than organizations as a whole. This has led to isolated systems where support platforms only interact with support data, CRMs remain disconnected from product teams, and engineering environments function independently.
According to DevRev, this fragmentation slows execution and creates inefficiencies. Computer is designed to overcome this by unifying structured and unstructured data, providing AI with complete business context.
Pandey emphasized that a major limitation of current enterprise search tools is their inability to bridge structured and unstructured information. "Search can't just crawl documents; it has to span structured and unstructured data in a permission-aware way," he said.
Built on patented foundations
Computer draws on two proprietary technologies DevRev has developed over the past five years:
* Computer Memory: a knowledge graph that turns enterprise data into a network mapping relationships between customers, teams, and products.
* Computer Airsync: a bidirectional synchronization engine that brings data into this network while preserving permissions, context, and compliance.
"Our proprietary knowledge graph turns enterprise data into a living network that maps complex relationships between teams, customers, and products," Pandey said. He added that Computer AirSync complements this: "Computer AirSync is a real-time, bidirectional sync engine that brings data into this intelligent graph and makes it AI-native."
These tools enable Computer to operate with the same systems of record teams already rely on. For example, if a bug affects key customers, Computer could create a ticket, assign it to the right expert, notify the product owner, draft communications for customers, and update Salesforce and Jira -- all while respecting data permissions.
Pandey argued that Computer's sync-first approach addresses weaknesses in legacy architectures that rely on dozens of runtime connections to disparate systems. "Relying on protocols that ping dozens of legacy systems at runtime is broken -- latency, security, and availability all suffer -- so we start with local memory and sync," he said.
How DevRev computer works
A recent product demonstration showed how Computer functions behind the scenes and in daily use. At its foundation, AirSync connects to systems of record like Salesforce, Jira, Zendesk, or in-house databases, and feeds that data into the Computer Memory knowledge graph. Permissions are preserved at every step, so employees only see information they are authorized to access. Once the graph is populated, AI agents can search, automate workflows, generate insights, and even execute updates across systems.
For employees, Computer can be accessed in multiple "surfaces," including a native desktop app, Slack, or Microsoft Teams. The agent recognizes the user's role and tailors its responses and skills accordingly. In a sales example, it can pull together account details from CRM, meeting records, support tickets, engineering work, and product usage data into a single, contextual overview. It can then generate a call plan, attach it to an opportunity record, and even schedule the associated meeting in Google Calendar -- all from within one conversation.
For customers, Computer can power front-end chatbots that go beyond scripted responses. In one demo, a customer noticed failed transactions. Instead of escalating the issue to human support, Computer reviewed error logs, diagnosed the problem, and corrected the account settings to allow the transactions -- all while staying conversational and context-aware.
The system also supports automation for tasks like employee onboarding, automatically provisioning accounts, devices, and benefits. In other contexts, it can draft and store documents such as RFP responses or reports in Google Drive or SharePoint, saving hundreds of hours of manual effort. DevRev supports multiple large language models, including GPT and Claude, but also allows enterprises to plug in their own fine-tuned models through existing AI gateways.
Several global companies, including Velocity Global, Bolt, Bill.com, and Uniphore, have tested Computer and its underlying technology. Reported results include:
* 85% of support tickets automatically resolved
* 50% reduction in support costs
* 10 hours saved per employee each week
Elec Boothe, director of support engineering at e-commerce checkout and finance service Bolt, told VentureBeat in an interview that adopting DevRev and its AI teammate required a wholesale rethink of support processes.
He explained that the company had been evolving its business into new areas, including marketplaces and gaming, while revitalizing its core checkout product with embedded AI. "We help conversion rates with abandoned carts, and we've recently moved toward a marketplace approach, diving into gaming and reigniting our checkout product with embedded AI," Boothe said.
Traditional systems proved inadequate for Bolt's needs. "Zendesk was just not cutting it for us. The complexity and automations didn't match the fast pace of our business, and their AI solutions were locked behind expensive subscriptions," Boothe noted. In response, Bolt rebuilt its support function in just two weeks, rethinking both its processes and automations from the ground up.
The transition allowed the company to introduce AI-driven efficiencies. Boothe described how DevRev's AI agent filtered out misdirected tickets, cutting through 3,000 irrelevant requests per month that previously flooded Bolt's European team.
Freed from repetitive triage, his staff could focus on higher-value work. He stressed that the goal was not to replace support agents, but to free them up to act more proactively -- reaching out to merchants to address potential issues before they escalated.
The practical effects were significant: Bolt cut its support response times from eight business hours to about ninety minutes, while merchants benefited from a help site powered by context-aware search rather than static keyword matches. Boothe said the results underscored a key lesson: "If you're switching systems, don't just replicate what you had. That wasn't working. Use the opportunity to rethink what your customers and team actually need."
Pandey added that Computer is designed to delight as well as deliver efficiency: "Delight matters: Computer is a keystroke away and knows your context -- calendar, documents, email, tickets, issues, opportunities -- while staying permission-aware."
Voices from DevRev
Michael Machado, corporate vice president of product and brand at DevRev, said Computer addresses long-standing challenges in enterprise workflows by transforming scattered data into context, disjointed processes into outcomes, and siloed departments into unified teams. He described the result as what DevRev calls Team Intelligence.
Pandey framed Computer as the third major shift in the history of computing. He drew a parallel to the rise of desktops in the 1980s and mobile computing in the 2000s, positioning conversational AI as the next defining era. He said Computer gives enterprises a new sense of agency by giving AI the full context it needs to collaborate with humans effectively.
Expanding on that point, Pandey explained that enterprises require more than a simple vector search to support this future. "Beyond vector search, you need a strong SQL engine and a powerful workflow engine so models can generate code, query data, and update systems," he said. He also positioned Computer as an enterprise-ready AI layer: "Think of it as a ChatGPT for the enterprise -- with the security, data-residency, and compliance DNA big companies require."
Pandey noted that onboarding has also been a focus. "With one-click AirSync, migrations that used to take a year can compress to three or four weeks -- moving data today and, going forward, code," he said.