close
close

Enterprise AI Copilot Moveworks surpasses $100 million in ARR

Moveworks has surpassed $100 million in annual recurring revenue, a milestone many high-profile AI startups have yet to reach.


When ChatGPT launched in November 2022, executives were scrambling to figure out how to use generative AI systems to make their companies run more efficiently. Now, after what feels like a marathon of flashy demos and announcements, they’re looking to cut through the noise and implement tools that Actually reduce costs and ensure a real return on investment.

Bhavin Shah, CEO and cofounder of conversational AI platform Moveworks, says his company helps businesses do just that. Moveworks’ AI tool automatically handles mundane tasks like troubleshooting IT issues and filing PTO claims, and it also helps employees quickly search through company documents like contracts, files, workplace policies, and calendars. In doing so, Shah says the company has helped businesses save millions of dollars by freeing up teams to spend less time answering basic questions and more time working on more important projects.

By automating the more mundane parts of HR and IT tasks, the company is seeing a level of revenue growth that’s rare among some of the most high-profile generative AI startups. Moveworks announced Tuesday that it has surpassed $100 million in annual recurring revenue (ARR), a calculation of revenue for the next 12 months based on current monthly subscriptions and contracts.

Moveworks’ revenue success comes as high-profile AI companies rack up huge valuations while revenues remain fairly small by comparison. Large language modeler Cohere, valued at $5.5 billion, was generating $35 million in ARR in July 2024, The Information reported. Enterprise AI search tool Hebbia, valued at $700 million, was generating $13 million in ARR in July, according to Techcrunch.

Meanwhile, Moveworks, a relatively unknown name in the AI ​​space, is used by about 5 million workers at more than 350 companies, including Hearst, GitHub, Toyota, and Salesforce. Their employees ask a chatbot questions like “How much can I spend if I take the team out for dinner?” and “If I have enough vacation time, can you book me the week after Thanksgiving?” The AI ​​system, which lives in Slack or Teams, connects to external tools like Workday, ServiceNow, ADP, and Microsoft, so employees can search information stored in different apps.

Founded in 2016, Moveworks is valued at $2.1 billion and has more than $300 million in venture capital from prominent investors like Lightspeed Venture Partners, Bain Capital Ventures, and Kleiner Perkins. The company started out by building a tool that automatically performed everyday help desk tasks like resetting passwords, ordering a new laptop, or providing access to programs or folders. Moveworks initially used statistical machine learning models for its chatbot and was one of the first companies to incorporate Google’s BERT model, a transformer-based natural language processing model, a technological breakthrough that gave rise to generative AI tools like ChatGPT.

But after ChatGPT, company executives began to realize that their employee-service chatbots seemed “ancient” by comparison, Shah said. Arif Janmohamed, a partner at Lightspeed Venture Partners who led Moveworks’ seed round in 2016, said the launch of ChatGPT was “a wake-up call for all businesses around the world.”

The explosion of generative AI has created new momentum for Moveworks, doubling its customer base in the past 18 months, Shah said. The software is now built on “dozens” of large language models, including OpenAI’s GPT-4o, Meta’s Llama 3.1, and its own “MoveLM” model. It’s trained on AI-generated “synthetic datasets” made up of 14 million employee conversations with chatbots, 500 million support tickets, and 400,000 web pages of enterprise-related information, such as manuals and FAQs—all of which were anonymized and annotated before being ingested into the AI ​​model.

As companies enter the market, Moveworks now faces a growing list of rivals, including the $4.6 billion enterprise AI search tool Glean and Writer, which is valued at $500 million to $700 million and is building a full-service generative AI platform for enterprises. Giants like Salesforce and Microsoft have also launched their own AI agents to automate tasks like lead identification and writing code.

Shah says there are challenges in building a system that understands the unique characteristics of each company. For example, the tool would need to be able to detect consonant-laden acronyms commonly used by defense contractors without getting confused and thinking it’s someone “keyboarding.” It would also need to be able to respond appropriately to prompts like “Where’s Michael Jackson?” by recognizing that Michael Jackson is probably the name of a conference room, not a person.

“AI companies that haven’t scaled yet have yet to see that it won’t work unless we tackle all of these hard, ugly, complicated problems, because those are the ones that make the world work,” he said.

MORE FROM FORBES

ForbesDuolingo’s Billionaire Founder Is Focusing on Artificial IntelligenceForbesMercor AI Interviewer has vetted 300,000 job candidates. Now valued at $250 million.ForbesSaudi Arabian Wealth Fund’s Big AI Bets Include Mistral, DatabricksForbesHow the Mayors of Quezon City and Freetown Are Tackling the Climate Crisis