Ema, a Mountain View startup that sells "AI employees" to run HR, IT and finance work, has raised $77 million in a Series B round. It is pitching the money as a bet against both traditional enterprise software and the IT services firms that implement it. The round, announced Tuesday, September 23, was led by Bengaluru-based Creaegis. Existing backers Accel, Section 32 (S32) and Prosus all put in more than last time. Ema has now raised $140 million in total, and the company says its valuation is more than four times what it was at its 2024 round. It declined to give an exact figure.
Ema's message to buyers is blunt. "Enterprises do not need more software. They need work to get done," said Surojit Chatterjee, the company's CEO and co-founder, in the funding announcement. Chatterjee, a former Google and Coinbase executive, started Ema in 2023 with ex-Okta executive Souvik Sen.
From pilots to production
Ema calls its product an AI employee: a system that coordinates several AI agents to carry out a multi-step business process from start to finish across the applications a company already uses. According to the company, these agents plan the work, carry it out, check their own output and send sensitive actions to a human for approval. Ema's pitch is that this avoids the long, services-heavy rollouts that have stalled many enterprise AI projects in the pilot phase.
The company shared a lot of traction data. Ema told TechCrunch it has more than 50 active enterprise deals, over 1 million active enterprise users and more than 5 million actions and queries handled. Named customers include NTT DATA, Hitachi, ADP, PwC, Google, KPMG, Wipro and Microsoft. Revenue has grown more than 50-fold over two years, from a base Ema did not disclose, and bookings have passed $150 million. Chatterjee said the bookings figure counts the full value of two- and three-year contracts, so it is not annual recurring revenue. He would not give a current run rate. He did say that more than 90% of customers have expanded past their first use case, and that net dollar retention is about 180%.
SiliconANGLE reported that Wipro uses an Ema-powered assistant for more than 240,000 employees in 65 countries. The assistant handles about 2.9 million queries a year across roughly 100 automated workflows. Ema's press release describes the same deployment without naming the customer, calling it a top-five global systems integrator. It says the assistant avoids about 60% of tickets entirely and lets the people-operations team run about 50% leaner. At Hitachi, Ema says the system went from concept to production in under four weeks, connected to more than 20 enterprise systems, and serves more than 40,000 employees. The company reports a 70% efficiency gain and a 30% drop in ticket volume there.
A shot at SaaS, and at services
Chatterjee's argument goes further than automating help desks. Ema says it first "wraps" around a company's existing applications. Over time, he argues, customers can depend less on some of those systems. "Many of our customers are already on the way to replace [large SaaS applications] completely, removing dependency on them, because they are mostly becoming like a database," he told TechCrunch.
He also sees AI doing the implementation, integration and consulting work that companies have traditionally paid IT services firms to handle. "A lot of the services companies are working with us," Chatterjee said, adding that they are rethinking their own business models. Pricing is also different from the usual software model. Ema does not charge per seat or per token. It bills based on completed tasks and business outcomes. Chatterjee says gross margins are close to 80%.
Anthropic and OpenAI have been moving into the same enterprise market, but Chatterjee does not see them as direct rivals. Ema's platform can use more than 150 frontier and open-source models, and the company focuses on domain knowledge, integrations and orchestration. "Progress in frontier models is actually very beneficial to us," he said.
Why It Matters
Ema's round is part of a larger fight over enterprise IT budgets. For two decades that money has been split between SaaS licenses and the systems integrators who install and maintain those products. Ema is going after both at once. Outcome-based pricing, if it catches on, would weaken the per-seat model that most SaaS revenue depends on. Pitching AI as a replacement for implementation work puts pressure on the labor-heavy services model. The customer list makes this sharper: Wipro, PwC, KPMG and NTT DATA are services firms themselves, and they are adopting agentic tools internally while working out how to sell them.
The numbers also need some caution. The 50x revenue growth has no disclosed base, the $150 million bookings figure covers multiyear contracts, and several of the most striking case studies come from Ema's own marketing. Still, a 180% net dollar retention rate and a Series B where every existing investor increased its stake suggest customers are expanding their use rather than stopping at a pilot. Creaegis's Prakash Parthasarathy told SiliconANGLE that Ema "stands apart," and that its platform already supports millions of employee interactions at very large organizations.
What to Watch
Ema says most of the new money will go to go-to-market spending after years of building product. It has already hired several senior leaders, including Jonathan Feldman as head of revenue, partnerships and solutions in August. The company has nearly 200 employees and offices in Bengaluru, London and Vancouver. Over the next year it plans to expand from the U.S. and Europe into Asia-Pacific, South America and parts of the Middle East. Key questions ahead: whether customers actually retire SaaS contracts, not just add a layer on top of them; whether outcome-based pricing holds up with procurement teams; and how Ema defends its orchestration layer as frontier labs and incumbent software vendors release their own enterprise agents.
“Enterprises do not need more software. They need work to get done.”— Surojit Chatterjee, CEO and Co-founder, Ema