A funding round built on a bigger claim
Ema, an enterprise artificial intelligence (AI) agent company based in Mountain View, California, closed a $77 million Series B round on September 23, 2026, taking its total funding to $140 million. Creaegis, a Bengaluru-based investment firm, led the round; existing backers Accel, Section 32 and Prosus all increased their stakes. The company was founded in 2023 by chief executive officer (CEO) Surojit Chatterjee, a former Google and Coinbase executive, and Souvik Sen, a former Okta executive. Ema says the new round values it at more than four times its 2024 valuation, though it has not disclosed the actual figure.
Agents, not chatbots
Ema's product is built around what it calls "AI employees": prebuilt agents aimed at specific HR, IT and finance functions, plus a workflow engine customers can use to build their own. The company draws a distinction between this and a chatbot: rather than answering a question and stopping, an Ema agent plans out the steps needed to complete a task, carries out those steps inside a company's existing software, checks its own work, and flags anything sensitive for a human to approve before it goes through.
The numbers, with a caveat
Ema says it now has more than 50 active enterprise deals, over a million active users, and handles more than 5 million agent actions and queries. Revenue has grown 50-fold over two years, and revenue bookings, the value of multi-year contracts rather than annual recurring revenue, have passed $150 million. The company also cites roughly 180 percent net dollar retention and close to 80 percent gross margins.
All of those figures come from Ema itself and have not been independently audited. That matters because the broader claim Ema is making, that AI agents are starting to displace conventional enterprise software, is easier to support with impressive-sounding numbers than to prove. Separate industry estimates put the share of generative AI pilots that reach measurable return on investment at scale in the single digits, a reminder that vendor-reported traction and typical market outcomes are two different things.
Named customers, concrete claims
Some of Ema's customer claims are more specific than its topline numbers. At Wipro, the company says its agents reach 240,000 employees across 65 countries, handle 2.9 million queries a year across 100 automated workflows, and have cut information technology (IT) support tickets by half. At Hitachi, Ema says a deployment went from concept to production in under four weeks, now touches more than 20 internal systems, and has cut support tickets by 30 percent while raising employee efficiency by 70 percent. The company also lists NTT DATA, ADP, PwC, Google, KPMG and Microsoft among its customers.
Betting against the subscription model
Chatterjee has framed Ema's ambition directly against the software-as-a-service (SaaS) model that dominates enterprise software today. He said customers are "already on the way to replace [large SaaS applications] completely, removing dependency on them, because they are mostly becoming like a database." The company is pushing toward outcome-based pricing, charging for tasks completed rather than per user seat or per token, a structure that only works if its agents reliably finish work without much human cleanup.
Chatterjee also argues Ema is not competing with the foundation model labs building the AI underneath it: "Progress in frontier models is actually very beneficial to us." Creaegis managing partner Prakash Parthasarathy struck a similar note about scale, saying Ema's platform "already supports millions of employee interactions at some of the world's largest organizations."
What it means in practice
For a business evaluating enterprise software today, the honest read is that agent platforms like Ema are mature enough to run real back-office workflows at large customers, not just pilots, and that outcome-based pricing is a genuine, growing alternative to the per-seat SaaS model. But the specific claim that agents will replace, rather than sit on top of, existing systems is still Ema's sales pitch more than an independently measured trend. Buyers evaluating this category should ask vendors for audited or third-party-verified metrics rather than headline growth figures, and should pilot outcome-based pricing on a narrow, measurable workflow before committing a full department's software budget to it.
