Ema Raised $77 Million to Sell Work, Not Software

Ema does not want to be the next piece of enterprise software. It wants to be the reason companies buy less of it. The Mountain View startup, which sells what it calls "AI Employees" for HR, IT and finance work, announced a $77 million Series B on September 23, led by Bengaluru-based Creaegis. Existing investors Accel, Section 32 and Prosus all increased their stakes.

The round brings Ema's total funding to $140 million and more than quadruples its valuation from its last round in 2024, according to TechCrunch. The company did not disclose the new figure. The raise is entirely primary equity, with no debt and no secondary sales, TechCrunch reported.

"Enterprises do not need more software. They need work to get done," co-founder and CEO Surojit Chatterjee said in the company's announcement.

The numbers Ema chose to share

Ema was more forthcoming on traction than most startups at this stage, with caveats. Chatterjee told TechCrunch that revenue bookings have passed $150 million. That figure is multiyear contract value, not annual recurring revenue, and the company declined to give an annualized run rate. It did say revenue has grown 50-fold over the past two years, net dollar retention is around 180%, and gross margins are close to 80%.

On deployment, Ema says it has more than 50 active enterprise deals, over 1 million active enterprise users and more than 5 million actions or queries handled, per TechCrunch. Customers it names include NTT DATA, Hitachi, ADP, PwC, Google, KPMG, Wipro and Microsoft.

The press release adds case detail. At a top-five global systems integrator, Ema supports more than 240,000 associates across 65 countries and fields 2.9 million employee queries a year, with 60% ticket avoidance and a 50% leaner people-operations team. At a top-five global BPO, it handles more than 1 million IT service management tickets and more than 1 million calls a year across 15-plus languages, with up to 95% accuracy and under 10% escalated to humans. A $50 billion conglomerate went to production in four weeks across more than 20 system integrations, serving over 40,000 employees.

The 180% retention number is the one to watch. Ema says customers typically double in value as they expand from one use case to two or three. That is the land-and-expand motion that made the last generation of SaaS companies valuable, now applied to work output instead of seats.

Agents that coordinate, priced on outcomes

Ema's product is not a single chatbot. An "AI Employee" is a coordinated set of agents that runs a multistep business process end to end, such as resolving an employee benefits question, triaging an IT ticket or handling a finance request. The platform supports more than 150 frontier and open-source models, per TechCrunch, which lets Ema route each step to whichever model fits the cost and accuracy target.

That model-agnostic design shapes how Chatterjee talks about the labs. Anthropic and OpenAI are both pushing into enterprise automation, but he framed them as suppliers rather than rivals. "Progress in frontier models is actually very beneficial to us," he told TechCrunch.

Pricing follows the same logic. Ema charges per task or outcome, not per seat and not per token. For a buyer, that turns the purchase into something closer to a services contract: pay for tickets resolved, not licenses provisioned. For Ema, it means every model price cut widens margin rather than shrinking revenue, which helps explain gross margins near 80%.

Coming for SaaS and for services

TechCrunch framed the round as evidence that AI is starting to eat into enterprise software and services budgets, and Chatterjee leaned into both halves of that claim.

On software, Ema's agents typically start by wrapping existing systems of record. Over time, Chatterjee said, customers can remove the dependency. "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.

On services, the pitch is more delicate, because several of Ema's biggest customers are services firms. Wipro, PwC, KPMG and NTT DATA sell exactly the kind of implementation, support and outsourced operations work that agents can absorb. Chatterjee said they see it coming. "A lot of the services companies are working with us. They are also dramatically changing or disrupting their own business models because they understand the human-forward model may not be the best model going forward," he said.

Creaegis sees the same thing from the investor side. Managing partner and chief investment officer Prakash Parthasarathy said in the release that Ema "has been building the default execution backbone for autonomous enterprise work."

Where the money goes

Ema has nearly 200 employees across Mountain View, Bengaluru, London and Vancouver, per TechCrunch. Chatterjee, a former Google and Coinbase executive, founded the company in 2023 with Souvik Sen, formerly of Okta. The new capital goes mainly to go-to-market, where Ema says it has already hired multiple senior leaders, and to continued platform work. Its business is concentrated in the U.S. and Europe today, with expansion planned into Asia-Pacific, South America and the Middle East.

Why it matters

The enterprise agent market is crowded with pilots. What makes Ema's round notable is the shape of its disclosures: retention, margin and named production deployments with ticket volumes attached, rather than a headline ARR number. The one gap is the one analysts will press on. Bookings of $150 million over multiyear contracts are not the same as annual revenue, and Ema chose not to say what that is.

Still, the direction is clear. If agents are priced on outcomes and sold as work delivered, the budget they draw from is not the software line alone. It is the much larger line that pays for people and outsourcers to operate that software. That is the budget Ema is going after, and its biggest customers are the companies that currently collect it.