Ema raises $77M Series B to replace enterprise software and IT services with AI agents

The round

Ema, a startup whose coordinated AI agents run multi-step processes across HR, IT and finance, has raised $77 million in a Series B round — a financing that signals where investors think corporate software and services budgets are heading.

According to TechCrunch, Bengaluru-based venture firm Creaegis led the round, while existing backers Accel, Section 32 and Prosus increased their positions. The financing was entirely primary equity, with no debt or secondary transactions. It lifts Ema's total funding to $140 million and more than quadruples its valuation relative to its 2024 round, although the startup declined to disclose the new figure.

What Ema actually sells

Founded in 2023 by Surojit Chatterjee, a former Google and Coinbase executive, and Souvik Sen, formerly of Okta, Ema markets what it calls "AI employees": systems that orchestrate multiple agents to carry out business processes spanning a company's existing applications, rather than completing one isolated task at a time.

The strategy is deliberately incremental. Chatterjee told TechCrunch that Ema first wraps itself around an enterprise's current application stack, after which customers can reduce their reliance on some of those products — or drop them entirely. He said many customers are already on that path, with large SaaS applications increasingly reduced to little more than data stores underneath Ema's agents.

The traction figures

TechCrunch reports that Ema has more than 50 active enterprise deals and over 1 million active enterprise users, and has handled more than 5 million actions and queries. Its customer list includes NTT DATA, Hitachi, ADP, PwC, Google, KPMG, Wipro and Microsoft.

Revenue has grown 50-fold over the past two years, and bookings have passed $150 million. Chatterjee cautioned that the bookings figure counts the full value of multiyear contracts, including two- and three-year deals, rather than annual recurring revenue, and he declined to share the company's current annualized run rate. More than 90% of customers have expanded beyond their initial use case, some across dozens of workflows, and net dollar retention sits around 180% — meaning existing accounts spend nearly twice as much with Ema over time.

Pricing and margins

Ema charges neither per software seat nor per AI token; pricing is tied to completed tasks and business outcomes. Chatterjee said gross margins are close to 80% and should improve as the systems learn from deployments and require less human support.

The company is also reaching into work traditionally billed by IT services firms — implementation, integration and consulting around enterprise software. Chatterjee said services companies are among those working with Ema while reworking their own business models, recognising that a labor-heavy approach may not hold up going forward.

The labs are not the competition

The funding lands as major AI players push into the same enterprise territory. TechCrunch notes that Anthropic has been extending Claude into companies' core operations, including financial and legal work, while OpenAI has fielded forward-deployed engineering teams to put AI into production alongside customers.

Chatterjee does not see the frontier labs as direct rivals. Ema's software can draw on more than 150 models, spanning frontier and open-source options, while the startup concentrates on the domain knowledge, integrations and orchestration needed to automate processes end-to-end. "Progress in frontier models is actually very beneficial to us," he told TechCrunch.

Where the new money goes

Much of the capital will fund go-to-market expansion, chiefly sales and marketing, after the startup's early years focused largely on building the product. Headquartered in Mountain View with nearly 200 employees and offices in Bengaluru, London and Vancouver, Ema has sold mainly into the US and Europe and plans to enter Asia-Pacific, South America and parts of the Middle East over the next year.

Why it matters

Ema's round is a concrete bet that AI agents are moving from copilot territory into replacement territory for both SaaS subscriptions and human-delivered IT services. The reported figures — 180% net dollar retention, task-based pricing and roughly 80% gross margins — sketch a model where software is paid for work completed rather than access granted. The financing also hints at an emerging division of labor in enterprise AI: frontier labs supply the models, while orchestration layers like Ema own the integrations and process knowledge that make those models useful inside large organizations. If the replacement dynamic Chatterjee describes spreads, the pressure will fall first on undifferentiated SaaS tools and the human-heavy services work wrapped around them.