Ema Raises $77M Series B for Enterprise AI Employees

The quiet line item in enterprise AI is everything required after the model answers correctly. Security reviews, system access, workflow ownership, exception handling, and the people paid to keep a clever demonstration from becoming an expensive orphan all arrive before production.

Ema has raised a $77M Series B to attack that operating layer. Creaegis led the September 23, 2026 financing, while Accel, S32, and Prosus increased their investments. Ema says the round brings total funding to $140M and more than quadruples its valuation from the prior round, although the company did not disclose that valuation.

The round is entirely primary equity, with no debt or secondary transaction, according to TechCrunch. The capital will support product development, go-to-market expansion, and growth across APAC, EMEA, South America, and the Middle East.

What Ema Is Building

Ema was founded in 2023 by Surojit Chatterjee, its founder and CEO, and Souvik Sen, its co-founder and CTO. Chatterjee previously led product at Coinbase and Google, while Sen held engineering leadership roles at Okta and Google. The company is headquartered in Mountain View, California, with offices in Bengaluru, London, and Vancouver.

Ema calls its software AI Employees. The label is designed to make a specific claim: these systems should execute multi-step work across business applications, rather than answer a question and leave the human to finish the process. Ema currently focuses its commercial offering on HR, IT, and finance, where a single request often crosses several systems and approval paths.

The platform combines workflow orchestration, integrations, organizational context, model selection, access controls, human approval points, and audit trails. Ema says EmaFusion can combine more than 100 models while its workflow layer connects to enterprise applications and builds context around how an organization works. Its public materials list SOC 2 Type II, ISO 27001, ISO 42001, HIPAA, GDPR, on-premises deployment, air-gapped deployment, and configurable governance controls.

The Production Evidence Behind the Round

The clearest public example is Wipro's deployment. Ema says its system supports more than 240,000 Wipro associates across 65 countries, handles more than 2.9M queries a year, and enables more than 100 actions across HR, IT, and payroll workflows. The case study reports that some resolutions moved from days to seconds, employee satisfaction increased 20%, and HR operations cost fell 50%.

Those are case-study claims, not independently audited financial results. Wipro's public disclosures do corroborate the scale of the operating environment, with a workforce above 240,000 across 65 countries. The useful signal is not that every enterprise will reproduce the same percentages. It is that Ema has a named deployment where agentic software is touching real policies, systems, approvals, and employee requests at global scale.

Ema also reports that revenue grew 50x over the past two years. TechCrunch says the company has more than 50 active enterprise deals, more than 1M active enterprise users, and more than 5M actions and queries. The publication also reports more than $150M in bookings, which Chatterjee clarified represents the total value of multiyear contracts rather than annual recurring revenue.

Other company-reported figures include net dollar retention of about 180%, gross margin near 80%, and more than 90% of customers expanding beyond the initial use case. Those numbers help explain the investor interest, but their durability will depend on whether customer expansion remains repeatable as Ema enters more regions and more complex operating environments.

Why Creaegis Led the Series B

Creaegis Managing Partner and CIO Prakash Parthasarathy framed the investment around production scale and product economics. The firm's thesis is that Ema can support large volumes of enterprise work without becoming a services business that requires a bespoke consulting team for every deployment.

That distinction carries much of the commercial weight in this round. Enterprise AI vendors can produce software-like gross margins on paper while quietly accumulating implementation work, custom integrations, and human support around each customer. Ema is arguing that its platform can learn from deployments, reuse domain expertise, and let a customer expand from one function into HR, IT, and finance on the same operating layer.

Accel's original investment thesis anticipated the same problem in 2024. The investor wrote that building production-grade enterprise AI was tedious, expensive, and messy, and backed Ema's attempt to abstract the data, security, authorization, model, and workflow complexity. Accel, S32, and Prosus returning for the Series B suggests the existing investors see enough expansion evidence to keep underwriting that thesis.

What the $77M Changes

The financing gives Ema room to move from product proof into a broader enterprise distribution fight. The company plans to expand sales and marketing after spending its first years primarily on product development, while continuing to invest in the platform and shorter deployment cycles.

The competitive field is wider than agent startups. DevCuration's coverage of Factory and Arcade.dev shows capital accumulating around the control and action layers of enterprise agents, while established software vendors and IT services firms reposition around AI implementation. Ema's position is that better base models help its system because it can combine them, while the durable value sits in domain context, integrations, governance, and execution across applications.

That turns the $77M round into a test of operating leverage. If each new geography and customer requires another layer of custom service, the product thesis weakens. If Ema can preserve governance and accuracy while customers add workflows on the same platform, it begins competing for money that used to be divided among software licenses, integration projects, service desks, and manual operations.

The Enterprise AI Handoff

Ema's Series B arrives as enterprise buyers become less impressed by demos and more demanding about outcomes. A Bain survey covered by DevCuration found that cost savings often trail expectations even as AI budgets rise, which puts more pressure on vendors to prove workflow results. A useful agent must survive identity controls, data boundaries, policy exceptions, human escalation, system changes, and the ordinary mess of a multinational company.

Chatterjee and Sen built Ema around the belief that software should move from being operated by people to completing work with people governing the system. The new capital lets the company carry that belief into more functions and markets. The next stretch of Ema's story will be written inside the requests that cross departments, where ownership becomes measurable and the difference between a product and a project is impossible to hide.