Enterprise AI is beginning to move beyond assistants that answer questions, summarize documents or help employees find information.
The next battle is over who actually does the work.
Ema has raised $77 million in Series B funding as it expands a platform designed to automate business processes across HR, IT and finance. The round was led by Creaegis, with existing investors Accel, Section 32 and Prosus increasing their investments. The financing takes Ema’s total funding to $140 million.
For Ema, the new capital arrives as enterprises start asking a more consequential question about AI: instead of adding another application to an already crowded software stack, can AI take responsibility for completing the workflows that those applications were built to support?
From AI Assistant to AI Employee
Founded in 2023 by Surojit Chatterjee and Souvik Sen, Ema describes its technology as “AI Employees.”
The idea goes further than deploying a chatbot inside an organization. Ema coordinates AI agents that can plan and execute multi-step processes across the applications a company already uses.
Rather than limiting an agent to one isolated task, the platform is designed to move between systems, perform actions, check its work and route decisions for human approval when required. Ema says its platform currently connects with more than 250 enterprise applications and systems of record.
That approach puts Ema into an increasingly important part of the enterprise AI market: the orchestration layer between AI models and the software businesses already depend on.
Ema can work with more than 150 AI models, according to Chatterjee. This allows the company to focus less on building the underlying foundation model and more on combining models, integrations and enterprise knowledge into systems capable of completing business processes.
Challenging the Traditional SaaS Model
What makes Ema’s strategy particularly interesting is what happens if those agents become capable enough.
Initially, Ema can operate across software companies already have in place. But Chatterjee argues that as more work shifts into the AI layer, some traditional applications could eventually become less important to the user experience.
In that model, employees would spend less time navigating individual applications and more time telling an AI system what needs to happen.
It is an ambitious proposition, and whether enterprises broadly replace major SaaS applications remains unproven. But Ema is already positioning its product around reducing software spending and consolidating workflows, while using outcome-based pricing rather than charging customers by software seat or AI-token consumption.
That changes the commercial model as much as the technology.
Traditional enterprise software largely monetizes access. Ema is trying to tie its economics more closely to completed work.
Enterprise Deployments Begin to Scale
Ema says its platform has reached more than one million active enterprise users and processed more than five million actions and queries. The company has more than 50 active enterprise deals, with customers including Wipro, Hitachi, ADP and PwC.
The deployments are also moving beyond small experiments.
At Wipro, Ema says its technology supports more than 240,000 employees across 65 countries and handles around 2.9 million employee queries each year. Another deployment handles more than one million IT service management tickets annually, while a Hitachi implementation moved from concept to production in four weeks and now serves more than 40,000 employees.
More importantly for Ema’s expansion strategy, Chatterjee told TechCrunch that more than 90% of customers have expanded beyond their original use case.
The company reports that revenue has increased 50-fold over the past two years, while revenue bookings have exceeded $150 million. Because that bookings figure includes the total value of multiyear contracts, however, it should not be read as annual recurring revenue.
AI Starts Reaching Into Enterprise Services
The disruption Ema is pursuing may not stop with software.
Enterprise applications often come with another significant expense: the consultants, systems integrators and IT services teams required to configure them and connect them with the rest of an organization.
Ema believes AI can automate portions of that work as well.
That creates an unusual competitive dynamic. IT services companies can work with platforms such as Ema while simultaneously confronting the possibility that agentic AI will automate some of the implementation and integration work those companies traditionally provide.
The distinction matters because enterprise AI spending may increasingly come from budgets that previously belonged to several different categories rather than a standalone AI budget.
Software, implementation, support and business-process services could gradually begin overlapping as agents become capable of executing more of the workflow themselves.
Taking Ema Into More Markets
Ema plans to direct much of the $77 million toward expanding its go-to-market organization while continuing investment in its platform.
The Mountain View-headquartered company has grown to nearly 200 employees, with offices in Bengaluru, London and Vancouver. After concentrating primarily on the United States and Europe, it is preparing to expand further into Asia-Pacific, South America and parts of the Middle East.
The funding therefore arrives at an important stage for the company.
Ema has already moved beyond demonstrating that AI agents can perform individual enterprise tasks. Its larger challenge now is proving that those agents can reliably take responsibility for complete processes across complex organizations.
If that transition continues, the impact of agentic AI may be measured by more than how many new AI products enterprises buy. It may also be measured by how much of the existing software and services stack companies decide they no longer need.