Mountain View-based Ema announced Tuesday it raised $77 million in Series B funding led by Bengaluru growth-equity firm Creaegis, with participation from Accel, Section 32 and Prosus Ventures. The all-equity round brings the company's total capital to $140 million. The startup's valuation more than quadrupled from its 2024 round, though Ema declined to share precise figures.
The financing arrives at what might be an inflection point for enterprise AI. While dozens of startups promise autonomous software agents, Ema has signed more than 50 active enterprise deals and logged over 5 million actions and queries across a user base it pegs at 1 million.
Ema builds systems that execute end-to-end workflows in HR, IT and finance departments. These are not copilots that suggest next steps. They handle tasks start to finish.
"Our customers are not running experiments; they are running their HR, IT and Finance operations on AI Employees at a scale of millions of interactions a year," CEO and founder Surojit Chatterjee said in a statement.
The company reported more than $150 million in bookings—the total value of multiyear contracts signed, not annualized recurring revenue, a distinction that matters when interpreting growth velocity. Net dollar retention sits at roughly 180 percent, according to TechCrunch. Gross margin hovers close to 80 percent.
Making the Case at Scale
Wipro deployed Ema's HR assistant to more than 240,000 employees across 65 countries. The system now automates over 100 workflows and handles roughly 2.9 million annual queries, according to the company. Response times collapsed from days to seconds. Employee satisfaction climbed more than 20 percent, Ema said in a case study published in September.
NTT DATA uses Ema's voice AI to front-end the IT service desk for a Fortune 100 insurer, fielding over 1 million calls per year in more than 15 languages. The company claims 90 to 95 percent end-to-end resolution.
Hitachi, ADP, Workday, Teleperformance, Allstate and PwC have also adopted the platform, according to Ema's Series B announcement.
The startup charges by task or outcome rather than per seat or token. The pricing model reflects a bet that software genuinely completing work, instead of assisting humans who do, requires different economics.
The Founders and the Org
Chatterjee served as chief product officer at Coinbase through its 2021 IPO. Before that he ran product for Google Shopping and Mobile Ads. He holds an MBA from MIT Sloan and an MS in computer science from SUNY Buffalo.
Co-founder and CTO Souvik Sen led engineering for data, machine learning and device teams at Okta. He worked on ML and security at Google before that.
COO Swati Trehan, who joined this year, held roles at Google, Shopify and McKinsey.
Ema now employs nearly 200 people across offices in Mountain View, Bengaluru, London and Vancouver, according to TechCrunch.
Why Creaegis Wrote the Check
Prakash Parthasarathy, managing partner at Creaegis, said in a statement that many agentic AI companies deliver compelling demos but few prove they can run in production at enterprise scale. The distinction matters. Pilots succeed when a vendor holds the customer's hand and fine-tunes prompts. Production means the system keeps running when the vendor leaves.
Chatterjee told TechCrunch that "many of our customers are already on the way to replace [large SaaS applications] completely."
He added that progress in frontier models helps Ema rather than threatens it. The platform's model-orchestration layer, called EmaFusion, taps whichever backend performs best for a given task. Ema currently orchestrates more than 150 AI models.
The Competitive Landscape
Moveworks automates helpdesk interactions. ServiceNow ships ITSM tools with Now Assist. RPA vendors UiPath and Automation Anywhere are layering generative AI into their platforms. Conversational AI vendors Kore.ai, Yellow.ai and Amelia also target employee-experience use cases.
The market remains crowded. Whether Ema's model-agnostic orchestration approach proves more durable than vendor-specific assistants will depend partly on how quickly enterprises decide they trust autonomous systems to run mission-critical operations without supervision.
Funding History and What Comes Next
Ema previously raised a $25 million Series A in March 2024, then extended it to $50 million total by July 2024. KPMG made a minority investment in October 2024 and began reselling Ema's platform.
The company has moved quickly. It launched publicly less than two years ago.
"Enterprises do not need more software," Chatterjee said. "They need work to get done."
That framing suggests Ema sees itself less as a SaaS vendor and more as a services replacement. If the unit economics hold at scale, the shift could reshape how companies think about headcount in back-office functions. The question is whether the technology can sustain the autonomy Ema claims when edge cases multiply and customer environments grow more complex.
