Ema raises $77M as its own release undercuts the 50% ticket figure

Ema’s own press release never says support tickets fell 50% at Wipro. It credits an unnamed customer, listed under the heading “Top 5 GSIs”, with avoiding roughly 60% of tickets entirely, and separately with running its people operations team about 50% leaner. But those are two different numbers, and SiliconANGLE’s write-up collapsed them into one.

The funding itself isn’t in dispute. Ema announced a $77 million Series B on 23 September, led by Creaegis, with existing backers Accel, S32 and Prosus all increasing their investments, according to the company’s release. That takes total funding to $140 million and more than quadruples the valuation from the previous round.

Ema declined to disclose the new number. It told TechCrunch the round was entirely primary equity, with no debt and no secondary transactions, and that the earlier round closed in 2024. The startup sells what it calls AI employees: groups of agents that plan and run multi-step HR, IT and finance work inside the applications a company already uses, check their own output, and route approvals to a human when a decision is sensitive.

Which customer gets credited with what

The release describes three deployments without naming any of them. SiliconANGLE attached names to two, and in doing so changed what one of the percentages measures.

Figure	Ema’s press release	As reported
60%	“avoided approximately 60% of tickets entirely”	Not used
50%	“enabled the people operations team to run about 50% leaner”	“a 50% reduction in the number of IT support tickets” at Wipro
20%	employee satisfaction up 20% at the same deployment	Not used
70% and 30%	credited to a “$50B global conglomerate”	credited to Hitachi
Models	Ema’s blog: output of “100+ LLMs” combined	Chatterjee: more than 150 models

Part of the attribution does hold up. Ema’s own blog post names Wipro against the 240,000 associates, 65 countries, 100 plus workflows and roughly 2.9 million annual queries. What that post doesn’t carry is the ticket arithmetic. It reports response times dropping from days to seconds and satisfaction rising 20%, and it leaves both the 60% and the 50% out.

Enterprises do not need more software. They need work to get done.

Surojit Chatterjee, CEO and co-founder, Ema, in the release

Even that line moves between documents. The blog renders the second sentence as “They need outcomes”, and has Chatterjee describing enterprises stuck in “pilot purgatory” where the release says “still stuck in the pilot stage”. The substance survives, though the wording doesn’t, which is worth remembering when a write-up quotes a founder.

The scale figures the release leaves out

The most quoted proof of traction never appears in the announcement at all. Chatterjee told TechCrunch that revenue bookings have passed $150 million, then clarified that the figure includes the total value of multiyear contracts, including two and three year deals, rather than annual recurring revenue. He declined to give the current run rate. Reading a round announcement for what it omits is the whole exercise, as we set out in how to read a funding round.

What both documents agree on is growth of 50 times over 24 months, off a base Ema hasn’t disclosed. TechCrunch adds more than 50 active enterprise deals, over 1 million active enterprise users, net dollar retention around 180% and gross margins close to 80%. The release puts customer expansion differently again, saying accounts typically double in value as they add two or three use cases, where the blog says each customer spends twice its initial spend within 12 months.

PYMNTS notes the round lands about two years after Ema topped up its Series A by $36 million to reach $50 million, having left stealth in March 2024. The pitch rests on being past the pilot stage that stalls many corporate AI projects, a pattern we’ve tracked in what actually reaches production.

Ema says the money goes to its go-to-market organisation, with multiple senior leaders already hired, plus expansion into Asia-Pacific and EMEA. The figure worth watching isn’t the $77 million. It’s whether Ema ever names the customers behind the 60%, the 70% and the 30%, or publishes the recurring revenue number it withheld this week.