Generative AI is firmly in control of public technology discussions today. His new San Francisco startup called Ema believes it's more than just a passing fantasy. The company has emerged from stealth today with its eponymous product, which it believes will open a new chapter in how AI, especially generative AI, will change the way we work.
“Our goal is to create a universal AI workforce,” CEO and co-founder Surojit Chatterjee said in an interview. “Our goal is to automate the mundane tasks that employees do every day in every company, freeing them up to do more valuable, more strategic work.”
The company and its investors are putting their money and profits where their mouth is. The company has already raised $25 million from an impressive list of backers, with customers it has quietly amassed while still in stealth to blow away vaporware accusations, including Envoy. Global, TrueLayer, Moneview.
What you can do with Ema: These companies use Ema for a wide range of applications, from customer service, including features such as providing technical support and tracking to users, to internal productivity applications for employees. doing. His two products at Ema, the Generative Workflow Engine (GWE) and his EmaFusion, are designed to “emulate human responses” but evolve with increased use with feedback.
As Chatterjee explains, it's not just robotic process automation (that's the 2010s), it's not just AI that accelerates certain tasks (and it goes back even further), and it's not just AI that accelerates certain tasks (and it goes back even further) It's not just another generation of AI accuracy failures waiting to happen either.
Chatterjee said Ema (short for “enterprise machine assistant”) leverages more than 30 large-scale language models and combines them with its own “smaller domain-specific models” in a patent-pending platform. In combination, it “addresses the challenges” mentioned and solves all the problems experienced so far in terms of accuracy, illusion, data protection, etc. ”
There are a lot of names added to Ema's cap table in this early round. It is co-led by Accel, Section 32 and Prosus Ventures, with participation from Wipro Ventures, Venture Highway, AME Cloud Ventures, Frontier Ventures, Maum Group and Firebolt Ventures. In addition to this, there are also notable individual supporters such as Sheryl Sandberg, Dustin Moskowitz, Jerry Yang, Dibesh Makan, and David Baszuki.
There are already dozens, perhaps hundreds, of companies building GenAI tools for the enterprise at this point, some working on solutions for specific verticals or use cases, others working on ambitious, home-run style solutions like Ema. Also includes a swing. If you're wondering why this particular He GenAI startup is getting attention from these investors, part of the reason may be due to the fact that they're already pumping up business. . But it also depends on the background of the team.
Prior to Ema, Chatterjee served as Coinbase's chief product officer through its IPO. Prior to that, he served as vice president of product for both Google's mobile advertising and shopping businesses. He himself has about 40 patents to his name in areas such as Machine Learning Enterprise, his software and ad tech.
Another co-founder, Souvik Sen, head of engineering at Ema, has an equally impressive experience. Most recently, he was VP of Engineering at Okta, where he oversaw data, machine learning, and devices. Prior to that, he was also at Google, where he served as his lead in data and machine learning engineering, with a focus on privacy and safety. He also has 37 patents of his own.
Together, these two experiences weigh the company's ambitions and its potential to execute them. But it also drops a lot of details that could reveal how it will evolve.
Take, for example, Chatterjee's expertise in e-commerce and ad technology. Given that these are the basis of how many businesses interact with customers today, it feels inevitable that businesses will figure out how Ema will evolve if it becomes operational. .
On the other hand, having founders who have previously had to address and take responsibility for data protection and privacy may make it more likely that startups won't mess them up. Or at least we can hope! After all, it's AI, and this is a Silicon Valley startup, and it's ultimately about the business at hand and how we use technology to achieve it. Focus.
At the moment, it's notable that ambitious startups are working on building products that cut across different LLM silos to achieve more advanced results. This is perhaps an early sign that over time LLMs will become more interchangeable and even commoditized than one might imagine.
And the ability to span different use cases gives the startup potential diversification, which could help grow its business and overall utility, investors say.
“While most Point Gen AI solutions offer high value for a specific use case, they are difficult to scale across or to adjacent use cases, and more importantly, large enterprises We are concerned about fragmentation and access to sensitive data by so many different applications,” Ashutosh Sharma, head of investments at Prosus Ventures in India, told TechCrunch. “Ema solves these problems and allows us to achieve high accuracy with optimal return on investment.”