Emergence AI, a generative artificial intelligence startup that’s focused on enhancing the productivity of business employees, said today it has closed on a $97.2 million funding round led by Learn Capital.

Adding to those funds, it has also secured lines of credit totaling more than $100 million, the startup said. For a company that says it’s only now exiting stealth mode, that adds up to a huge amount of money, and it suggests that the company has, at the very least, grand ambitions.

Those ambitions fall in the area of generative AI “agents” that go further than simply responding to user’s queries and prompts, and actually perform various work-related tasks on behalf of human users.

The startup, which is a division of the educational chatbot company Merlyn Mind Inc., is working on an “agent-based” AI system that can perform many of the jobs commonly performed by knowledge workers. It does this not with its own AI models, but by understanding what must be done and outsourcing those tasks to third-party models, such as OpenAI’s GPT-4o.

Emergence’s main project, known as “Agent E,” is aimed at automating tasks such as completing forms, searching for products on retail websites and finding content on streaming services such as Netflix. The company has already created and open-sourced a prototype of Agent E that’s trained on both human-annotated and synthetic data.

Orchestrating LLMs

Emergence co-founder and Chief Executive Satya Nitta told TechCrunch in an interview that Agent E is best described as an “orchestrator,” which acts like an intermediary, processing users’ requests and finding the most suitable large language model to complete that task. For example, if a user requests Agent E to write an email, it will understand that request and then find the best model to do it, based on a curated list of LLMs.

The CEO said developers who integrate Agent E in their applications will have a lot of flexibility in terms of what they can do. For instance, they can “switch to the latest open source or generalist model on demand, without having to worry about issues such as cost, prompt migration or availability,” he said.

Nitta told TechCrunch that he started Emergence after realizing that some of the technologies developed at Merlyn Mind could be applied to automating various business tasks. He explained that while existing LLMs have extremely powerful capabilities in terms of language understanding, they lack the planning and reasoning skills needed to perform complex automation tasks, even if they understand the user’s request.

Although Agent E is open-source, Emergence eventually wants to monetize its orchestrator tool through a premium, subscription-based application programming interface. That service will be capable of performing almost any work-related task, Nitta said. For instance, it will be able to process claims and documents, manage information technology systems, handle customer service duties and more.

The startup has already entered into strategic partnerships with Samsung Electronics Co. Ltd. and Newline Interactive Inc., and is planning to integrate its AI orchestrator tech into those companies’ products. Initially, those partnerships will focus on display technologies, and we can expect to see AI productivity functions in Samsung’s WAD interactive displays and Newline’s Q and Q Pro displays, Nitta said.

Emergence claims to be doing something novel, but it appears to have quite a few rivals that are also focused specifically on creating more capable AI agents. Those rivals include the likes of Google LLC and OpenAI, as well as startups such as Sierra Technologies Inc., which is developing conversational AI agents that can take actions on behalf of employees. Others include Orby AI Inc., which is training AI agents to work with a range of desktop software, and Adept AI Labs Inc., another AI automation startup that claims to be focused on “general intelligence.”

It remains to be seen which of these companies will ultimately come out on top in the nascent field of AI agents, or indeed, if any of them will even succeed. It’s notable that Adept AI, despite raising a bumper $350 million funding round in March 2023, is reportedly now short of funds and desperately seeking a bailout from companies including Microsoft Corp. and Meta Platforms Inc.

Image: SiliconANGLE/Microsoft Designer