Adept Takes on Collaboration with AI

By Adam Pease

Veterans of OpenAI, Google, and other market innovation leaders have come together to found a new AI company called Adept focused on using intelligence to transform enterprise collaboration. Adept has stealth launched with $65 million in Series A funding led by Greylock and Addition. In this blog, we discuss Adept and the future of AI.

Who Is Adept?

The Adept core team, led by CEO David Luan, CTO Niki Parmar, and Chief Scientist Ashish Vaswani is made up of engineers who spent time at Google working on one of the most critical innovations in the last decade of AI research: transformers. Transformers are a kind of neural network that takes an input sequence and transforms it into an output sequence. At OpenAI, Luan worked on the now-famous General Purpose Transformer 3 (GPT-3), which has turned heads with its ability to convincingly output original, human-like writing from any given input string.

At Adept, these engineers hope to pivot away from transformers towards more active, action-oriented AI models. While transformers are very good at reading and writing information, they are not quite as effective when it comes to taking action in software environments. While OpenAI famously demonstrated the power of its algorithm to beat competitors in the online game Dota 2, it is not so obvious the AI would perform quite as well when measured against an expert in Adobe Photoshop. Adept seeks to change this.

Is AI-Enabled Collaboration the Future of Work?

Adept has set its sights on the enterprise AI market with a lofty promise: an algorithm that can operate any software in the world. Rather than accepting inputs and delivering outputs, Adept seeks to build a system that will be able to dynamically use applications to complete any task that a human being could. It dubs this vision, “universal collaboration,” and describes a world in which a human knowledge worker could turn to their computer, ask for it to complete a composition in InDesign, and watch as results populate the screen.

In other words, Adept aims to deliver a collaboration co-pilot for the modern knowledge worker. Rather than seeking to replace human workers with computers, Adept aims to offload many of routine software tasks to computers. In this way, it follows in the footsteps of still-emerging programs like GitHub Co-Pilot, which leverages GPT-3 to help programmers write code. Aragon has already predicted the rise of generative content in the enterprise, and if vendors like Adept can deliver, knowledge workers are sure to benefit from this dramatic transformation of the way work gets done.

Bottom Line

The open question is how well Adept’s leaders will be able to take their AI knowledge to market with this ambitious product. Algorithms like GPT-3 have yet to find deep adoption in the enterprise, and Adept envisions an even more striking upheaval in our collective approach to work. Nevertheless, Adept’s launch is turning heads.