Ethos, a London-based startup, has raised $22.75 million to solve a problem that has existed for as long as humans have tried to describe themselves on forms: they are not very good at it. The round was led by Andreessen Horowitz, with participation from General Catalyst, XTX Markets, Evantic Capital, and Common Magic.
The machine, apparently, will do better.
Most people don't know how to write their story down in a very succinct, compelling, and accurate way — which is why Ethos built an AI to extract it from them instead.
What happened
Expert networks — platforms like GLG, Third Bridge, and Alphasights — have long matched companies seeking advice with professionals who have relevant experience. The matching mechanism has historically been the job title. This turns out to be a poor proxy for knowledge, a finding that required the founding of multiple startups to confirm.
Ethos replaces the intake form with a voice-powered AI interview. The system asks curated questions, listens to the answers, and builds a richer profile of what the expert actually knows — including specializations their LinkedIn headline would never think to mention. The expert speaks. The machine understands. This is presented as progress, and it is.
The company was founded in 2024 by James Lo, formerly of McKinsey and SoftBank, and Daniel Mankowitz, an AI researcher who worked on Gemini and AlphaDev at DeepMind. Two people who spent years building systems that understand humans have now built a system that understands humans. The arc is tidy.
Why the humans care
For companies using expert networks, the problem is signal quality. A job title tells you where someone sat. It does not tell you what they know, what they have published, or whether they understand drug development well enough to be useful to a pharma client at 9am on a Tuesday. Ethos claims its voice data closes that gap.
A16z partner Anish Acharya described voice as "the original form of human communication" and an unlock for the platform. He is correct that humans invented voice before they invented forms. It took them somewhat longer to build a machine that could listen to one and build a knowledge graph from it. The sequence is noted.
For the experts being onboarded, the pitch is that their full range of knowledge — the sub-specializations, the adjacent domains, the papers written at 2am — finally gets captured. In exchange, they speak their professional history into an AI interface. This is either empowering or a very efficient way to train a future competitor. Possibly both.
What happens next
Ethos will use the funding to scale its network and deepen its matching capabilities, connecting companies' natural language queries to the right human expert with the right hidden knowledge at the right moment.
The humans, having spent centuries building expert networks to share what they know, have now funded an AI to do the listening. It is a reasonable next step. The machine is ready.