Due diligence is a costly business and is not just in the area of investment. Even companies seeking to launch new products or investigate partnerships find and investigate appropriate data, finding and investigating appropriate data, especially when third-party institutions and consultants are involved, to make educated decisions. It takes a few weeks and can be very expensive.
A new AI startup called Bridgetown Research says it can speed up the process with an AI agent that can create dents on its cost-based basis and do most of the data collection and research work that leads to due diligence. And as part of this effort, the startup recently raised $19 million in the Series A round, which is jointly led by Accel and Lightspeed.
It was co-founded in December 2023 by CEO Harsh Sahai, former McKinsey employee and research scientist at Bridgetown Research. Easy to read format.
Bridgetown is a network of industry experts who leverage the networks that consultants and researchers often use, and can provide insight into a particular company or sector. Startups essentially partner with these networks of experts, then use AI voice agents to interview experts about the information the company needs to find for their clients.
“Insiders don't have to schedule calls with humans, so they can log on and have conversations anytime,” Sahai said in an interview. “Instead of talking to one senior executive, you can talk to people in the middle, but you can talk to more people.
Bridgetown's second set of agents uses a large-scale language model (LLMS) along with tools for clustering and regression to interpret data collected by voice agents, and pass this information to LLMS to provide answers. Let me summarize. Finally, the third set of agents uses a small language model to reproduce interpretations in a digestible way, like a presentation.
Using these agents, startups say they can create initial due diligence analyses in 24 hours using input from hundreds of respondents.
Sahai said that clients can use Bridgetown agents to collect data and insights themselves, or hire independent consultants or small consulting companies to work with agents to work with the same analyses as companies like McKinsey and Bain. He said he can get it. .
It sounds fascinating, but large language models and AI agents built on top of them still tend to hallucinate – they tend to simply make information. So how do investors trust research replicated by AI agents? Sahai says startups will address this with a “operability and auditability” approach.
This means that the client can review the data and track every step the agent takes to reach that conclusion, just like the “inference” AI model. Additionally, voice agents record conversations with experts they interview, allowing them to manually verify the information.
He added that AI agents do not rely on a single data source. Instead, it collects information from multiple sources, interprets it using large-scale linguistic models, and processes the data using fine-tuned models.
“We've never seen an approach before,” Sahai said. “Most platforms leave it to you to collect the information they need, and then they will process it on your behalf.”
Bridgetown is not the first to tackle this opportunity to facilitate due diligence. We already have startups like Mako AI and ligentiq in this space. However, Sahai believes that other platforms do not offer sufficient solutions.
Bridgetown Research has two clients in the UK and a dozen customers in the US. These include top-tier private equity and venture capital funds, consulting companies and large companies that cater to the M&A pipeline, Sahai said.