Finding an edge against the robots needs top human expertise

What happens when the service you sell is affected by the rapid adoption of new technology? For instance, a research tool built on top of a large language model?

The impact of generativeAI on the global industry for expert insights is one such example. Professionals, academics and policymakers with specific knowledge sets help people buying companies to make the right call.

Expert networking companies reach out and collect experts as new client briefs come in. Speed and accuracy are important, but the real value is discovering something important that no one else knows.

Gerson Lehrman Group, of the United States, is one of the most prominent, but British outfits including Third Bridge, founded in 2007, and AlphaSights, set up in 2008, have established themselves as significant players, while new names such as ProSapient are jostling for business. The latter was recently included on an accelerator programme called Future Fifty, organised by Founders Forum and counting almost a third of the UK’s billion-dollar start-ups as alumni.

Expert insights can be a lucrative trade, as customers tend not to be price-sensitive. AlphaSights, for instance, made a $200m (£153 million) pre-tax profit on revenues of $584 million in the year to last December. Third Bridge posted $250 million in revenues but a pre-tax loss of $43 million. ProSapient hit revenues of £33.1 million in 2024, up 17 per cent, and narrowed its losses to £636,000.

There are all responding to the impact of generative AI on the research industry. Given the propensity of humans to show off when they know something, particularly in academic papers and on social media, tools such as Claude, Gemini and ChatGPT are making light work of digging the internet for their thoughts, as well as conducting more traditional market research.

It makes insights on specific people, technologies and companies that are not freely available even more valuable. And this is where the expert networks are increasingly focused.

• AI is teaching a university course. We need to stop it

Margo Polishchuk, co-founder and president of ProSapient, accepts that the advent of the AI assistants “makes generally available information easily consumable”.

“There is a new base level of knowledge among our customers and they want to really quickly pivot into something that is unique,” she says. “General knowledge becomes commoditised, but what is premiumised is something that is locked in someone’s mind that is not readily available. This is the type of information that is most impactful when making an investment into a private company.”

ProSapient’s AI agents and team cruise the world for experts with tactical and strategic insights, matching them against briefs from clients. Some may be drawn from their specific expertise and experience; other insights can purely be through personal connections and contacts.

Polishchuk, 37, says: “You want to find an expert who is unique and nobody else has and then you want to unpack what is in their mind. Do they have strategic insights or is it more tactical? Who are these insights useful to?” They then bring out that insight at the right time to help a client. “It gives them the competitive edge that they need on their deal.”

• AI accused me of killing three women. Now I’m suing Google

She provides a specific example: “We come across an expert who can say, ‘There is this company that is family-owned but the couple is going through a divorce. The company is going to become available for a sale.’ If you know your customer is interested in that sub-sector, it is highly unlikely other people know this company is going to become available. A private equity firm could approach the company, have a bilateral deal. That sort of insight is highly valuable.”

What is far less valuable are views and “facts” provided by people who don’t know what they are talking about. The expert advisory market is as prone to being taken for a ride by people who embellish what they know, just as in any other industry.

“As wonderful as AI is, it is making a dog’s dinner out of fraud,” says Polishchuk. People have a habit of not telling the truth, which AI isn’t brilliant at spotting, yet. Polishchuk has also spotted “experts” typing ProSapient’s questions into ChatGPT and feeding back the answers in the hope of earning a quick buck.

This isn’t new: expert witnesses used in court trials to assess evidence have been known to misrepresent the depth of their expertise. “With genAI, a tool has been given to people to really embellish what they know. It is our job to make sure only authentic insights get to our clients,” Polishchuk says.