In late 2022, as the world became captivated with OpenAI’s ChatGPT — the most impressive generative artificial intelligence chatbot yet available to the public — asset managers were banning employees from using it at work.
Investment firms feared that their analysts, portfolio managers, and other personnel might share something proprietary with ChatGPT, which combines user input with information that’s readily available on the internet to answer questions.
“There’s concern around that because that’s your proprietary and intellectual property — all the goodies you don’t want out there in the world,” said Andrew Smith Lewis, the founder and CEO of Alai, a new AI venture studio in Los Angeles.
Companies also have no oversight or control over what employees are putting into publicly available AI tools — or what those tools are spitting out. There’s no way to track how successful, or detrimental, using them is.
Asset managers needed a safe AI sandbox to play in — so Smith Lewis started Alai nine months ago to create one. Smith Lewis previously co-founded the adaptive learning software company Cerego, and built a learning platform for the alternative investments platform CAIS.
His new bootstrapped company is developing AI tools that work in conjunction with existing large language models, such as OpenAI, Google’s LaMDA, or Microsoft’s Orca, that can also be kept isolated within a company or amongst certain users. Alai is also a consultancy that helps companies implement AI tools, then trains their employees to use them with little or no software coding. And the company has its own venture fund that will serve as a sidecar for its partners.
Alai has only a few employees, but it’s growing, has some patents pending already, and has gained some traction in the asset management industry. Smith Lewis said Alai already counts one multitrillion dollar asset manager, a major custodian, and three multibillion asset managers as clients. He declined to name which firms specifically.
Some asset managers have been leveraging AI for a long time, while others are now racing to take advantage of what BlackRock has termed a “mega trend.” Big firms can afford their own AI-focused employees, but there is a shortage of true expertise, and it could take months or longer to hire those teams, according to Smith Lewis.
Only 3 percent of investment professionals say they are proficient in artificial intelligence and machine learning concepts, according to a survey by the CFA Institute, which created a Data Science for Investment Professionals Certificate in April.
“You’ve got thousands and thousands of great employees within your firm. And if you can quickly unlock some thinking and capability there, why not do it?” Smith Lewis said. “If we can help kick that off, and we show value, I’m sure there’s staying power there.”
Initially, Alai’s multitrillion dollar manager client wanted the AI firm to visit its office and give a presentation. Smith Lewis thought that would be a flop. “How unexciting for people in the audience to sit there and listen to some guy from the outside come and talk about AI for an hour and a half, when you could find 15 talks on YouTube that were made that week that did the same thing?”
Instead, Smith Lewis suggested that Alai work with some employees at the asset manager who were already interested in using artificial intelligence. Alai gave the employees access to its Spark tool, helped them become fluent in how to use it, and effectively empowered them to come up with creative uses for it. Those employees then presented their ideas to executives at the firm, who were sold on the value in continuing to pursue AI.
The asset managers and custodian working with Alai “recognize that the real power is going to come when you figure out how to unlock your own data, your own expertise, your own research that you’ve been sitting on, as well as what’s in the heads of your employees,” Smith Lewis said.
Eventually, everyone will have access to some third-party AI offerings, like the Bloomberg Terminal’s ChatGPT-esque tool. But the real value, Smith Lewis argued, will be found in tools that can recall and synthesize an investment manager’s entire library of data and research far faster and better than any person could. There could be trends and insights that otherwise might go unnoticed.
Right now, Alai is focused on asset managers. But in the future, Smith Lewis expects it will work with education, entertainment, and lifestyle companies to help them better understand their customers and themselves.
After 20 years working at the intersection of AI and cognitive and neuroscience, Smith Lewis is confident that the “real magic” of AI is not replacing humans. “It’s basically helping them unlock their superpowers,” he said.
