Brightwave’s AI agent helps asset managers find signal, and it’s fundraising fast

The finance world is constantly looking for signal: indicators of a buying or selling opportunities in the markets. Some seek out signal in pricey analyst reports, while others might find it in backdoor conversations with insiders. But sometimes, investors just need help seeing such indicators hiding in plain sight.

That’s what Brightwave promises its AI agent can find: signal in public data.

“There’s a hell of a lot of signal in the world that is not currently being transacted upon,” Brightwave CEO Mike Conover (pictured above; right) told TechCrunch. “The core job of an asset manager is to know something about the world that nobody has seen in order to identify a mispriced asset, and it’s not clear that is a task well-suited to the human attention span.”

Conover argues the task may be better suited for his company’s AI system.

Brightwave’s AI agent can generate research reports for asset managers, using generative AI to summarize news events, stories or market reports. In an example on Brightwave’s website, the AI agent summarizes a 30-page report on the state of AI from Goldman Sachs into roughly five pages, highlighting key numbers and forecasts.

The startup says customers can highlight snippets of text in its AI-generated reports to see the source for any sentence, or dig deeper into specific topics. Brightwave wouldn’t offer TechCrunch a demo of this feature, but it doesn’t sound too dissimilar to the follow-up questions feature offered by Perplexity, which recently also introduced its own financial research tool.

There’s no shortage of startups working on AI agents to help with financial research, but Conover says one of Brightwave’s biggest differentiators is its knowledge graph: a structured way of representing real-world entities and their relationships — a simple knowledge graph, for example, could label Elon Musk as the CEO of Tesla in a way automated systems could understand.

Conover worked on knowledge graphs during his PhD, while at LinkedIn, and he holds several patents in the field. He thinks more complex knowledge graphs could unlock better performance for today’s AI systems.

Brightwave this summer raised an oversubscribed $6 million seed round led by Decibel Partners, joining many other AI startups who secured seed funding this year. But unlike its peers, Brightwave has already closed another round just four months later: On Tuesday, the startup said it raised a $15 million Series A, once again led by Decibel, with participation from OMERS Ventures.

Two rounds in four months is pretty quick even for an AI startup today. So, what gives?

Sometimes quick fundraises can help when an AI startup burns through its funding on model training or hiring talent. However, the investor who led both rounds in Brightwave, Decibel partner Alessio Fanelli, says this was not the case. Instead, he says the startup has seen great traction, quadrupling its revenue in four months, and Decibel wanted to close the round quickly to make sure a larger fund didn’t swoop in and box them out of the Series A.

“We don’t really want you to go out there spending too much time talking with other investors and distracting yourself from building the business,” Fanelli told TechCrunch. “Maybe some other VC will tell you to just wait and raise at a higher price later on […] We just want to give you this round today.”

This sentiment may represent a broader pattern of thinking amongst investors today: The competition is intense to land the hottest AI startups, so VCs must promise lots of capital and do it quickly. This could help explain why some AI startups are raising money at unprecedented rates.

Conover previously co-created Databricks’ open-source AI model, Dolly, and his co-founder Brandon Kotara previously led machine learning projects at Workday. The startup wouldn’t disclose much about the models it uses, and also refused to share details about the public and licensed data the startup uses to generate its financial research reports.

However, it seems more than likely that Brightwave’s AI agent is pulling some of its information from news articles. Other AI startups, such as Perplexity, have been accused of repurposing information from journalists into products that compete with media companies.

Conover does not see news organizations as a competitor, saying his startup “would never sidestep paywalls.” For now, the CEO says Brightwave is excited to work with the world’s best news organizations in a way that respects the rights of content producers.

Source link

Leave a Reply

Your email address will not be published. Required fields are marked *