A highly customizable generative AI platform for alternative asset managers is becoming available to a broader client base this week. Built to meet the needs of VCs and other private investors, BlueFlame AI helps clients make use of emerging AI technologies to improve their operations and generate resource and cost savings while ensuring security, privacy and compliance.
Offerings from BlueFlame and other developers promise to provide VCs with the benefits of AI without having to hire experts to build custom solutions. Firms that are building their own AI solutions to help with deal sourcing, operations and due diligence include Alumni Ventures, AngelList, Connetic Ventures, EQT Ventures, Headline, H/L Ventures and M13. (Read our September cover story about how AI is transforming the venture industry.)
As alternative investment managers ponder how to effectively deploy AI technologies, BlueFlame aims to address their questions and help firms apply generative AI to expedite manual workflows, accelerate data quality improvements and make better use of their technology investments, the company said in a statement.
The idea is to improve clients’ access to their siloed data systems, including their portfolio companies’ data rooms, vendor contracts, legal agreements and portfolio oversight tools by using natural language commands, which can drive mass automation, CEO and co-founder Raj Bakhru tells affiliate title Venture Capital Journal.
“That will speak differently to different people within the firm,” he says. “But that has a tremendous opportunity of making the data quality going into those systems better because you’re able to interface in a natural language manner and allow for cross-system interactions through natural language commands.
“We layer on top of large language models that are already out there. We’re LLM-agnostic, so we can go to OpenAI, Cohere or Anthropic,” among other platforms.
By providing a layer of tailoring and context, BlueFlame AI makes a response to a query “industry-specific and actually firm-specific,” Bakhru notes. If a query is related to a client’s fifth fund, for example, “that needs to have context within your firm. So we’re helping provide a lot of that context and interactions into your niche and proprietary systems.”
Rather than building an LLM from scratch, which does not make sense when huge LLMs already exist and the costs to run them are enormous, “we can take the existing LLMs that are out there and tailor them so they understand our space and that client a lot better,” he adds.
Previously available only to a beta group of customers comprised mostly of close friends and family of the founding team who are in the alternative assets industry, BlueFlame is being rolled out in the US and Europe.
For now, the company, which has offices in New York and London, has bootstrap financing from its management team and members of its advisory board. They include a partner at private equity firm TJC, which has a growth equity fund, and the chief technology officer at Kayne Anderson, which has a venture fund. Both firms are focused on the mid-market alternatives space.
“They’ve joined the advisory board with the hope of at some point being able to invest [in BlueFlame],” says Bakhru. The advisory board has been assembled mostly to provide insight from a range of perspectives into the needs of the various departments within a client firm.
Unlike AI vendors that want clients to work on their platform, BlueFlame meets clients where they already work. “We’re looking to seamlessly support existing workflows and help [clients] do what they already do, but faster, as opposed to helping them do it differently,” says chief operating officer and co-founder Henry Lindemann.
“We’re going to be natively integrated into [Microsoft] Teams and Slack,” Lindemann adds. “There will be a web interface if you want to go to it, but we’re not asking people to fundamentally change their day. A lot of the learning that has to happen when you do onboard a new system is probably a reason why data quality can suffer.”
Giving senior partners access to a natural language layer makes them more likely to use the platform themselves and probably get about 80 percent of what they’re seeking, says Lindemann. “If they need that 95 percent cut with the analyst’s overlay and judgment, they can get that or the analyst can use the tool themselves in order to aggregate the information from them from the get-go.”