New AI Tool Aims to Assist Alternative Investment Managers

Artificial intelligence tools are popping up in the hands of asset managers and asset owners, however for many, the most common AI use case is using the technology for clerical purposes, such as recording and transcribing meetings and other administrative tasks.

Startup BlueFlame AI seeks to offer new artificial intelligence products aimed at alternative investment managers. CEO Raj Bakhru speaks with CIO about the use case of these tools and what they can offer asset managers and asset owners.

AI for Alternatives Managers

Blueflame.AI, publicly launched in October, is working with asset managers and allocators across hedge funds, public markets, venture capital, private equity and credit.

For alternative managers, BlueFlame AI’s tools can organize unstructured data from numerous sources, and it can also assist managers with CIM analysis. Bakhru says that AI will not be 100% creating a one pager to show to an investment committee, but it can certainly create 60-70% of one.

BlueFlame AI’s tools are also useful for hedge funds and public market participants. The company’s tools speed up the review of earnings transcripts, expert network calls, sell side research and quarterly and annual reports.

“On the public [market] side, working through earnings transcripts, working through expert network transcripts, even on the private side working through expert network transcripts, working [with] market research reports being able to summarize and extract bullets and key data points out of those taking board decks and then other investor documents and pulling out numbers, feeding them into models. It’s really great at all of those types of things,” Bakhru says.

The company’s tools are also useful for investor relations, it can draft responses to RFPs and due diligence questionnaire requests and is also useful for fundraising from LPs, according to BlueFlame AI’s website.

“With deal sourcing on the private side. It’s really great at identifying targets … working with LP databases and investor databases and finding people who are either allocating into strategies that you operate or running those strategies if you’re on the allocator side,” Bakhru says.

Data Management

Bakhru says that asset allocators are finding even more value from BlueFlame AI’s tools than many asset managers, specifically when it comes to data management. Allocators often receive a tremendous amount of unstructured data from multiple sources, and it is often time consuming to sift through that kind of data and compile it into one system.

“If you have 100 allocations out to hedge funds and private equity managers, every month, you’re getting unaudited monthly returns from each of them,” Bakhru says. “You’re getting capital calls from the private one and you’re getting sector allocations on a lagging basis and maybe some research on a monthly basis from the hedge funds. So just a huge amount of unstructured data that’s been emailed to you… and synthesizing all this content takes quite a lot of time.”

Allocators often spend hundreds of thousands of dollars just on outsourcing operations to sift through and manage this unstructured data, however, Bakhru says that BlueFlame AI’s tools can manage and upload data to specific accounting systems with some level of manual review, something that can save allocators time and money.

“AI is really good at working through all of that content to synthesize it into something much more digestible and allows you to look cross sectionally against that dataset,” Bakhru says. Even the basics of ‘I have to take each of these monthly return estimates and stick it in my own accounting system, so I have an accurate view on what my portfolio holdings are.’ It’s really good at pulling out those numbers from the PDFs that are being emailed to you every month.”

Workflow Implementation Challenges

Security, privacy and compliance concerns raise the biggest implementation challenges Bakhru says.

Another challenge is the learning curve of new AI tools. While BlueFlame AI’s products are consumerized and do not need specialized talent to run, the company’s client success team works to train its clients to help with things like prompt engineering, a skillset necessary to make the most out of these tools.

“AI tools that help you with your work aren’t necessarily going to be simple chat bots that you can just interact with the same way as you can with ChatGPT, you have to know how to use them to get the work output that you want. Whether that’s something as simple as prompt engineering, all the way to if you have complex templates that you want to build out, getting the AI system to understand how to build those templates.”