Human-centered artificial intelligence firm Vianai Systems Inc. today launched what it says is a first-of-its-kind AI chat assistant for institutional and retail investors, investment banking analysts and other professionals.

The assistant, called “hila,” is available to use for free in beta for a limited time, and Vianai claims that it will set a new standard for conversational AI with its ability to serve a very tangible business need.

Hila is based on a large language model and is designed to address an industry use case that demands extreme accuracy. To that end, hila provides AI-generated insights for investors that are based on reliable and trustworthy data from fully vetted sources, governed by a patent-pending technology that aims to eliminate issues with AI “hallucination,” which is a phenomenon where AI assistants may sometimes fabricate answers in the event they lack the data to provide a factual response.

Vianai said hila’s underlying LLM is trained on verified datasets for S&P 500 firms and the NIFTY 50, including things such as earnings transcripts, financial statements, income statements and balance sheets. The model is continually improved, with new financial data, corporate information and companies added every week. It can be used to surface immediate and actionable insights for investors for use in due diligence, earnings call preparation, investing decisions and more.

There are built-in safeguards, with hila always citing its sources and the text that it draws its insights from, complete with a confidence score for each statement. According to Vianai, this is key because it eliminates the need for analysts to confirm the key insights they’re basing their decisions on.

Andy Thurai, principal analyst and vice president of Constellation Research Inc., told SiliconANGLE that AI has had surprisingly little impact when it comes to analyzing things such as company earnings announcements, conference calls and videos. He said the issue is that most of this information comes from unstructured data, such as video, audio and PDF files.

“Vianai’s hila aims to solve this problem, enabling financial analysts to ask questions about very specific items from call scripts, 10K documents, income statements, cash flow and balance sheets,” Thurai explained. “While hila is still in beta mode, it seems to do a decent job with these kinds of searches. Another benefit is that analysts can upload their own PDFs and search for specific items inside of those.”

Vianai said hila’s reliability is one of the main reasons why it’s already being used by almost 2,000 investors since its launch in private beta test. Early adopters have been using hila to gauge management sentiment and priorities, identify analyst concerns, ask about quarterly earnings and more.

Mike Ostroff, an analyst with Maverick Capital Ltd., said hila has vastly improved his research process. “I can rapidly search 10-Ks and earnings calls to find if anything related to my thesis is hidden inside,” he said. “More importantly, I can do it quickly without wasting time skimming irrelevant topics or pinpointing keywords.”

Vianai said it will soon roll out more advanced, premium capabilities for hila, such as the ability to expand its dataset with user-uploaded documents and data.

Image: Vianai Systems