HerculesAI Pioneers Large Language Models Before Trend

HerculesAI, formerly Zero Systems, has been automating professional services since 2017, initially in the legal field, and has long developed advanced language models, positioning itself perfectly with the rise of ChatGPT in late 2022

The company announced a $26 million Series B investment today to maintain its recent momentum.

Alex Babin, the CEO and co-founder of the company, stated that they had been developing small models with 500,000 to 2,000,000 parameters since approximately 2020.

They were running these models on edge devices for compliance purposes. However, before ChatGPT was introduced, people should have paid more attention to this aspect of their solution.

“I recall speaking to our clients and explaining to CIOs what an LLM is, and no one cared,” Babin told TechCrunch, approximately eight to nine months before the launch of ChatGPT.

Naturally, this would undergo a rapid transformation by November of that year, and the concept abruptly garnered widespread attention. This change has been instrumental in the business’s swift expansion over the past year.

Currently, the organization maintains numerous models executing three critical functions: data transformation, extraction, and verification. The initial step is conventional and entails extracting data from documents.

The second component automatically establishes a set of norms and structures around the data, but the third component, verification, is of particular significance, according to him.

Babin stated, “It is truly the Holy Grail when you can compare the extracted information and subsequently transform it into the source of truth, whether it be regulations, policies, contracts, laws, or anything else.”

This guarantees that employees are promptly alerted to discrepancies between the source materials and their use.

Additionally, the startup has been able to construct a multi-agent system on top of these services to assist in automating all of these activities. This was made possible by the three containers.

“Those multi-agent systems can be implemented in high-value, continuous processes or workflows that necessitate automated decision-making,” he stated.

This is of particular significance to his principal customers in the regulated industry. Today, this encompasses legal services, financial services, and insurance.

The company has reported a fourfold increase in growth over the past year, suggesting that its AI strategy is effective.

Thirty percent of the top one hundred law firms in the United States are their clients. Additionally, they serve numerous Fortune 500 clients, including State Farm, Standard & Poor’s, and Mercer.

Despite the additional funding, Babin intends to maintain a compact organization and prioritize improving internal processes over recruiting new personnel.

The organization currently employs approximately 75 individuals. “I fail to comprehend the necessity of employing additional personnel.”

In reality, we will allocate additional resources to automation and internal processes. He stated, “To increase our scalability, we must consume our dog food and utilize our products.”

Proof VC, Thomson Reuters Ventures, Alumni Ventures, and various industry entrepreneurs participated in today’s funding, which Streamlined Ventures led.