In the current landscape, many companies face significant challenges in extracting meaningful insights from their data. According to Forrester, a staggering 60% to 73% of data remains untapped for analytics within the average business. This underutilization is often due to data being siloed or restricted by technical and security constraints, hindering the effective use of analytical tools.
Recognizing this pervasive issue, Anna Pojawis and Tyler Maran, former engineers at Y Combinator-backed startups Hightouch and Fair Square, decided to take action. They observed that numerous businesses, particularly those in regulated sectors such as healthcare and finance, were unable to leverage analytics due to engineering barriers.
“We’ve identified that a large portion of the market, especially in regulated industries, faces significant hurdles with data analytics,” Maran stated. “Most corporate data today doesn’t fit neatly into databases; it encompasses sales calls, documents, Slack messages, and more. Given the scale of these enterprises, standard data models are often inadequate.”
To address this challenge, Pojawis and Maran founded OmniAI, a comprehensive suite of tools designed to convert unstructured enterprise data into formats compatible with data analytics applications and AI. OmniAI integrates with a company’s data storage solutions and databases, such as Snowflake and MongoDB, prepares the data, and enables the execution of chosen models, including large language models. OmniAI’s operations can be conducted in the company’s cloud, OmniAI’s private cloud, or on-premises environments, ensuring enhanced security.
“We believe that large language models will become integral to company infrastructure within the next decade, and consolidating everything in one place just makes sense,” Maran commented.
OmniAI offers out-of-the-box integrations with models from Meta’s Llama 3, Anthropic’s Claude, Mistral’s Mistral Large, and Amazon’s AWS Titan. Applications range from automatic data redaction to building AI-powered solutions. Customers engage with OmniAI through a software-as-a-service contract, managing models within their existing infrastructure.
Despite being in its early stages, OmniAI recently secured a $3.2 million seed round led by FundersClub, achieving a $30 million valuation. The company already boasts 10 customers, including Klaviyo and Carrefour, with annual recurring revenue projected to reach $1 million by 2025.
“We’re a lean team in a rapidly growing industry,” Maran noted. “Our strategy is that over time, companies will prefer running models alongside their current infrastructure, and model providers will increasingly focus on licensing model weights to existing cloud providers.”
