Most businesses struggle to derive value from their data. A few years ago, Forrester reported that 60% to 73% of the data belonging to the average company goes unused for analytics. This is because the data is siloed or classified by technical and security considerations that make it difficult (if not impossible) to apply analytical tools.
Previously working as engineers at Y Combinator-backed startups Hightouch (a data sync platform) and Fair Square (a health insurance tool), Anna Pojawis and Tyler Maran saw that many companies were being “locked out” of their analytics strategies due to engineering roadblocks, so they decided to take on the challenge of solving the data value problem.
“We found that a significant portion of the market, especially regulated industries like healthcare and finance, struggles with data analytics,” Malan told TechCrunch. “The vast majority of corporate data currently doesn’t fit into databases — sales calls, documents, Slack messages, etc. And given the scale of these companies, off-the-shelf data models are often insufficient.”
So Pojawis and Maran founded OmniAI, a set of tools that transform unstructured enterprise data into something that data analytics apps and AI can understand.
OmniAI syncs with a company’s data storage service or database (such as Snowflake or MongoDB), prepares the data internally, and allows the company to run any model of its choice on the data (such as large-scale language models). Maran says that OmniAI does all of its work in its cloud, OmniAI’s private cloud, or on-premise environments, which obviously improves security.
“We believe that large-scale language models will become essential to enterprise infrastructure over the next decade, so it makes sense to host them all in one place,” Maran said.
OmniAI offers out-of-the-box integrations with models such as Meta’s Llama 3, Anthropic’s Claude, Mistral’s Mistral Large, and Amazon’s AWS Titan to address use cases such as automatically removing sensitive information from data and generally building AI-powered applications. Customers can enter into a SaaS agreement with OmniAI to manage models on their own infrastructure.
It’s still early days, but Omni, which recently closed a $3.2 million seed round led by FundersClub at a $30 million valuation, claims to already have 10 customers, including Klaviyo and Carrefour, and expects annual recurring revenue to reach $1 million by 2025, Maran said.
“We’re an incredibly lean team in a fast-growing industry,” Malan said. “We believe that over time, enterprises will choose to run models in parallel with their existing infrastructure, and model providers will focus on licensing model weights to existing cloud providers.”
