The Sacramento Inno Awards recognize some of the year's most talented and successful players in the tech and startup community. This year, the Business Journal recognized companies, products, and leaders in the in the innovation space. Unstructured.io is this year's Innovative Company of the Year.
Sacramento-based Unstrucutured.io is developing technology that makes it easier for users to access the promise of large language models like ChatGPT-4 by creating clean and curated data for them to use.
Unstructured offers technology to make it easier for customers to access data no matter the file type, document location or layout, so that the user can better use artificial intelligence.
Unstructured was launched in July 2022. By the time the company raised its $25 million round in July this year, Unstructured had 100 customers. By early October this year, the company had more than 1,000 customers using its open-source product, said founder and CEO Brian Raymond.
Unstructured helps a user user put their own digital information into a format that is readable by programs that use generative artificial intelligence like Open AI Inc.’s ChatGPT-4.
The popular large-language model that can create new content with generative artificial intelligence was released in September 2021.
But ChatGPT-4 knows nothing about the world after that date that unless you as a user load it, feed it, and make sure it can absorb the information, Raymond said.
Also, ChatGPT-4 only knows publicly available information about a user, its documents or the organization itself. For a user to be able to get better performance from the platform, a user must input the correct information, no matter where or how that information is stored. And how that data is stores is a big deal.
Unstructured says some 80% of enterprise data lives on formats like HTML, PDF, PNG, CSV and others, and those formats can be difficult to access by ChatGPT-4 without manually transferring the data. The alternative to using Unstructured is literally loading documents by hand, he said.
In addition to converting data, Unstructured allows ChatGPT-4 users to precisely choose data.
The promise of ChatGPT-4 is that it has potential to be a giant productivity boost in many fields, but to really be useful, it also needs to not make errors, not fabricate details and not hallucinate scenarios, which are things generative AI does tend to do now, Raymond said.
Data that is based on random connections and hallucinations is not useful, he said.
One way for a customer to quality control their AI is for the customer to use its own vector database, which allows the user to park a collection of data in a private channel. The user then forces the large-language model use that database. Going a step further, a user can then force the software to account for its reasoning, he said.
The AI model can then be programmed to give a bibliography and to cite sources and data. Going a step further, a user can fact check the AI potentially by using AI.
Raymond started Unstructured last year after working for three years as a vice president at PrimerAI, a San Francisco-based company specialized in AI-powered data analysis. Raymond has been working in artificial intelligence for the past six years.
The company is based locally because Raymond grew up in Roseville and his wife is from Loomis. When the pandemic hit, they moved home from the Bay Area.
The company’s team of 35 employees works fully remote, he said. The employees are almost all engineers, data scientists and software engineers, he said.
Unstructured.io's technology is open source, and it is currently accessible for free. It will likely pivot to offering an upgraded toolkit for a fee. The company doesn’t release revenue, he said.
The users so far of Unstructured include hospitals, insurance, industry and the military, he said.
Raymond earned a bachelor’s and a master’s degree, both in political science, from the University of California Davis. He was in the doctorate program at UC Davis for political science when he was recruited to the Career Analyst Program with the Central Intelligence Agency. He worked on foreign policy in the Middle East, and specifically Iraq, for five years. He then moved on to work on the National Security Council for a year in the second term of President Barack Obama.
The funding round earlier this year into Unstructured was led by Seattle venture capital firm Madrona Venture Group, with participation from New York-based Bain Capital Ventures, San Francisco-based M12 Ventures (formerly Microsoft Ventures), Los Altos-based Mango Capital, New York-based MongoDB Ventures and San Francisco-based Shield Capital, Madrona said. As part of the financing, Madrona Managing Director Karan Mehandru and Bain Capital Ventures Partner Enrique Salem joined Unstructured’s board of directors.
