Many companies using artificial intelligence are focused on applications in writing and communications, but a Pasadena-based company is using AI technology to help clients quickly analyze complex data, and it’s starting to prove itself interesting to clients, from small businesses all the way up to the Department of Defense.
Virtualitics Inc. is focused on using AI to analyze millions of items of tabular data, with AI-generated visualizations of data relationships and written explanations of the insights it finds. Its platform can process and explore datasets to pull out relationships in information without the client needing to perform heavy coding.
Virtualitics has 21 contracts so far through the Department of Defense, and closed a $37 million series C funding round in August. The company is now looking to expand its client base into the commercial sector, starting with new partner Citibank N.A., which participated in Virtualitics’ recent funding round.
Virtualitics was born out of more than 10 years of research done at the California Institute of Technology, where staff scientist Dr. Ciro Donalek worked on the machine learning and visualization generation technology that Virtualitics is based on. Donalek is now the chief technology officer of Virtualitics and co-founded the company in 2016 with Michael Amori.
According to Amori, who is now Virtualitics’ chief executive, many clients suffer from a gap in the tooling side of analytics. The typical process of data analysis begins by removing or tweaking entries that are incorrect or incomplete. Then, usually using a business intelligence tool such as Tableau Software or Microsoft Inc.’s Excel, the analyst would set up predictive models to inspect historical and current information.
“It’s in that second part of that process that we found that the traditional tools just were not good enough for increasingly complex data,” Amori said.
“Let’s say that you had to analyze a set with 100 columns and a million rows, and you’re trying to figure out what’s going on in that data. You’re going to be spending a lot of time trying all the different combinations in Tableau until you find some relationships that really makes sense. Whereas you can use AI and automatically generate visualizations to be able to quickly zoom in on what’s important in the data.”
Virtualitics’ AI can be given single or multiple datasets and be asked in plain English to find correlations or patterns. It then pulls relevant insights from the information provided and creates 3D informational graphics and charts of its findings, which are accompanied by annotations and explanations for clarification. The platform’s visualizations and annotations can then be used internally to guide company development or inserted into reports to share with stakeholders. Amori said the system can handle spreadsheets with millions of rows.
The company’s 21 active contracts with the Department of Defense include a partnership with the Air Force’s Global Strike Command and Nuclear Weapons Center. Global Strike Command officials said the organization is using the Virtualitics platform to predict aircraft maintenance needs and potential failures, manage inventory and optimize allocation of manpower. Non-federal clients include shipping company A.P. Moller-Maersk, the Institute for Systems Biology and electronic payments solution North American Bancard.
Virtualitics charges for the use of its software via a yearly subscription, and while the company would not disclose pricing, it said that prices are based on the amount of users on an account and number of workflows – the series of steps needed to complete a given task or project – used.
SMBs benefit
Another company working in the data-exploration sector of AI technology is Manhattan-based Polymer Search Inc. According to Polymer’s chief executive and co-founder, Yasser Ansari, AI tools for data science are particularly beneficial for small- to medium-sized businesses that may not possess the budget or capacity to have a data analysis team in place. He said that if a company doesn’t have a devoted data scientist or sophisticated technology system on staff, it can get caught working with simple data on a surface level and be unable to identify insights such as key relationships in customer outcomes.
“There’s a great, exciting area where very talented and very technical people do amazing things with data and with writing code in the data science realm,” Ansari said. “But if you’re not a data scientist, if you don’t have a lot of technical expertise in that domain, it’s very hard for you to work with data, especially the more sophisticated data that can be more illuminating around your business and the work that you’re doing.”
Amori said that for investor Citibank, for example, a workflow could include exploration of credit card debt. The client could import data about credit cards owners, such as their educational background, geographic location, income and age, and then task Virtualitics’ system with identifying what individual factors correlate with higher levels of debt. The Virtualitics platform will, according to Amori, save that client a tremendous amount of time and effort, and the informational graphics produced will clearly explain the findings of the analysis.
“I think what’s most exciting about these breakthroughs is giving people who traditionally have not had these powers and these capabilities the ability to do some really sophisticated stuff,” Polymer’s Ansari said. “Whether it’s writing code or doing data analysis and visualizations … if you don’t know how to work with the data to manipulate it, to normalize the data, prepare the data and clean the data, there’s lots of steps associated with working with data that you kind of have to do before you can even get an answer.”
Amori said that Virtualitics’ new funding round would help it gather resources needed to expand its portfolio of commercial clients, continue growing its Department of Defense contracts and increase its go-to-market team.
