Data discovery startup Explorium Ltd. revealed today that it has landed $75 million in a new round of funding that brings its total amount raised to more than $127 million.

Insight Partners led the Series C round, which included participation from existing investors Zeev Ventures, Emerge, F2 Capital, 01 Advisors and Dynamic Loop Capital.

Explorium offers an unusual service for companies that need more data for their artificial intelligence models and predictive analytics projects. The company provides external data from relevant sources to enhance enterprises’ internal datasets, so they can create more accurate AI models and predictions.

The company argues that although organizations already have a fair amount of data to work with, they can benefit from extra outside data to bulk up the information they already have. The COVID-19 pandemic has created a realization among enterprises that the data they do have is not enough, Explorium said. It cites a report from McKinsey & Co. that revealed the limitations of predictive models based entirely on internal data that were exposed by the pandemic.

“In a few short months, consumer purchasing habits, activities, and digital behavior changed dramatically, making preexisting consumer research, forecasts and predictive models obsolete,” the report noted. “Moreover, as organizations scrambled to understand these changing patterns, they discovered little of use in their internal data.”

Dave Menninger, an analyst with Ventana Research Inc., stressed the importance of third-party data in machine learning projects too, saying that almost 80% of the companies his firm works with are using external information to enrich their models. “Just a few years ago, enterprises were trying to get machine learning up and running. Now they’re realizing the importance not just of data quality, but data diversity,” he said.

Explorium provides that diversity by gathering data from thousands of publicly available sources, co-founder and Chief Executive Maor Shlomo told SiliconANGLE, giving companies a competitive edge over rivals that don’t have access to external data.

“Explorium is the missing link: It helps them automatically discover thousands of relevant external data signals, understand the impact of those signals and use that data easily in their predictive models,” he said.

Shlomo gave the example of an online lender that’s trying to use machine learning to detect fraud among the thousands of loan applications it receives. He said that a lender that only uses its own data will be able to feed its model a more limited amount of information. That means it won’t be as accurate as one trained on 10 times the amount of data through external sources.

“One online lending platform had almost $3 for every $1 dollar tied to a fraudulent application,” Shlomo said. “With Explorium, it generated new features for a model that could dynamically detect fraud and boosted its application fraud detection rate to 92%.”

The company works with its customers to analyze their existing datasets first to work out what kind of external information is most likely to benefit them, Shlomo explained. It then sets about collecting that data and provides it to the customer in a ready-to-use format that complies with data regulations.

When it comes to training AI models, a lot of companies overcome the lack of data diversity by using what’s known as “synthetic data” to bulk out their existing datasets. But Shlomo said Explorium’s external data is superior in many instances, especially when it comes to business analytics that inform decision making.

“Explorium provides real-world data that enriches actual data sets in production that businesses use to make mission-critical decisions about credit, fraud or market targeting,” he explained. “Unlike synthetic data, it’s not just for training models.”

Explorium said it has more than doubled its customer base and quadrupled its revenue since the start of the coronavirus pandemic. The company recently added a second product to its lineup too. The Signals Studio announced last month complements its External Data Platform by enabling data teams to find and integrate quickly the most relevant data signals to their analytics pipelines.

“Machine learning is key to our competitive strategy, and external data is what powers it,” said Elad Zoldan, head of data at Melio Payments Inc. “Before we started using Explorium, we had to search for data source by source and strike separate deals with different providers. It wasn’t a scalable or flexible solution. With Explorium, we have a universe of external data at our fingertips, and an incredibly efficient way to figure out what data will make the biggest impact.”

Looking ahead, Shlomo said Explorium will use its latest funds to add new features to its platform and provide wider data coverage. In addition, it plans to double its team across all functions and expand its U.S. presence by opening a new office in New York.

Image: kreatikar/pixabay