When it comes to digital transformations — or even simply trying to use sensor data to optimize a business process — there aren’t enough data scientists to go around. And even if there were, some problems probably aren’t worth the time and cost of getting an expert statistician involved.
This is where a bevy of startups offering no-code solutions come into play. One of these, Elipsa, was formed two years ago with an eye toward providing data analysis for financial data, but by the end of last year had switched to providing analysis for IoT data.
Elipsa makes software that takes in data and runs it against several statistical algorithms and neural networks to figure out what math is most effective at finding anomalies or trends in said data. To use the software, an employee at an Elipsa customer provides the data and then tells the software what they are looking for, such as certain anomalies. Or perhaps they want to forecast inventory availability, or pricing trends.
After sharing the data and defining the types of results, Elipsa’s software runs through several different algorithms to figure out which ones give the most accurate prediction. The software shows how confident the chosen algorithm is in the prediction and which pieces of data are most important in reaching the prediction.
During the demo, I was impressed at how easy it seemed to be to use and how well it communicated both the prediction and where the predictions might break down. Each prediction gets turned into an API that the customer can then use to export the models and predictions to other software or services.
By making it so easy, individual machines could potentially get their own individual APIs and tailored predictions as opposed to having a manufacturer try to apply one algorithm to a bunch of different machines, each running under very different circumstances. Elipsa charges per model, which means that every additional machine or API feed generates revenue.
Jeff Kimmel, a co-founder and the CEO of Elipsa, said that demand skyrocketed during the pandemic because so many companies tried to add remote capabilities and sensor-based tracking systems for employees. Plus, companies in the industrial sector or asset tracking are underserved by data scientists when compared to the finance industry, which also drove the pivot.
Elipsa is a small company with only five employees and plans to raise funds. However, it’s already inked partnerships to provide its software along with Losant, Software AG, and Crosser. Brandon Cannaday, the chief product officer at Losant and one of its co-founders, said that at least one customer is testing the Elipsa product in its operations.
Losant provides equipment and limited services for customers building out edge computing. Cannaday said that most customers have some type of AI aspirations, but they don’t have the data scientists on hand. In many cases, they may not even yet know what they want. With Elipsa, they can start down the path of getting insights in an easy and replicable way without spending a lot of money on data scientists or consultants. For many customers, this is enough, especially as they start their digital transformation efforts.
It’s worth noting that what Elipsa does wouldn’t be possible without first establishing trust in both the data and the methodologies for creating algorithms. And once companies build trust in the IoT, others can then abstract some of the hard work of turning data into insights. If a business can rely on the data and the math that drives a decision, then it’s much easier to hand that over to experts and let executives or line operators keep doing what they do best.