Xpanse AI is bringing data science to businesses at light speed

1 Apr 2019

Maciek Wasiak. Image: Xpanse AI

Our Start-up of the Week is NDRC-based Xpanse AI, creator of a predictive analytics tool that gives organisations the ability to build data science models much faster.

“Xpanse AI provides an easy-to-use platform that allows businesses to build their own predictive models,” explained Xpanse AI co-founder and CEO Maciek Wasiak. “For example, they can create a retention model to predict which customers are going to leave, or a model to predict which customers are interested in a new product.

“It allows a non-technical user to deliver AI-driven predictive models from start to finish with just a few clicks.”

‘While data scientists are typically limited by time to a couple of hundred of hypotheses, Xpanse AI investigates millions of hypotheses within minutes’

Xpanse AI recently emerged as the winner of the autumn cohort at the NDRC Investor Showcase, securing €30,000 in follow-on funding.

“Our core team is an experienced group of data scientists who have boiled their 30 years of experience down into one tool which automates the bottlenecks in the model building process.

“As a result, our clients build models within one day instead of waiting weeks or months for manual data science delivery. This enables them to adapt to the market much faster than the competition and to forecast customer needs based on fresh, high-quality insights.”

The market

Xpanse AI is primarily targeting direct marketers in highly competitive industries where managing large customer bases is a significant driver of revenue, such as companies in telecoms, SaaS or utilities.

“The need we are seeing from this group is that targeting the right customer is a key part of their job, and predictive analytics is the gold standard for this task.

“Marketeers ask many questions about their customers and they want answers fast, while typical data science delivery takes at least a couple of months to find answers,” said Wasiak.

The founders

Two men standing on a street with sun shining between them.

From left: Maciek Wasiak, CEO of Xpanse AI, with Ben Hurley, CEO of NDRC. Image: Shane O’Neill/SON Photographic

Wasiak has more than 15 years of commercial experience delivering data science solutions for telecoms, banks, insurance and airline industries. “I also have a PhD in AI, which comes in handy in this line of work. Prior to Xpanse AI, I was a head of advanced analytics in O2 and a director of advanced analytics in Distinct Business Consulting.”

Shane Teehan worked for a number of years in consulting before moving to a role in O2. He became head of advanced analytics in Three and later helped set up the advanced analytics team for Dublin Airport Authority.

Ger Harte is a data scientist who worked for several years as an analytics consultant, then senior data scientist in Three and later analytics director in FTI Consulting.

The technology

When it comes to explaining Xpanse AI’s technology, Wasiak said that there are two main parts: the data science challenges and the engineering challenges.

“For the engineering side of things, we host our service on Amazon Web Services, which gives us flexibility and scalability. We have also an on-premises solution.

“In terms of the science challenges, at the core of what we do is automated data preparation linked with machine learning. It is generally accepted that more than 80pc of a predictive modelling project will be consumed by data prep, where data scientists hypothesise about what in the data can correlate with the event that they are trying to predict.

“More so, while data scientists are typically limited by time to a couple of hundred of hypotheses, Xpanse AI investigates millions of hypotheses within minutes. This cuts months off delivery times and we are first to develop this kind of technology.”

Wasiak said that there were additional challenges that needed to be solved, such as how to link the machine-learning capabilities, present the insight in a user-friendly way or translate the automated findings into English.

“We are helping companies to unlock the full potential hidden in their data. Companies are currently using predictive analytics only for a fraction of the use cases, mainly because the delivery process is slow and therefore expensive.

“What we see in the near future is a business user who has a complete insight into customers’ behaviours, and AI-driven targeting models at their fingertips – instantly.

“This will have a transformational impact on business performance for companies. Our goal is to move fast and enable these capabilities to as many companies as possible by the end of 2021,” Wasiak said.

Crunching the numbers

Wasiak said that Xpanse AI is enjoying considerable momentum at the moment.

“We have acquired several clients, some of which are already looking into increasing the number of licences. What’s interesting [is], despite our focus on customer-centric applications of the technology, we are getting traction in a quite varied space. One of our clients is a large EU-based telecom, using Xpanse AI to detect fraud. Another is a US-based consultancy forecasting spikes in electricity prices, something we did not even think Xpanse AI can do.

“At the moment we are running proof-of-concept projects for healthcare, predicting the length of stay for patients, or in the internet of things, predicting equipment failure.

“And this is on top of other clients who use Xpanse AI for more ‘typical’ applications, such as selecting high-quality leads in the sales process or customer retention. This breadth of applications, most of which are coming inbound, to us is very exciting.”

When it comes to investment, Wasiak said: “We are looking for the right fit: an investor with adequate understanding of the applications of deep tech but also willing to support us actively in scaling up globally.

“For the businesses, data science seems to be complicated enough and something close to a magic black box. Convincing them that actually a software can do automatically what a data scientist can do manually is a challenge. Like with any automation, there are always some who don’t trust it, or feel threatened by it.”

From industry to start-up

The founders of Xpanse AI are relatively new to the start-up scene. “We were all working in industry while developing and bootstrapping to date. When you’re working in industry, the world of start-ups can be somewhat hidden,” said Wasiak.

“Our first real exposure to the scene was from joining on the NDRC accelerator last September, and it has been invaluable. We joined other entrepreneurs and companies that have been on the same journey. No one has all the answers and everybody is figuring things out, but people always seem happy to share the bits they have figured out over a coffee.

“Ireland is such a small place with a high concentration of talent that a thriving start-up start up scene seems almost inevitable.”

His advice to fellow founders is do the research. “I cannot stress enough the importance of researching and validating that what you are trying to build will be attractive enough on the market. There [are] plenty of reasons why a seemingly good idea will not take off. Ideally you would be solving a problem that you and many others experience on a regular basis and there is no solution in sight.

“The good news is, there are a lot of problems like that.”

Want stories like this and more direct to your inbox? Sign up for Tech Trends, Silicon Republic’s weekly digest of need-to-know tech news.

John Kennedy is a journalist who served as editor of Silicon Republic for 17 years