Facebook sets the standard for smart products based on data such as Strava. The social network giant has more than two billion customers around the world generating hundreds of petabytes of data, which they can analyse in intricate detail to work out behavioural patterns.
Few organisations can compete with Facebook's resources, but they nonetheless are being pushed to deliver smart solution.
There is a shortage of talent in advanced coding, new recruits and also senior staff who struggle to keep up with the rate of change.
There also often isn't the capacity to create these smart products.
"I believe we're at a really pivotal moment in time and there's this conflict where on the one hand we're all used to this ever-increasing pace of technology, but on the other, we're battling with the business change," says Fiehn. "Now's the time to act."
Changes at Markerstudy
Fiehn is helping Makerstudy act on these developments in a number of ways.
The company created a virtual data map of its IT infrastructure using software developed by Dockland that will act as the foundation on which they can identify future AI projects.
Two years ago, the IT team used the data to predict when the service desk would receive calls. Their internal information was combined with external factors such as weather, and stock market changes to map the correlations that affect the likelihood of calls to the service desk.
The results can help them understand the type of smart systems and services that are needed in the insurance sector.
Markerstudy has partnered up with cutting-edge tech companies to harness the power of their data to develop these products. Their collaborators include Nexthink, which uses machine learning to understand when a member of staff will likely have an issue, and Darktrace, which applies algorithms to network traffic that can spot anomalies and shut them down as soon as they emerge.
Traditional IT department typically moves in an orderly but slow fashion. To make the team at Markerstudy more agile, Fiehn flattened the hierarchy of its structure.
"We consciously have created small freely autonomous teams who move in a different direction," says Fiehn.
"For this truly to work we've had to really push down the decision-making so the teams can make a lot of the decisions themselves to move with the utmost velocity."
Machine learning and automation
Markerstudy has also partnered with DataRobot to improve the machine learning capabilities that determine its insurance pricing. DataRobot takes all of the algorithms available on the open source market and provides a platform on which they compete against each other.
As data enters the system, the algorithms are displayed on the interface in order of accuracy, allowing Markerstudy to choose the best predictive model for a specific dataset.
Sign up for CIO Asia eNewsletters.