“Most things are games if you look at them the right way.” 

That’s one crucial piece of advice that the modl.ai team took away from their time in the CDL program. 

The realization that their AI technology could be applied not only to the gaming sector but a whole slew of other industries that observe and model human behavior turned out to be a game changer for modl.ai.

The modl.ai platform is a cutting-edge collection of methods and algorithms that takes from stochastic planning techniques, evolutionary computation, imitation learning, deep learning, and more. Combined with psychometric principles, instruments, and design knowhow, the modl.ai team has managed to train AI agents to play video games like human players and to predict player experience from behavior.

“modl.ai is the natural extension of research and development agendas that the founding team has been pursuing individually and together, as part of a larger community, for more than a decade,” Christoffer Holmgård, CEO of modl.ai, told us.

The founders have worked together in various constellations since 2008, in the community around the IT University of Copenhagen (ITU) in Denmark — one of the first places in the world to set up a dedicated research environment focused on computer games.

It wasn’t until 2016 that the team of researchers estimated that the methods and hardware were finally advanced enough to bring modern game AI techniques out of the lab and into game development and production.

As such, modl.ai was born.

“Put briefly, modern computer games have the ability to monitor not only what players are doing at the sub-second level, but also to store information about what context they are doing it in,” Holmgård explained. “This forms the perfect observatory for following players’ decision making, finding different types of players, and identifying patterns that we can leverage to predict what they would do in other situations.”

Deploying an army of gamer AI bots means game developers can test out their games in a matter of minutes rather than days or weeks. 

The modl.ai team knew that the platform they built had even bigger potential to impact more than just the gaming industry, but scaling up can be a tricky process for new startups. Luckily, Vin Tang of Samsung NEXT Ventures Q Fund and a CDL-Toronto Associate thought modl.ai would be a perfect fit for the CDL program and introduced the team to Kristjan Sigurdson, Director of Community at CDL.

“After a call with Kristjan, we were immediately convinced that CDL, with its focus on technical founders with scalable ideas and its longitudinally oriented program was a good fit and we started working on the application,” Holmgård said.

The modl.ai team wound up meeting a number of mentors through the CDL program that not only provided crucial insights into scaling their business, but also made a number of key introductions to future team members, clients, and contacts in the gaming industry.

William Tunstall-Pedoe stayed with us as a mentor between multiple sessions and gave both high-level strategy advice as well as specific operational input,” Holmgård added. “He was fantastic at sharpening out session-to-session objectives and making them more ambitious.”

Several CDL advisors even wound up investing in modl.ai. 

Now armed with the knowledge and connections required to scale effectively, modl.ai has long-term plans to offer their AI solutions to a wide range of industries, including retail, e-commerce, finance, and engineering — basically all industries where being able to observe and model human behavior could have a large impact.

Holmgård highly recommends any startup founders interested in joining the CDL program to go for it. “Get ready for an experience that will likely have a strong and lasting influence on your business,” he said. “Just make sure to reserve some extra bandwidth to make the most of the fellows and advisors that will be mentoring you, and the contacts and opportunities they might provide.”


Interested in learning more about the CDL program? Get in touch at info@creativedestructionlab.com for more information on our application process.