Professor Game Podcast, Episode 436

Using AI to Build Better Learning Games

Guest: Leif Sorensen (Founder of Acti)
Host: Rob Alvarez
Date: March 16, 2026

Episode Overview

This episode centers on how artificial intelligence (AI) is changing the game for the learning industry—specifically, how AI supports the design and use of engaging learning games. Rob Alvarez interviews Leif Sorensen, founder of Acti, a global learning games platform, about AI’s impact, the biggest practical challenges, and actionable strategies to maximize learning impact using games. Leif shares hard-won lessons, the critical importance of focus, and the pitfalls of “magic powder” thinking in gamification. The conversation is lively, tactical, and accessible both for L&D professionals and innovators new to this space.

Key Discussion Points & Insights

Leif’s Background and Daily Work

[02:40 - 05:50]

• Leif is deeply passionate about learning—games are his tool, not his end-goal.
  • “Learning is actually the core of my skill set. That’s where my love actually lies... Games are just such a powerful tool in that toolbox.” (Leif, 03:10)
• A typical day:
  • Starts with stand-up meetings with Acti’s developers
  • Focused on product development, client engagement (especially outside Denmark), and making game building faster and more accessible through AI
• Global reach: Only 30% of Acti’s business is in Denmark—the rest is worldwide.

The Classic Failure: Too Much in One Game

[06:14 - 14:35]

• Leif’s most common failure: letting games become too complicated at a client’s request.
  • “They think it’s a magic powder that you can spread over and then everything will be solved… and that’s simply not the case.” (Leif, 01:05 / 06:48)
  • "The biggest mistake is actually going with that flow... not adding too much into the game." (Leif, 06:48)
• Real-world story:
  • Leif agreed to condense a two-day L&D certification workshop into one day for a major telecom client.
  • Outcome: 32 of 35 participants left frustrated and incapable of using the game.
  • “I should never, ever, ever have said yes to that situation. And I was so much right… nobody was capable of using the game afterwards.” (Leif, 10:15)
• Key Lessons:
  • Hard conversations upfront with clients are essential to managing expectations.
  • It’s easier to negotiate before you say yes—after that, all the burden is on the provider.

COVID, LMS Failures, and the Need for Quality

[15:40 - 18:02]

• COVID-19 spurred a wave of investment in LMS/LXP systems, but engagement lagged.
  • “Nobody is visiting the LMS and the LXP. It was a total failure of investments into a system where there was no engagement.” (Leif, 15:40)
• The key: focus on quality (depth of learning) over overwhelming content or quantity.

Process of Designing Learning Games with AI

[18:19 - 22:02]

• Pre-AI: Building a custom learning game took a month; required deep interviews and manual scenario design.
• With AI:
  • Now, clients can use prompts to quickly generate games based on chosen theories/models (e.g., Daniel Goleman’s emotional intelligence).
  • The more context provided, the better the AI-generated game.
  • The role of friction: Real learning games need true dilemmas—obvious “right-answer” games don’t engage or teach deeply.

The Very Real Challenges of AI in Game Building

[22:02 - 27:34]

• AI’s output is only as good as its prompts; non-experts can get lackluster results.
• Major technical challenges:
  • Game balance: AI sometimes had characters use all resources up front or move through the game in a boring, predictable way.
  • Needed to ensure choices led to meaningful outcomes—some options must be stronger/weaker, creating tension and real decisions.
• Partnership with Roskilde University enabled Acti to spend a year refining their AI-based game engine for context, theory connection, and balanced gameplay.
  • “It took us a year… but that actually allowed us to create a really great engine.” (Leif, 27:11)

Inspiration and Recommendations

[27:34 - 31:24]

• Recommended Guest: Sune Gul Dixon (Professor at Danish Design School, LEGO Serious Play expert)
  • “He is an amazing nerd in the game and... always inspiring.” (Leif, 27:55)
• Book Recommendations:
  • For transfer of learning:
    • Transfer of Learning by Robert E. Haskell (English)
    • Transfer by Velgrain & Aarkrog (Danish)
  • Why: “Scrap all the noise and go with the one thing... If you want people to take something from one context to the other, it’s actually very difficult… that comes from the transfer theory.” (Leif, 28:59)

Personal Favorite Game

[31:31 - 34:39]

• Leif’s current favorite: Derby (“Horse Race”), a complex, engaging board game about capitalism—like Monopoly, but with more depth and story twists.
  • “It had enough elements of coincidence... all the elements combined game elements-wise, it was a really, really fantastic, balanced game.” (Leif, 33:37)

Notable Quotes & Memorable Moments

• On keeping games focused:

  “The biggest mistake is making games too complicated, we add too much into them or we expect too much of them.”
  (Leif, 06:48)

• On “magic powder” thinking:

  “Clients… think it’s a magic powder that you can spread over and then everything will be solved, almost like a wand that you can just put on anything… and that’s simply not the case.”
  (Leif, 01:05/06:48)

• On AI challenges:

  “AI is giving you some output because it’s a specialist in giving you output, but not necessarily in giving you very good output.”
  (Rob, 22:02)

• On what makes a great game:

  “You have to have a game where you have dilemmas, where there’s really something at stake for the player. So you don’t know what to choose.”
  (Leif, 18:19)

• On new facilitators using games:

  “Don’t worry about what is said in the game, what is the options… Focus on how you lock people in. Focus on how you open up reflections… Ask the ones who played, it’s them who just played it. It’s not you.”
  (Leif, 34:57)

Timestamps for Key Segments

• Leif’s background & daily work – [02:40–05:50]
• Biggest failure: overcomplicating games – [06:14–14:35]
• COVID and the lesson of LMS disengagement – [15:40–18:02]
• Process for building learning games (pre- & post-AI) – [18:19–22:02]
• AI’s unique challenges in game creation – [22:02–27:34]
• People and book recommendations – [27:34–31:24]
• Leif’s favorite game (Derby) – [31:31–34:39]
• Final advice for educators and L&D professionals – [34:57–37:10]

Final Advice from Leif

Leif urges new facilitators not to worry about mastering the content or theory of a game before using it—instead, focus on the flow, reflection, and engagement.

“Don’t worry about what is said in the game, what is the options… Focus on how you lock people in. Focus on how you open up reflections… Ask the ones who played, it’s them who just played it. It’s not you.” (Leif, 34:57)

He also emphasizes that the hardest part for newcomers is the first couple of uses, but practice and engagement make it far easier very quickly.

Learn More

• Explore Leif Sorensen’s work and the Acti platform: https://acti.dk
• Further resources and professional community at Professor Game: Professor Game Community on School

For anyone building games for learning, this episode is a masterclass in the do’s, don’ts, and nuances of both old-school and AI-driven approaches.