339. The Most Honest AI Adoption Story in Gaming

InnoGames didn't reach for AI to chase a trend. They reached for it because the alternative was shutting down a game people love.Thomas Lehr walks through how InnoGames kept Sunrise Village alive after the business case stopped working, taking the team from 25 people down to 3 and stabilizing a title that was slipping. This is the rare AI adoption story told with actual numbers, actual workflows, and no hiding the parts everyone else leaves out. No people pushed out, no game handed to the agents, no hype. Just a methodical breakdown of where the time went and what AI could realistically take off the table.Topics Covered:• Why the game was declining and what options were on the table before AI• What happened to all 25 people on the team• Going from 25 humans to 3 humans plus AI without killing the game• Decomposing the production pipeline to find where time actually goes• The six-step AI stage designer covering story, quests, balancing, maps, placement, and cinematics• Why building custom tooling beats subscribing to 15 shiny products• Sticking with GPT-4o for years while the pipeline just worked• What it actually costs to run the pipeline every month• The 84% of developers who reject AI at any stage, and the pushback inside the company• A direct playbook for a stuck studio making 500k a month