Side and Modl.ai ditch source-code access to unblock enterprise AI QA outsourcing
Summary
Side (formerly PTW) and Danish AI firm Modl.ai signed a binding 6-month MOU for AI-enabled QA
The system operates without source code or SDK access, removing the security blocker to publisher adoption
Modl.ai's prior integration-based model required full builds and plugin access, stalling enterprise deals
"No one's really found a great solution for that. We think we have."
01
Source-code access was the structural blocker to AI QA outsourcing
Modl.ai's prior model required developers to submit full builds via plugin and SDK
Founder Christoffer Holmgard concedes the integration model "became difficult over time"
Publisher IP security concerns made enterprise-scale adoption commercially unviable
This friction is why AI-augmented QA has stayed niche despite sustained publisher demand
02
Visual-input architecture eliminates the IP exposure surface
The AI observes screen output, simulates keypresses, and logs the session - no internals accessed
A second AI layer reviews the data, flags bugs, and escalates to human QA specialists
"We take the product as you would give it to a QA tester... our solution plays the game by looking at the screen"
03
Natural-language "game knowledge" priming replaces code-level context
Developers upload assets, design documents, and images, then brief the AI in natural language
Human QA specialists shape the prompting layer that instructs the AI agent
Testers label reports ("bug" vs "not a bug") to train disambiguation over time
"All of this is in natural language. We just chat with the system about our game"
04
Current scope limits the addressable pilot market to slow-mechanic titles
COO Oliver Deslandes names the fit: "puzzles, card games, stuff that's not too fast"
Fighting games and fast-action titles are explicitly flagged as "further down the line"
Early adopters are publishers willing to co-develop, not clients demanding immediate savings
"If you are at the table now with us and maybe doing some pilots with us, then you will see where this tech goes"
05
Side bets AI oversight roles can reverse QA offshoring
Side positions the AI layer as a career track: testers move into AI engineer roles at higher pay
McBride argues jobs offshored on price could return once the role shifts to AI oversight
Automation is scoped to smoke tests and build verifications, not gameplay assessment
"Instead of just doing testing... they can become an AI engineer. It allows our people to get paid more money"
What this means
For Publishers & Developers: AI QA outsourcing is now viable without handing over source code - reopen vendor conversations previously blocked by security review.
For Service Vendors: Side's value-chain repositioning sets the competitive template; pure-play offshore QA pricing loses defensibility as oversight roles move onshore.
For Investors & VC: Modl.ai's pivot from SDK integration to visual input is the unlock event - diligence AI QA startups on whether their architecture requires build access.
Events
Companies
Games
—
Locations
—
COMPILED BY GAMES ATLAS EDITORIAL
RELATED ARTICLES
EXPLORE MORE
Comments (0)
No comments yet. Be the first to share your thoughts!