naoo AG Launches METIS: Real-Time Experimentation Infrastructure Completes Proprietary AI Stack

EXECUTIVE SUMMARY

naoo AG (Düsseldorf Stock Exchange: NAO) announces the successful deployment of METIS, a real-time experimentation and decisioning infrastructure that represents the third major AI initiative within the naoo platform stack. METIS operates as the optimisation layer above GAIA (the data platform) and ModelKnife (the machine learning framework), completing a proprietary three-layer AI architecture designed to enable continuous, measurement-driven product development.

The launch of METIS follows the Generation V platform relaunch completed in late March 2026, which delivered a significant increase in user activity: feed impressions approximately doubled and time spent on the platform increased by more than 50% as a direct result of the rebuilt platform architecture. METIS does not account for these gains — they are attributable to Generation V. METIS is the infrastructure now in place to systematically build on, optimise, and scale these improvements going forward.

KEY OPERATIONAL DATA POINTS

• Generation V impact (platform relaunch, late March 2026): feed impressions approximately doubled; time spent on platform increased by more than 50%. Both metrics are attributable to the Generation V rebuild, not to METIS.
• METIS role: real-time experimentation infrastructure deployed to optimise and scale the performance gains generated by Generation V on a continuous, measurement-driven basis going forward.
• AI infrastructure stack: METIS is the third proprietary initiative, following GAIA (data layer, May 2025) and ModelKnife (machine learning layer, December 2025). Together they form a closed-loop architecture: data ingestion → model training → real-time experimentation → optimisation.
• METIS scope: real-time testing and optimisation of all platform elements including content ranking, feed algorithms, AI models, engagement mechanics, gamification logic, and monetisation flows — significantly broader than conventional A/B testing infrastructure.
• System components: real-time event ingestion, structured event catalogue, reusable metric definitions (conversion rates, engagement KPIs), experiment management with statistical power and significance tracking, and a decision queue for product rollout decisions.
• Operating model shift: product development, AI optimisation and monetisation logic can now be continuously tested and refined based on real-world performance, reducing reliance on assumption-driven decision-making.
• Expected outcomes: management expects METIS to contribute to further increases in user engagement, improved retention and more efficient monetisation over time.
  
  

"At scale, the ability to continuously test and compare different versions of every feature is what differentiates high-performing platforms. With METIS, we can systematically evaluate what works best and continuously apply those improvements across the entire platform."

— David Liu, Lead Data Scientist & Head of AI, naoo AG

"The strong increase in impressions following the Generation V rollout demonstrates the effectiveness of our new platform, and METIS enables us to systematically scale these improvements."