
Accedo will showcase its new AI-driven orchestration platform, Accedo Compose, at the International Broadcasting Convention (IBC) next month, the company announced on Thursday.
Accedo Compose is set to launch next year and is designed to help streaming providers replace static customer journeys with continuously adaptive experiences. It uses Amazon Web Services (AWS) technology, including Amazon Sagemaker Unified Studio and Amazon Bedrock, to enable real-time testing and optimization across playback, messaging, monetization and engagement strategies.
Developed with cloud-native data and video specialists Merapar, the platform introduces modular “impact agents” that monitor viewer behavior, test interventions and autonomously evolve touchpoints based on business objectives.
Instead of relying on manual configurations and fragmented analytics, Compose orchestrates real-time adjustments aligned with goals such as reducing churn, maintaining quality or increasing customer lifetime value.
Examples include a churn-risk agent that identifies disengaged viewers and initiates retention flows, or a UX agent that adjusts layouts and messaging based on contextual signals like time of day. A monetization agent can deliver upsell offers when users are most likely to convert. Each of these agents can work independently or in sequence, allowing for greater flexibility than static personalization models.
“As streaming becomes more competitive, media companies need to make every product decision count,” Accedo Group CEO Michael Lantz said in a statement on Thursday. “Compose closes the loop between strategic intent and user-level execution. It enables streaming services to adapt in real time, with intelligence baked into the experience itself.”
Accedo Compose integrates with Accedo Assemble, the company’s modular OTT platform, and can be deployed as a standalone solution or within Accedo’s managed service framework. Accedo said additional impact agents will be introduced over time as new optimization opportunities emerge through data analysis.