Redis has announced a major expansion of its Iris platform, introducing six capabilities designed to help businesses move artificial intelligence agents from experimental projects into reliable, production-ready systems.
Announced on 4 November 2026, the latest development builds on Redis Iris, the company’s real-time context engine for AI agents, by introducing new technologies for memory, context management, routing, observability and communication.
The expansion aims to address some of the challenges organisations face when deploying autonomous AI agents in operational environments, including access to relevant information, complex system integrations and the costs associated with repeated AI processing.
Bringing AI agents into production
Although AI models and supporting infrastructure have advanced considerably, many organisations continue to experience difficulties deploying AI agents at scale. According to Redis, agents frequently lack access to the right information at the right time or struggle to retain relevant context between interactions. These limitations can result in unnecessary processing, repeated calls to large language models and complex integrations between different software systems.
Rowan Trollope, Chief Executive Officer of Redis, said “Developers shouldn’t need five infrastructure products to make agents work.”
The company believes that combining memory, context, routing and communication within a single platform can simplify the infrastructure required to develop and operate enterprise AI agents.
Six capabilities supporting enterprise AI
The expanded Redis Iris platform introduces six capabilities, which developers can use independently or combine depending on their requirements.
Redis Intent Router
Intent Router identifies the meaning behind a user’s request and directs it towards the most appropriate destination. This could include an existing answer, an AI model, a specialised agent or a human operator. The technology also supports model routing, guardrails and more flexible decision-making processes.
Redis Agent Playbooks
Agent Playbooks allows organisations to reuse established answers and procedures rather than repeatedly generating responses from scratch. The capability distinguishes between requests that can be answered using existing information and those requiring current data or fresh execution. Redis says this approach can reduce processing requirements, response times and AI token consumption.
Redis Agent Memory
Agent Memory enables AI systems to build and organise memories using information from user interactions, documents and other data sources. The technology incorporates inspection, auditing and data provenance controls, allowing organisations to understand how information has been collected and used. It also supports the removal of information that should not be retained, including personally identifiable data.
Redis Context Gateway
Context Gateway provides a governed access point connecting AI agents with external information sources and tools. Organisations can integrate their existing Model Context Protocol (MCP) servers and reuse previous results to improve consistency between interactions. The capability is intended to reduce repeated data retrieval and simplify integrations between AI agents and enterprise systems.
Redis Lens
Lens introduces observability capabilities designed to help developers understand how AI agents reach particular outcomes. It enables organisations to trace decisions back to the relevant procedures, routing choices, stored memories and original interactions. The technology also provides visibility into operational measures, including task success, token consumption and response latency.
Redis Agent Mesh
Agent Mesh supports communication and coordination between multiple AI agents. The technology uses messaging streams that can recover from connection interruptions, helping agents continue tasks and exchange information more reliably. It also supports delegation and handovers between agents operating across different workflows.
Availability and enterprise deployment
Four of the six capabilities — Intent Router, Agent Playbooks, Agent Memory and Context Gateway — are generally available. Redis Lens and Agent Mesh are currently available in beta.
Redis Iris originally launched in May 2026 and has already been adopted by organisations developing production AI applications. One example highlighted by Redis is Paradigm, which uses Redis technology to support the memory and orchestration requirements of its AI agents. The expansion reflects Redis’s wider strategy to support enterprise AI development by bringing essential infrastructure capabilities into an integrated platform.
Supporting the next generation of AI applications
Redis believes successful AI deployment requires more than access to advanced language models. Production systems must also manage unpredictable user requests, maintain appropriate context, coordinate interactions and provide reliable access to information.
The company has designed Iris around four principles: making context navigable, ensuring memory is selective and adaptable, improving processing efficiency and enabling systems to learn from previous interactions. By combining these capabilities, Redis aims to help organisations build AI agents that operate more reliably while reducing the complexity and cost of maintaining supporting infrastructure.
The latest expansion highlights the growing importance of dedicated software infrastructure as businesses move towards more widespread deployment of autonomous AI applications.
