AI agents are intelligent operators that can understand user intent, interpret goals, create plans, and take action — either independently or alongside people. Agents can work autonomously or semi-autonomously, using skills like planning or reasoning and interacting with tools. They also utilize memory, drawing on experience and behavioral data to enhance their performance.
Adobe Experience Platform Agents are a new class of enterprise intelligence designed to work alongside marketing and creative teams. Purpose-built to support customer experience orchestration, the agents are focused on automating high-value tasks, surfacing actionable insights, and creating capacity for individual contributors so marketers and creatives can spend more time on strategic work.
Agentic skills are the abilities that let an AI agent act on its own and make choices. These skills help the agent take purposeful actions that are guided by context and aimed at achieving specific objectives.
Adobe Experience Platform Agents can operate both autonomously and semi-autonomously. When proactively providing insights and recommendations, the agents act on their own. In other cases — such as when co-ideating with a user — they seek user input before taking further action, acting semi-autonomously. Regardless of the mode, Experience Platform Agents will always act within the guidelines and limitations set by the user and only at the user’s direction.
Adobe Experience Platform Agents can be accessed through a conversational interface — such as AI Assistant in Adobe Experience Platform — where a user can submit prompts in natural language to complete tasks. In the background, the reasoning engine steps in. It looks at the prompt, assesses what the task is, and creates a plan to complete the task. This plan may use one or more expert AI agents - specialized in one or more skills that are required to execute the steps involved in completing the task such as pulling data, generating insights, creating content and more. These agents also tap into the knowledge base to gather any information they need to complete an objective.
Throughout the process, Agent Orchestrator makes sure that the agents’ work can be checked, validated, and repeated so you get reliable results every time.
The reasoning engine is the “brain” behind Adobe Experience Platform Agents. It uses Large Language Models (LLMs) and a knowledge base to generate responses and carry out tasks. To train these LLMs, Adobe relies on synthetic data — like sample prompts created by Adobe’s product and engineering teams — so the models can learn effectively. Adobe does not use customer prompts, customer data, metadata, or outputs for training the LLMs. Plus, we ensure that third-party LLM providers are not allowed to use Adobe customer data for their own training or manual review.
Adobe Experience Platform Agents are created for the specific workflows and tasks needed to deliver exceptional customer experiences. This is why they are:
- Grounded in real-time customer experience data and content. Agents are powered by Adobe Experience Platform and its semantic understanding of journeys, content, behavior, and performance signals. Actions are context-aware and aligned with business goals.
- Designed for orchestration at scale. Agents do not operate in silos. Through Adobe Experience Platform Agent Orchestrator, they coordinate with other Adobe agents or agents from third parties across workflows, applications, and systems to deliver results in real time.
- Extensible by design. The agent framework supports a wide range of use cases. Customers, partners, and developers can build, configure, or integrate agents to match the specific needs of their teams and industries.
Organizations should manage AI agents’ security by grounding them in governed enterprise data, enforcing strict privacy and consent controls, and applying clear responsible AI governance practices.
Adobe Experience Platform Agents run on Adobe Experience Platform’s enterprise‑grade trust layer, with built‑in data access controls, consent and data‑usage management, and standardized review processes informed by Adobe’s responsible AI practices. This helps organizations deploy AI agents with confidence while protecting sensitive customer data, supporting regulatory compliance, and maintaining customer trust at scale.