On the occasion of the Adobe Summit 2025, the American giant of San José unveiled a new strategy focused on agency. Customer experience, marketing … Artificial intelligence infused in all products.
After the generative AI with Firefly, place at the agentic AI. Adobe unveils this March 2025 Summit 2025 new solutions focused on autonomous artificial intelligence. More than just a chatbot, the agentic AI is capable of performing actions with different degrees of autonomy. Adobe notably announces a fundamental platform, Adobe Experience Platform Agent Orchestrator, which allows you to create and manage specialized AI agents. On this basis, the company deploys 10 operational agents covering content production, marketing and optimization of the customer experience. Finally, Adobe presents Brand Concierge, a complete application allowing brands to configure personalized conversational agents to interact directly with their customers.
AEP Agent Orchestrator: the suite dedicated to agents
This is the main novelty of the Adobe Summit 2025. Adobe Experience Platform Agent Orchestrator allows companies to create and manage IA agents that automate complex marketing tasks. The continuation is based on the Adobe customer data management platform, Experience Platform. Thanks to the deep understanding of the data of the platform, agents are able to understand data and act accordingly. The link with Experience Platform makes it an agental suite Privacy by design. All data governance policies are respected in Agent Orchestrateur. Practical.
The continuation is also quite open and compatible with its launch with the other agentic ia ecosystems of Microsoft, AWS, IBM, Genesys, SAP and Workday. More concretely, Adobe agents can communicate and exchange information with them. For example, a Genesys customer service agent could share relevant information with Adobe’s marketing agent to create a coherent customer experience. Welcome interoperability for companies that are often customers of several platforms for different segments.
10 ready -to -use agents
AEP Agent Orchestrator arrives at its launch with 10 operational agents. Each agent specializes in a precise function of the customer experience.
- Optimization site Allows you to analyze websites to detect and correct SEO and performance problems. The agent then automatically implements his corrections, directly from the Adobe ecosystem.
Happy production is used to assess digital assets, identify obsolete content, and can adapt, crop or automatically update visuals. - Day Automation The creation of multichannel customer journeys. Example: the agent generates an onboarding course for new customers with welcome email, then reminder SMS if the email is not open.
- Data Engineering Manages complex integration, cleaning and data security tasks. The agent cleans, transforms and automatically integrates customer data from different sources.
- Workflow Optimization intervenes directly in the work processes. The agent can, for example, automatically reassign late tasks or simplify approval circuits when a bottleneck is detected in a team workflow.
The other agents presented, such as audience, Data Insights, Product Advisor, Living in qualification or experience, are more similar to intelligent assistants enriched with RAG which analyze the data and formulate recommendations, but without real capacity for autonomous action.
Adobe Brand Concierge, a new customer experience based on AI
Brand Concierge asserts himself as the general public window of Adobe’s new agent strategy. Unlike conventional conversational agents limited to answer simple questions, Brand Concierge offers a complete multimodal experience (text, voice, images) capable of guiding consumers from the discovery phase to the purchase. Concretely, the tool allows brands to quickly deploy an intelligent conversational interface on their website or mobile application, capable of naturally interacting with customers while reflecting the visual identity and the tone of the brand.
“Instead of having a tree structure with a whole ramification to navigate and search for information, imagine that the first page is just a quick. And as requests are not a simple search engine, but a structured response with elements composed on the fly,” explains an Adobe spokesperson. The user can then either continue the conversation to refine their request, or click on the elements generated to switch to a more traditional navigation. This hybrid approach can seduce the retail and travel sectors, whose customers are already used to interacting with generative AI rather than with conventional search engines. Adobe claims to have anticipated the referencing issues that this new approach raises for the sites concerned.