Technology

Adobe Expands AI Agent Capabilities and Third-Party Integrations

Adobe Expands AI Agent Capabilities and Third-Party Integrations

Adobe announced major expansions to its local Model Context Protocol integrations at the Adobe Max conference on Tuesday, allowing users to leverage third-party AI agents for autonomous creative work across select Creative Cloud applications.

Expanding AI Integrations and User Control

Model Context Protocol, known as MCP, is an open-source protocol designed to connect artificial intelligence systems with external applications and databases. This technology functions as a two-way street, enabling advanced bots or AI agents to complete tasks independently. Through this system, users can direct AI applications such as ChatGPT to run agents that edit designs directly in the Photoshop beta desktop app, eliminating the need to navigate complex toolbars manually.

These new MCP integrations are currently available in preview, offering users greater flexibility by allowing them to work within Creative Cloud apps using their preferred AI assistants. Over the past year, Adobe has integrated its creative technology into ChatGPT, Claude, Gemini, and Copilot as plugins, while also incorporating external models from various AI companies into its Firefly platform.

Shilpa Anand, vice president of global strategic partnerships, told CNET in an interview that bringing outside agents provides users with greater control. Anand stated that these new models enable people to work faster, experience the full power of applications, and make software more accessible to non-experts.

New Capabilities and Always-On Agents

Adobe has been steadily advancing toward agentic artificial intelligence, laying the groundwork with AI assistants in Photoshop and Adobe Express before officially launching creative agents in June. Anand told CNET that safety and security are fundamental requirements for Adobe, noting that the company has faced no issues aligning with external AI partners on safety precautions.

As part of the latest updates, the Firefly AI assistant is gaining the ability to handle more complex, multistep tasks and features a new “plan mode” in preview that allows users to review an agent’s plan before execution. Furthermore, new always-on Firefly agents can operate continuously to manage predefined tasks, such as assessing image quality in newly uploaded assets, and are also in preview. Upgraded AI assistants are additionally rolling out across Illustrator, InDesign, Premiere, After Effects, and Frame.io.

Adobe is also broadening its collaborative network by introducing a built-in Adobe Express integration for Microsoft Photos, optimizing Adobe apps for Microsoft’s Arm-based Windows PCs, and deepening a relationship with Disney to bring Firefly AI technology to Disney’s creative teams.

Anand told CNET that software partnerships and agentic capabilities will remain a primary focus for Adobe moving forward. Anand stated that agentic software will help the company reach a wider audience and fulfill its mission of empowering everyone to create.

You May Also Like

Leave a Comment

Your email address will not be published. Required fields are marked *