Adobe Is Giving Its AI Agents Even More to Do

Would you use an always-on AI agent such as Claude Duet or OpenAI’s Dot to edit your photos? At Adobe Max on Tuesday, the company announced it’s expanding its local MCP integrations with more third-party companies to enable more autonomous creative work.

MCP, or Model Context Protocol, is an open-source protocol for connecting AI systems to external apps and databases. AI agents and MCPs act as a two-way street. Agents, or fancy bots, can complete tasks independently.

For example, if you wanted your OpenAI agents (or Dots, whatever you call them) to edit a project in Photoshop, the MCP pathway would allow ChatGPT to run agents to edit the design in the Photoshop beta desktop app — no need to search through the toolbar if you’re not familiar with it.

More from CNET

The new MCP integrations are in preview now. The idea is to give people more choice by letting them use their preferred AI app to help them work in select Creative Cloud apps.

Over the past year, Adobe has been working with AI companies to bring its creative knowledge and tech to ChatGPT, Claude, Gemini and Copilot as Adobe plugins. It’s also brought external models from AI companies into Firefly, Adobe’s AI hub. Now, Adobe is deepening its relationships with third-party AI vendors to adapt to the agentic AI era.

Bringing your own agents offers a degree of user control, according to Shilpa Anand, vice president of global strategic partnerships. “These new models allow me to actually do things faster, experience the full power of the application and really make it more accessible to a broader set of users who are not experts in our workflow,” Anand told me.

Adobe’s agentic AI journey

Agentic AI has been a huge focus for tech companies this year. It’s the next step for companies that want to move beyond mere chat and information-gathering AI capabilities.

Adobe has been driving toward this vision for a while. It released AI assistants in Photoshop and Adobe Express that set the foundation for June’s official launch of Adobe creative agents.

But AI agents from frontier labs such as OpenAI and Anthropic caused trouble this summer when they escaped testing environments and hacked real websites. Anand told me that safety and security are “table stakes” for Adobe and that there haven’t been any issues aligning with other AI companies on safety precautions.

When considering partnerships, the company also considers which AI models will be most useful for Adobe’s creative customers, which is why Firefly includes models from big players including Google, as well as more specialized AI companies, such as ElevenLabs for AI audio tech. (Adobe acquired Topaz Labs, which makes AI photo editing software, earlier this year.)

Those creative AI agents are getting new capabilities across Creative Cloud apps, too. In Firefly, the company’s AI hub, its AI assistant should now be better at multistep or more complex tasks, and it has a new “plan mode” in preview that lets you review the agent’s plan before it starts running.

New “always-on” Firefly AI agents do exactly what the name implies: They’re always working and can handle predefined tasks, such as assessing image quality in new asset uploads. They’re in preview now, too. And across Illustrator, InDesign, Premiere, After Effects and Frame.io, those AI assistants are getting upgraded capabilities as well.

Adobe is also expanding its partnerships. Microsoft Photos will have a built-in Adobe Express integration, and Microsoft’s Arm-based Windows PCs are getting “optimized” Adobe apps. Disney is also deepening its relationship with Adobe, bringing Adobe’s Firefly AI tech to its creative teams.

Partnerships with AI companies and the integration of agentic capabilities will continue to be a focus for Adobe next year, Anand told me. “Agentic software is really going to enable us to reach more users and empower everyone to create, which is our mission.”

Reporter 2

Katelyn is a reporter with CNET covering artificial intelligence, including chatbots, image and video generators. Her work explores how new AI technology is infiltrating our lives, shaping the content we consume on social media and affecting the people behind the screens. She graduated from the University of North Carolina at Chapel Hill with a degree in media and journalism. You can reach her at kchedraoui@cnet.com. See full bio