Announced 10 Nov 2026 · Sources checked
What Adobe announced at Max 2026
Adobe held its annual Max creativity conference in Miami Beach on 10 November 2026, and the company used the event to make its most aggressive AI agent push yet.
The announcements span three major areas: opening Creative Cloud apps to third-party AI agents via the Model Context Protocol (MCP), previewing always-on AI agents that handle background creative work, and expanding Firefly into a full AI creative studio.
The company also announced a partnership with Google to bring Adobe AI tools into Gemini and Google Workspace.
The announcements were covered by TechCrunch (https://techcrunch.com/2026/11/10/adobe-adds-more-ai-to-its-products-and-launches-an-mcp/), The Verge (https://www.theverge.com/tech/1030593/adobe-max-ai-agents-mcp-firefly-assistant-updates) and Adobe's own newsroom (https://news.adobe.com/news/2026/11/adobe-firefly-expands) on 10 November 2026.
Adobe's vice president of Firefly and agentic AI, Forest Key, said in the press release that the company built Firefly to be the AI creative studio where the best models, the tools creators already trust and an agent that acts as a creative partner come together.
MCP integrations let third-party AI agents edit Creative Cloud files directly
The most technically significant announcement is Adobe's local Model Context Protocol (MCP) integration for Creative Cloud tools. MCP is an open protocol developed by Anthropic that lets AI agents interact with external tools and data sources.
Adobe's implementation, previewed at Max, exposes endpoints in Photoshop, Illustrator, After Effects and Premiere that agents from OpenAI's ChatGPT, Google's Gemini and Anthropic's Claude can call to make direct changes to native project files on a user's computer.
Adobe vice president Deepa Subramaniam told TechCrunch that the company is still figuring out what capabilities to expose through MCP and will rely on user feedback.
"We are still figuring out where and when and how to draw the line," Subramaniam said. "What we're putting out there is a number of endpoints and capabilities across Illustrator, Photoshop, After Effects, and Premiere.
We're just going to see how people use it."
The MCP integration is not available yet. Adobe says a private beta will launch in early 2027.
When it arrives, it could significantly change how creative professionals use AI agents, because rather than generating outputs that need to be imported manually, the agents will be able to work directly inside the professional tools designers already use.
Always-on agents handle background creative work
Adobe also previewed always-on agents that can run recurring creative tasks continuously in the background.
A user could assign an agent to organize incoming assets, run quality checks, or prepare content for different social media channels, and then review the agent's plan and approve the final outcome.
Adobe says this feature will come to Firefly AI Assistant first, though the company has not specified when.
The always-on agent concept extends the Adobe Creative Agent, which powers Firefly AI Assistant. Since its introduction earlier this year, the Creative Agent has been used for tasks like batch retouching portraits and converting product shots into short social videos.
New capabilities announced at Max include Plan Mode, which lets users review and approve the agent's plan before it starts working; a Model Selector that lets users choose which AI model to use for a project; and customization options for formats, colors, brand guidelines and characters that apply across every project.
The Creative Agent is also being added to Premiere's mobile editing app, allowing conversational video edits on the go.
The always-on agent preview is Adobe's answer to a question that has been hanging over the AI agent space: can agents do useful work without constant human supervision?
Adobe's approach is to let the agent run independently but require human approval at the final stage, similar to how [Figma's AI features](/blog/figma-specs-markdown-agent-context-code-layers-open-beta-config-india/) work in Config India and how other creative tools are experimenting with agentic workflows.
Firefly expands into an AI creative studio
Adobe is repositioning Firefly from an AI image generator into a full AI creative studio.
The new Firefly brings together Adobe's creative tools, third-party AI models from companies including OpenAI, Google, ByteDance, Kling, ElevenLabs and Black Forest Labs, and the Creative Agent as a unified workspace.
New capabilities include Firefly Design, an all-new design editing space that combines illustration, image editing, animation, 3D and print on one canvas; Firefly Video with expanded video editing features including a simplified timeline, caption generation and AI-generated transitions and b-roll; and Firefly Graph, a workflow builder that lets creators build repeatable multi-step processes using different AI models and Adobe tools.
Firefly Builder, in private beta, is a design-to-code prototyping tool that lets users transform ideas into interactive prototypes like websites or apps using natural language, with no code required for the initial prototype.
Users can then edit any element directly on the canvas while a design system keeps the project consistent.
This directly positions Firefly against Figma, whose co-founder Dylan Field recently said MCP is not a perfect standard and benefits frontier labs more.
Adobe also announced Quick Actions, a gallery of ready-to-use workflows from one-click fixes like Remove Object and Thumbnail Generator to guided projects like Build Product Ad.
New image models from Adobe's recent acquisition Topaz Labs, called Wonder 4 for upscaling and denoising and ImageHDR for HDR conversion, are also available in Firefly.
Adobe-Google partnership brings Firefly to Gemini
Alongside the Max announcements, Google announced a partnership with Adobe that integrates Adobe's creative and productivity tools into Google Workspace and Gemini.
Users will be able to edit PDFs in Drive using Adobe tools and access Adobe AI capabilities through Gemini.
The partnership is separate from the MCP integration and represents Adobe's strategy of making its tools available wherever users work, not just inside Creative Cloud apps.
Adobe has been building toward this agent-first strategy for months. In April 2026, Adobe's creativity head David Wadhwani said the Adobe Creative Agent would eventually surface in "virtually everything we develop."
The company has already added AI features across Photoshop, Express, Premiere and Illustrator throughout 2026, and in October it [expanded its partnership with NVIDIA](/blog/adobe-nvidia-expand-collaboration-rtx-spark-photoshop-premiere-frame-io-nemotron/) to optimize Creative Cloud apps for RTX Spark hardware.
The bigger picture: Adobe versus Figma in the AI era
Adobe's agentic AI push comes as competition with Figma intensifies. Figma's Config India event in October 2026 introduced Specs, a Markdown-based home for agent context, positioning Figma as the AI-native design tool.
Adobe's response at Max 2026 is to bet on open standards like MCP and on its existing user base of professional designers who are already invested in the Creative Cloud ecosystem.
The MCP strategy is notable because it makes Adobe's tools available to any AI agent, not just Adobe's own Firefly Assistant.
That could make Creative Cloud apps more attractive to developers building AI-powered design workflows, but it also means Adobe is competing on the quality of its tools rather than on lock-in.
The always-on agent preview suggests Adobe sees a future where creative work is increasingly delegated to AI agents that run in the background, with humans acting as editors and art directors rather than hands-on operators.
The announcements also signal that Adobe is betting heavily on agents as the next interface for creative software.
Rather than clicking through menus and toolbars, users describe what they want in natural language and an agent orchestrates the tools to deliver it.
Whether that vision matches how professional designers actually work will determine whether Adobe's agentic bet pays off.
Three takeaways
Adobe's MCP integration is the first major example of a creative software company opening its core tools to third-party AI agents.
If the private beta launches in early 2027 as planned, it could set a template for how professional creative tools work with AI agents.
Always-on agents represent Adobe's answer to the autonomy question.
Rather than full autonomy, Adobe's approach keeps humans in the loop as approvers while letting agents run in the background, which may be more practical for professional creative work where quality control matters.
Firefly's expansion into an AI creative studio positions Adobe to compete with Figma, Canva and other web-based design tools on their own terms, while retaining its existing professional user base through deep Creative Cloud integration and access to third-party AI models.
Common questions
When will Adobe's MCP integration be available?
Adobe says a private beta of the Creative Cloud local MCP integrations will launch in early 2027. The MCP integration lets third-party AI agents like ChatGPT, Claude and Gemini directly edit Photoshop, Illustrator, After Effects and Premiere files.
What is the difference between Firefly AI Assistant and Adobe's MCP integration?
Firefly AI Assistant is Adobe's own AI agent, powered by the Adobe Creative Agent. The MCP integration is an open protocol that lets third-party AI agents from other companies access Adobe's tools.
Users can choose either approach.
Sources & further reading
How this story was made
Written by Kristian Kostov with AI assistance and checked against the linked sources. Company performance claims are attributed to the company. Analysis reflects AiLookout’s interpretation; we have not independently tested the products discussed. The cover is an AI-generated editorial illustration, not a photograph or evidence of an actual event.
Our editorial standards
