Adobe is expanding its artificial intelligence footprint across its major product ecosystem, including Photoshop, Premiere, Illustrator, and InDesign, during a prominent feature announcement at its annual MAX event in Miami. The company is actively scaling up the AI capabilities within its creative design tool, Firefly, while simultaneously launching a Model Context Protocol that enables external chatbots like ChatGPT, Claude, and Gemini to interact directly with Adobe features.
Builder Prototyping Tool Challenges Figma
The Firefly application is venturing into territory traditionally dominated by Figma by introducing a new prototyping tool currently in private beta named Builder. This tool allows users to rapidly assemble and showcase design concepts. Additionally, the update brings a dedicated feature for generating on-brand graphs alongside an advanced image-editing function that allows users to modify specific sections of an image without altering neighboring objects or elements.
Enhanced Video Editing and Proactive Agents
A newly integrated AI-driven video tool within Firefly enables users to construct complete video sequences using raw clips alongside auxiliary elements like still images and sound effects. Furthermore, the company is refining its virtual assistant to be proactive and continuously active, allowing it to gather deeper context regarding ongoing projects and remain accessible for any required action. Routine editing tasks, such as background removal or automated video thumbnail generation, are also being streamlined for faster execution.
App-Specific Upgrades Across the Creative Suite
Adobe has already integrated a wide array of AI functionalities into applications such as Photoshop, Express, Premiere, and Illustrator. Continuing this momentum at the MAX event, Photoshop is receiving a novel relighting tool alongside dedicated sliders for prompt-based generative styles. Premiere is gaining a professional-grade color grading system and a video object removal feature, while Lightroom is introducing tools to correct blinking eyes and eliminate unwanted blemishes.
Third-Party Models and Recent Acquisitions
The platform already supports numerous third-party models originating from developers like Google, Runway, ElevenLabs, Black Forest Labs, OpenAI, Kling, and ByteDance. Capitalizing on recent acquisitions such as Topaz Labs, the company is rolling out two specialized image models named Wonder 4, designed for upscaling and denoising, and ImageHDR, which converts standard photographic images into high dynamic range formats.
Local MCP and Developer Flexibility
While existing integrations with ChatGPT and Claude already allow users to access select product functions, Adobe is now deploying a local Model Context Protocol to expose a greater breadth of its Creative Cloud application features. Adobe vice president Deepa Subramaniam noted in an interview that the firm is currently evaluating which capabilities to expose, choosing to rely heavily on ongoing user feedback.
Subramaniam explained that the company is still determining where and when to establish boundaries by releasing multiple endpoints and capabilities across Illustrator, Photoshop, After Effects, and Premiere. She emphasized that observation of user behavior will guide future rollouts without imposing unnecessary restrictions on the MCP endpoints, ensuring that creative workflows are fully supported regardless of the chosen method.
In contrast, during an event held in India last month, Figma co-founder Dylan Field expressed skepticism, stating that MCP is an imperfect protocol that primarily benefits frontier AI laboratories.
