Adobe moved Photoshop into Gemini and Acrobat into Claude. The app is becoming a tool call.

On September 24, 2026, Adobe announced two expansions of its push into AI assistants: creative tools from Photoshop, Lightroom, Adobe Express and Firefly are now available inside Google Gemini, and Acrobat tools were added to Adobe's existing Claude plugin.

Adobe says the goal is to let users complete creative and document tasks conversationally without switching apps. You describe the outcome; Adobe picks the tools and runs the steps. Everything in this article traces to Adobe's own announcement, which is the only primary source available. No independent test of either integration was located, and we ran none.

What was announced

Adobe in Gemini covers Photoshop, Lightroom, Adobe Express and Firefly. Adobe says it is rolling out globally from September 24 and is available across all Gemini plans. Setup runs through Gemini's Personal Intelligence settings: select Adobe under Connected Apps, then sign in with an Adobe account or create one for free.

Adobe for Claude is the bigger change. Acrobat tools join the existing creative tools in one unified plugin, which Adobe says now spans more than 80 tools across Acrobat, Express, Photoshop, Illustrator, Premiere, Lightroom, InDesign and Adobe Stock. Two additions matter more than the tool count: a new interactive PDF editor inside Claude (reorder pages, edit text, add comments and highlights) and a Document Review skill that analyzes a document and annotates it. Express also gets a layer-based interactive editor for editing individual design elements without regenerating the whole asset.

Claude users can start as a guest or sign in. Adobe says signing in unlocks "the full plugin experience, including higher usage limits, additional tools and continuity across sessions." The update is rolling out on Claude and Claude Code across desktop, mobile and web, and the previous connector updates automatically.

This continues a strategy Adobe has run all year. Its creative tools came to Claude earlier in 2026, and Adobe lists ChatGPT, Slack and Copilot as existing integrations. Gemini is the new platform; Claude is the deepening one.

Why it matters

The unit of software is shifting. The app used to be the unit: you opened Photoshop, learned its interface, and did the work. The connector model makes the tool call the unit: you describe an outcome in a chat window, and the assistant decides which Adobe functions run in the background.

For operators, that changes where workflows live. High-volume, low-stakes work (batch photo touch-ups, template customization, PDF triage and redaction) can plausibly move into the assistant. The assistant becomes the intake layer; the app becomes the place you go for final control.

The interactive editors are the most honest part of the announcement. Pure conversation is a poor interface for precision work, and Adobe knows it: it had to build hands-on editing back into the chat experience. The model that emerges is conversation for intent, direct manipulation for refinement. That is a tacit admission that "just describe what you want" was never the whole story.

What Adobe gains

Distribution inside chat surfaces, and an account for every user who connects one. Gemini requires an Adobe account; Claude's full experience requires sign-in. Every connection converts an assistant user into an Adobe account holder, including many who would never have installed Creative Cloud. Adobe cites Gemini's "more than one billion monthly users" as the prize. That figure is Google's, repeated by Adobe, and should be read as marketing.

Adobe also gets positioning where creative work increasingly starts. If people begin projects in a chatbot, the company whose tools are callable there has an advantage over one that requires an app launch. The defensive reading is straightforward: this is distribution insurance.

What the announcement does not answer

Three gaps matter for anyone deciding to adopt or build on these integrations.

First, the data question. Your images and PDFs pass through the assistant and Adobe's services. Adobe's announcement says nothing about how files are processed, where they are retained, or for how long. It also says nothing about which assistant-side privacy controls apply. Before routing sensitive documents through these connectors, you have to answer that yourself from both providers' policies. The announcement leaves it open.

Second, the limits question. Adobe mentions "higher usage limits" for signed-in Claude users but gives no numbers. It does not say which tools require a paid Creative Cloud or Acrobat plan versus a free Adobe account. The pricing boundary between free account, sign-in tier, and paid plan is unspecified.

Third, the evidence question. The capabilities are described, not independently demonstrated. "Adobe orchestrates the right tools and steps to get you there" is a vendor claim about an agentic routing layer you cannot inspect. The use cases (product photo collections, seasonal campaign restyles, the photographer and creative director examples) are Adobe's own illustrations.

Adobe itself draws the boundary. Its announcement says to continue in the flagship apps "when your work calls for even greater depth and precision." The sensible read: first drafts and volume work in the assistant, final control in the app.

What remains unknown

• Usage limits for guests versus signed-in users, and what happens when you hit them. Adobe named the tier difference but published no numbers.
• Which tools require a paid plan. A free Adobe account gets you in the door; the paywall map is undisclosed.
• Data processing and retention terms for files handled through the connectors, on both the assistant side and Adobe's side.
• Real-world reliability and output quality. This analysis rests on Adobe's announcement; no independent test exists yet, and we ran none.