OpenAI Collaboration Tools: ChatGPT Slides and Decisions API

ChatGPT Slides can begin with a conversation, uploaded files, an outline or a template. Users can also provide an example presentation to guide the deck’s style and layout. That makes it possible to use ChatGPT to turn existing work material into a starting presentation rather than building every slide manually.

The collaborative layer is designed for the revision stage, when colleagues often need to align on wording, charts, visuals and the order of ideas. OpenAI collaboration tools can bring those revisions into the presentation itself, keeping feedback close to the content being discussed.

OpenAI says users can present directly from ChatGPT. On the web, decks can also be exported to PowerPoint, PDF or Google Slides. Exporting lets teams continue working in familiar software, though they should inspect the exported version to confirm that formatting and content have carried over correctly.

The product’s official feature page describes shared editing, feedback and AI-assisted updates as core parts of the experience. The aim is to reduce the manual work between producing a first draft and getting a presentation ready for an audience.

Who Can Access Collaborative ChatGPT Slides?

OpenAI’s help page says collaborative Slides is rolling out to eligible ChatGPT Pro, Business and Enterprise users with Space access. Availability may vary by account, workspace configuration and platform, so the feature may not appear for every user at once.

Users can create a presentation through ChatGPT Work and open it in Space, or start a deck directly in Space using a blank slide or template. They can then share it with teammates and grant editing access.

There is also a platform caveat for some Windows desktop environments. OpenAI says that when the required component is unavailable, local slide creation and editing in the desktop app may use PowerPoint files rather than the collaborative Slides experience. The web version or cloud mode can provide another way to work with collaborative decks.

For organisations evaluating OpenAI collaboration tools, the rollout details matter. Teams should confirm their plan, Space access, permissions and supported environment before relying on shared editing for a deadline-sensitive presentation workflow.

For businesses considering OpenAI collaboration tools, access is only the first step. Teams should establish who can create presentations, who can edit shared slides and who has final approval. These simple rules can help OpenAI collaboration tools fit into existing review processes without creating confusion over ownership or conflicting edits. Before a wider rollout, organisations should also test sharing permissions and confirm that colleagues can access the right workspace.

Decisions API Gives Developers Structured AI Choices

Alongside Slides, OpenAI introduced the Decisions API, a developer-facing feature announced in its DevDay 2026 recap. It is designed for situations where an application needs an AI model to choose from a defined set of answers rather than generate an unrestricted response.

Developers supply context using text or images and specify the questions and finite answer options. OpenAI says the resulting decisions can be used to classify content, route requests or determine an agent’s next action.

For example, a customer-support application could use a model to classify an incoming message into a set of support categories, then direct it to the appropriate team. An AI agent might use a similar decision to choose which tool or workflow to invoke next.

This is a different kind of API interaction from asking a chatbot to explain a topic. The application defines the choices in advance, and the model evaluates the supplied context against those options.

OpenAI collaboration tools now span both human-facing work and developer workflows: Slides helps colleagues edit shared presentations, while the Decisions API is intended to connect AI judgments to software actions.

One advantage of OpenAI collaboration tools is their potential to support both collaborative office work and software development. The Decisions API gives developers a way to connect predefined model choices with application logic. In practical terms, OpenAI collaboration tools can support a workflow in which employees collaborate on information while software uses structured outputs to classify requests, select actions or route tasks. Developers should still validate model responses and define fallback behaviour.

What the Decisions API Preview Means for Developers

OpenAI’s DevDay recap described the Decisions API as available in limited preview, with a broader release planned in the following days. That announcement provides the initial access status; developers should consult current OpenAI API documentation for availability, supported models, limits and implementation requirements before building a production dependency.

The finite-answer design may make model outputs easier to integrate into conventional application logic. It can be useful when software needs a classification, a routing result or a selection among actions.

However, structured output does not guarantee that the choice is correct. Developers still need to design appropriate answer options, provide relevant context, evaluate performance and decide what the application should do when the model is uncertain or encounters an unusual input.

The API also does not replace an application’s broader rules, workflow engine or safety controls. If a decision could affect a customer, employee or sensitive process, teams should determine when human review is necessary and how errors will be detected and corrected.

When assessing OpenAI collaboration tools, developers should treat the Decisions API as an integration component—not a complete autonomous decision system.

Developers evaluating OpenAI collaboration tools should begin with a narrow use case and a clearly defined set of possible outcomes. Test results against representative examples, including ambiguous inputs and unusual cases, before connecting the API to important workflows. OpenAI collaboration tools may simplify parts of an AI-assisted process, but the surrounding application remains responsible for permissions, validation, error handling and appropriate human review. This testing is especially important when a model’s selection can affect customers or business operations.

OpenAI Is Building a Shared Workspace Around ChatGPT

The two announcements fit into OpenAI’s wider effort to make ChatGPT useful for shared work, not only individual conversations. At DevDay, the company also highlighted ChatGPT Space, Pages and team-oriented agents—features intended to bring colleagues, files and AI assistance into a common environment.

Presentations are a natural place to introduce collaboration because they frequently require input from multiple people. A team may need to combine research, metrics, design and executive feedback before a deck is ready. OpenAI collaboration tools could reduce the friction of moving between drafting, reviewing and revising those materials.

The Decisions API addresses a different but related challenge: business software often needs AI outputs that fit a specific operational path. A model that selects among predefined responses can be easier to connect to an application than a long, free-form explanation.

That does not mean every team should move its work into ChatGPT or automate every decision. Organisations still need to consider data access, permissions, governance, quality assurance and the cost of maintaining AI-assisted workflows.

The broader direction of OpenAI collaboration tools is to bring AI closer to everyday work rather than leaving it in a separate chat window. Shared presentations can help colleagues develop and review ideas together, while structured APIs can connect model outputs to software workflows. For organisations, OpenAI collaboration tools will be most valuable when they reduce duplicated effort, make responsibilities clearer and work reliably with existing systems. Their practical impact will depend on how well teams adopt them.

What the Updates Mean for Businesses

ChatGPT Slides offers teams a shared place to create and revise presentations, with comments, simultaneous editing and AI-assisted changes. The Decisions API gives developers a constrained way to use model judgments in applications.

The common theme is practical integration. OpenAI wants AI to participate in the work itself—helping people refine shared material and helping software make structured choices. OpenAI collaboration tools may reduce repetitive handoffs, but their value will depend on accuracy, usability and how well they fit existing processes.

The features also have different maturity and access conditions. Collaborative Slides is rolling out to eligible paid users with Space access, while the Decisions API was announced in limited preview. Businesses and developers should check the latest official documentation, test the features against their needs and keep human oversight where mistakes could have meaningful consequences.

Before adopting OpenAI collaboration tools across a department, businesses should run a pilot and measure the results. Track presentation preparation time, the number of revision rounds, export quality and user feedback. For API-based workflows, measure classification accuracy, failure rates and the amount of manual intervention required. These checks help businesses decide whether OpenAI collaboration tools deliver measurable improvements rather than simply adding another platform to manage. A successful rollout should also include staff guidance, access controls and a clear process for reporting problems.