OpenAI DevDay 2026: what the new ChatGPT features mean for small businesses

Published: 2 October 2026. Availability reflects the documentation checked on this date.

You finish a client meeting with useful notes. Then you still have to write the follow-up email, update the project plan and make sure someone acts on what was agreed. ChatGPT can help with each step, but you may still be moving information between tools yourself.

OpenAI's DevDay announcements on 29 September 2026 address more of that work. Dots, shared workspaces and new automation capabilities are intended to help AI carry tasks forward using connected business information.

For small businesses, the opportunity is less time spent organising and repeating work. Whether that becomes a worthwhile saving depends on access, setup and how much checking the output needs.

What did OpenAI announce at DevDay 2026?

The most relevant announcements for small businesses were Dots, ChatGPT Space and Pages, a Meetings plugin, recurring Team Tasks, and GPT-6.1 Sol. Developer updates also expanded the options for building agents and connecting tools inside ChatGPT.

These products have different purposes. Dots are ongoing assistants. Space provides shared context for collaboration. Meetings captures notes and action items. GPT-6.1 Sol is a model available through ChatGPT Work, Codex and the API, rather than a separate business application.

OpenAI's DevDay recap sets out the announcements and their availability. Several features are rolling out gradually, while others remain previews or are scheduled for later release.

Dots: AI assistants that can carry work forward

Dots are persistent agents powered by GPT-6 Astra. They have their own cloud computer and can use apps you connect. You can give them goals, provide feedback and inspect their work, without having to start a fresh conversation for every task.

In its 29 September launch announcement, OpenAI describes an early tester whose Dot noticed he had forgotten to invoice a publication. It prepared the invoice and sent it after his approval.

That is a useful illustration for a freelancer or small agency. Missed administration can delay payment, and spotting an unfinished task may be as useful as generating the document itself. However, this is an anecdote reported by OpenAI, not independent evidence of reliable invoice management.

Other possible applications include preparing social posts from an approved interview transcript or revising draft launch materials when requirements change. These are workflows to test, rather than outcomes to assume.

There is also a distinction between background research and taking action. OpenAI says Dots' proactive research uses read-only tools. Permissions and review rules govern actions, and consequential work still needs checking.

Can UK businesses use Dots?

OpenAI's current setup guidance excludes the UK from the Pro rollout. It says Business Premium users can access Dots across supported ChatGPT regions. Enterprise access is a beta that administrators must enable.

UK Pro subscribers should therefore check eligibility before planning work around Dots. Ordinary Business access should not be confused with Business Premium, and rollout may take several days to reach an eligible account.

ChatGPT Space: less repeated briefing for small teams

If several colleagues use AI separately, they can end up maintaining different versions of the same context. One has the client brief, another has the latest feedback, and someone else knows which deadline changed.

ChatGPT Space provides a shared environment where teammates and AI can work with common information. Pages are collaborative documents where people can edit, comment and ask AI to help with the next step. Space is available on Pro, Business and Enterprise plans.

Consider a small team preparing a campaign. A shared page could hold the agreed brief and receive feedback as the work develops. The potential benefit is reducing how often people reconstruct the background before asking for help.

It still needs an owner. Someone must resolve contradictory instructions and confirm which decisions are final. Sharing outdated information more efficiently will not improve the result.

Collaborative slides were announced for the coming weeks. Businesses should distinguish features they can use now from those still promised.

Meetings and Team Tasks: practical opportunities to reduce admin

The Meetings plugin creates summaries and action items in ChatGPT Space. It is currently a beta in the macOS desktop app for Pro and Business users.

For a consultant or service business, a sensible first use is a client follow-up. Review the meeting summary, ask ChatGPT to draft the email, then check names, commitments and dates before sending it. This is a bounded task with an output you can inspect easily.

Team Tasks support recurring work, while plugin events can trigger automations when something happens in a connected app. A weekly project update is a plausible use: gather information from approved tools, prepare a summary and flag items needing attention.

Each workflow needs clear instructions about what information to use and what happens if that information is missing. Start with preparation and review before granting permission to change records or contact customers.

This is where practical AI automation needs to fit your existing process. Decide what should happen at each step, including when the system should stop and ask someone.

GPT-6.1 Sol: what lower model costs mean

OpenAI released GPT-6.1 Sol with standard API input and output prices at one-fifth of GPT-6 Astra's. OpenAI reports performance approaching Astra on several evaluations, including coding and professional work.

That may matter if you commission a custom workflow that processes documents repeatedly. Lower model charges can improve its economics, provided the model completes your tasks adequately.

It does not mean your ChatGPT subscription becomes 80% cheaper, or your whole automation costs 80% less. Development, integrations and checking still contribute to the bill. Longer tasks and retries also affect usage.

At launch, Sol was available in ChatGPT Work, Codex and the API, but not ordinary Chat. For most owners, the useful decision is whether a workflow produces an acceptable result at an acceptable total cost.

What evidence supports business benefits?

The announcements are too recent to establish sustained results for UK small businesses. Existing research supports testing AI on specific tasks, with care over how widely the findings apply.

The ONS reported in July 2026 that 28% of businesses with zero to nine employees used at least one AI technology, compared with 49% of businesses with 250 or more employees. This measures broad adoption, not successful ChatGPT automation.

A peer-reviewed customer-support study, published in the May 2025 issue of the Quarterly Journal of Economics, found AI assistance increased issues resolved per hour by 15% across 5,172 agents. It studied one large business and an earlier AI tool. It is evidence that well-chosen assistance can help, rather than a forecast for Dots.

The practical implication is to measure your own workflow. A feature demonstration cannot tell you how often your business will need to correct the output.

How should a small business get started?

Choose a recurring task that consumes noticeable time and has a clear finish. Drafting meeting follow-ups is a better first experiment than asking an agent to manage all customer relationships.

Record how long the task takes today. Then test the AI-assisted version on a small batch, counting setup, review and corrections. Check whether it saves time while preserving the standard your customers expect.

Give the system only the access the task needs. Keep approval before external messages or changes to important records until you have evidence that the workflow is dependable.

If choosing the task is proving difficult, Glow's guide to getting started with AI without the hype provides a useful starting point. Business AI consultancy can help you assess a process before committing to a build.

Frequently asked questions

Do I need a developer to use the new ChatGPT features?

Features such as Space and Meetings are designed for direct use. Custom integrations and API-based agents may require technical support. Your needs depend on the tools and actions involved.

Will these features replace staff?

The evidence does not establish that conclusion. ONS reports that AI adoption has not yet produced widespread changes in workforce size. Start by assessing which tasks staff could spend less time on.

What should I try first?

Try one draft-and-review workflow, such as a meeting follow-up, using features available on your account. Expand only when the saving survives the time spent checking and correcting the result.