AI Breaking News is an AI-generated alert, curated and reviewed by the Kursol team. When major AI developments happen, we break down what it means for your business.
OpenAI announced Dots at DevDay on September 29, 2026, and began rolling them out to Pro and Business Premium users on October 1. Dots are always-on agents powered by GPT-6 Astra that run on OpenAI's cloud infrastructure, can connect to over 4,000 applications, and operate continuously in the background pursuing user-defined goals with minimal oversight. Available through ChatGPT, Slack, and Microsoft Teams, Dots represent a fundamental shift in how companies deploy AI agents—moving from infrastructure-heavy self-hosted solutions to managed, cloud-based automation that any OpenAI subscriber can activate immediately.
Why OpenAI Built Dots
The traditional way to deploy an AI agent requires your team to run its own servers, manage the handoffs between different parts of the system, keep track of what the agent is doing, and build custom connections to the other tools it needs. OpenAI's approach eliminates that overhead. Each Dot gets its own cloud compute, runs continuously, and maintains conversational context across hours or days. When a user defines a goal—"manage my calendar and flag conflicts with my top clients"—the Dot operates autonomously, checking the calendar, cross-referencing client lists, and notifying the user when conflicts emerge. No custom code required to connect your apps. No separate database to track what the agent is doing. No deployment on your own hardware.
This is distinct from ChatGPT's existing capabilities. ChatGPT requires a human to send a message and wait for a response; Dots work in the background, initiating actions without prompting. They can delegate tasks to sub-agents, handle recurring work, and follow up across days. For context: at DevDay, OpenAI demoed a Dot scheduling a meeting by sending calendar invites, receiving refusals, and proposing alternative times—all without human intervention between the initial request and the final confirmation.
What This Changes for Your Team
Dots affect three categories of agent deployment:
1. Workflow automation that currently requires custom integration. If your team uses Zapier, Make, or custom-built webhooks to automate business processes (expense approvals, report generation, lead qualification), Dots can handle that natively. The Dot reads your Slack messages, understands context, takes action across connected apps, and reports back. No middleware required. For operations teams, this is the difference between "build an automation in two weeks" and "enable a Dot in one afternoon."
2. Personal AI assistants that are too heavy for native ChatGPT. Teams deploying OpenAI models internally often end up with custom orchestration layers to keep assistants stateful across sessions. Dots run on OpenAI's infrastructure, which means your team can ship always-on assistants without maintaining a backend. Slack integration arrives in October; email and messaging expansion follow. This is particularly valuable for distributed teams where a single Dot can coordinate across timezones—scheduling, summarising, flagging exceptions.
3. Enterprise AI productivity, with vendor lock-in implications. If your automation strategy centres on OpenAI—ChatGPT, GPT-6, fine-tuned models—Dots tighten that integration. You're not choosing a tool alongside OpenAI; you're extending OpenAI's platform deeper into your operations. That's a strategic decision, not a technical convenience.
What to Do This Week
If you're on OpenAI Business Premium, start with a single Dot and a bounded use case. Try automating one recurring task: weekly report assembly, expense categorisation, client follow-ups. Don't build an Execution Centre around Dots yet; test the behaviour and integration surface first. Does the Dot reliably understand your domain? Do integrations with your existing tools work? Can you rely on the Dot to act when you're not watching?
Document your governance assumptions. Dots have autonomy—they make decisions and take actions without approval. If your compliance posture requires audit trails, approvals, or change control, Dots demand operational rules you haven't written yet. Before deploying Dots in production, work through: What actions can a Dot take without escalation? What requires human sign-off? How do you audit Dot decisions? This is the kind of operational AI due diligence embedded engineers help with—understanding your governance model, testing agent behaviour against your risk profile, and building the monitoring and escalation that keep AI agents operating within your organisation's boundaries.
Check if your team has existing agent infrastructure. If you already run agents through custom-built code or specialised agent-hosting platforms, Dots offer a different trade-off: less control, less choice over where your data is stored, but no infrastructure to operate. Evaluate whether Dots can replace your existing stack for certain workloads, or whether your data governance requires self-hosting. Both are valid answers; the decision depends on your compliance posture and infrastructure constraints.
The Bottom Line
Dots eliminate the infrastructure barrier to agent automation. That changes who can deploy agents and how fast they can move. For companies on OpenAI, it removes an engineering decision—you no longer choose between "build agent infrastructure ourselves" or "use a third-party orchestration platform"; you just activate a Dot. That's a massive simplification. It also means your agent strategy is now inseparable from your OpenAI commitment, which is worth understanding before you've shipped production automation on Dots.
If you're rethinking your agent deployment strategy in response to this change, take our free AI readiness assessment to understand where your organisation stands.
AI Breaking News is Kursol's rapid analysis of major artificial intelligence developments — focused on what actually matters for your business. Subscribe to our RSS feed to stay informed.
FAQ
Function calling lets you define actions for a model to take within a single request-response cycle. Dots are stateful, long-running agents that operate across multiple sessions, maintain their own cloud compute, and can schedule actions to happen in the future. You might use function calling to build a single Dot, but Dots are the operational wrapper around that capability.
Dots rolled out first to Pro and Business Premium users. OpenAI has signalled broader rollout later this year, but timing is not confirmed. If you're on the ChatGPT free tier, you'll need to upgrade to access Dots.
Partially. Your Dots run on OpenAI's infrastructure and are powered by GPT-6 Astra. If you want to switch the underlying model to Anthropic, Mistral, or Google, you'd need to migrate your Dots to a different orchestration platform. This is worth evaluating as part of your vendor strategy: is the convenience of managed Dots worth the switching cost if OpenAI's model quality falls behind competitors?
Kursol