Quick answer: OpenAI Dots are always-on AI agents inside ChatGPT that keep working after you close the conversation. Each dot belongs to one person and runs only on OpenAI’s GPT-6 Astra model. GoSearch takes the same proactive approach company-wide: it surfaces what needs attention for every employee, grounds answers in permission-aware company knowledge, and lets teams choose their AI model.
OpenAI announced Dots at DevDay on September 29, 2026. Dots are part of a wider shift in workplace AI, from tools that answer prompts to agents that act without one. People don’t want to open a chat, explain a project, paste in files, and repeat the setup tomorrow. They want AI that knows the work and moves it forward without a prompt.
Dots answer that need for individual users on premium ChatGPT plans. Bought one seat at a time, personal agents can also add to AI sprawl. Enterprise teams need something broader: proactive AI that every employee can use, built on company context with the same access rules as the systems it comes from. GoSearch closes that gap today.
What Are OpenAI Dots?
OpenAI Dots are persistent ChatGPT agents that run in the background on GPT-6 Astra. Each dot gets its own cloud computer and browser, and users can reach it through ChatGPT, Slack, or Microsoft Teams.
The main change from a regular chat is continuity. A dot carries context from conversations, ChatGPT memories, and connected apps, so it can work toward ongoing goals between sessions. OpenAI also designed Dots to notice when something needs doing and bring it to the user, rather than waiting for a prompt.
Access is limited at launch. Dots are rolling out to ChatGPT Pro and Business Premium users in eligible markets, with a beta for eligible Enterprise workspaces. Free, Go, and Plus users don’t have access.
OpenAI has also outlined what comes next: specialist dots with organizational responsibilities, teams of dots, and paid options to add capacity.
OpenAI Dots vs. GoSearch: Personal Agent or Company Context Layer
Dots and GoSearch both move AI from answering prompts to keeping work moving. The difference is scope: a dot works for one person, while GoSearch works across the organization.
| OpenAI Dots | GoSearch | |
|---|---|---|
| Built for | One user’s ongoing goals and tasks | Every employee, plus shared team agents and workflows |
| Availability | ChatGPT Pro ($100/month, includes one dot) and Business Premium, plus an Enterprise beta. Pro isn’t available in the EEA, Switzerland, or the UK. | Free, Pro, and Enterprise plans |
| Proactive features | Background research and task follow-through for the user who owns the dot | Proactive Intelligence pulse cards for every user |
| Where it reaches | 4,000+ third-party apps, such as Google Drive and GitHub, through plugins each user connects in ChatGPT | 100+ connectors built around a shared, permission-aware company index |
| Permissions | Shares app permissions with ChatGPT, Work, and Codex. Background research is read-only. | Answers respect each source system’s access rules |
| AI models | GPT-6 Astra only | OpenAI, Anthropic, Google, and xAI models |
| Usage costs | Chats don’t count toward limits. Deeper work and Codex or Work tasks draw on usage allowances. Long-term pricing is pending. | Included in plan pricing |
| Governance | Microsoft Agent 365 integration announced, not yet available | SOC 2 Type II, GDPR, HIPAA, and CCPA compliance; zero data retention with LLM providers; bring-your-own-cloud option |
Dots details reflect OpenAI’s launch announcements as of October 2026.
GoSearch Proactive Intelligence Brings the Work to You
OpenAI pitches Dots as AI that spots work before you ask. GoSearch already does this for every employee with Proactive Intelligence, a set of daily AI pulse cards on the GoSearch homepage:
- Morning Brief: Items that need a decision or reply, plus decisions and shipped work on your projects
- Waiting for Me: Open requests, @mentions, approvals, and tickets assigned to you
- Meeting Prep: Upcoming meetings with related emails, messages, docs, and prior notes
- Slipping Through the Cracks: Commitments you made and haven’t finished yet
Each card draws on every connected app, not just one inbox and calendar. Users can refresh any card when their day changes, or widen its view to the week ahead or the past two weeks. Cards respect source permissions, so each person sees only what they can already access.
Personal Memory Isn’t Company Context
A dot knows only what one user tells it and what’s in that user’s apps. Enterprise work runs on knowledge spread across the company, often in systems that user never opens: decisions in Slack, customer history in Salesforce, implementation details in Jira, and policies in Confluence.
GoSearch connects those systems into one searchable layer. Employees ask questions in natural language, and GoAI, the GoSearch AI assistant, returns answers with links to the sources behind them. For example:
- “What did we decide on enterprise pricing, and who signed off?”
- “Which Jira issues are blocking the Q4 launch?”
- “What did our account team discuss with Acme last quarter?”
How GoSearch Puts Company Context to Work
Custom AI Agents for Every Team
Any team can use the GoSearch no-code agent builder to create an AI agent for a recurring task, grounded in that team’s own sources. An IT team can build a support agent that searches approved runbooks and ticket history. An HR team can build a policy assistant that answers from current benefits documents. Owners set each agent’s instructions, data sources, and access, so a department gets one consistent assistant instead of every employee writing the same prompt.
GoAI Artifacts Turn Answers Into Deliverables
GoAI artifacts produce documents, spreadsheets, PDFs, and code from connected company data. Users review each artifact and its sources in an output panel before they share it. A sales rep can build an account brief from CRM records and call notes, and an engineering lead can turn Jira and GitHub activity into a release summary.
GoSearch Workflows Automate Recurring Tasks
GoSearch Workflows run AI tasks once or on a schedule. Each workflow pulls from connected apps, uploaded files, web search, or agents, then delivers the result by email or Slack. Common uses include weekly Gong call summaries, competitive intelligence reports, and engineering release updates.
A dot works on behalf of one person. A workflow runs on a schedule and delivers results to a whole team.
Use Cases by Team
- Sales: Pull up deal history, customer objections, and past call notes in one account brief before a customer call
- Engineering: Find architecture docs, past incidents, and design decisions made in Slack when debugging an issue or getting up to speed on a codebase
- IT and Support: Pull up runbooks, similar past tickets, and recent product changes before replying to an employee or customer
- HR: Let employees find policy and benefits answers on their own, drawn only from HR-approved documents
ChatGPT Dots may be a sign of where personal AI workspaces are heading. GoSearch is built for where enterprise AI needs to go next: from finding knowledge to using it.
Get GoSearch Free or schedule a demo.
Search across all your apps for instant AI answers with GoSearch
Schedule a demo
Frequently Asked Questions
Is ChatGPT Dots publicly available?
No. As of October 1, 2026, ChatGPT Dots is not publicly available, so its final features, pricing, security controls, and release timeline have not been confirmed.
What is ChatGPT Dots expected to do?
Public information is limited because the product has not been released. The discussion around Dots suggests a focus on persistent AI workspaces, project context, and a more continuous relationship between AI conversations and ongoing work.
How is GoSearch similar to ChatGPT Dots?
GoSearch provides persistent, context-rich AI assistance through connected company knowledge, custom AI agents, structured artifacts, and recurring workflows. It is designed for shared enterprise use rather than only individual project organization.
How is GoSearch different from ChatGPT Dots?
GoSearch focuses on enterprise search and execution. It connects to more than 100 workplace applications, respects source permissions, supports multiple LLM providers, and lets teams create agents and workflows across business systems.
Can GoSearch create custom AI assistants?
Yes. GoSearch includes a no-code AI agent builder that lets teams create specialized assistants for functions such as HR, IT, sales, customer support, engineering, product, and marketing.
Can GoSearch automate recurring work?
Yes. GoSearch Workflows let users schedule AI-powered processes that pull information from connected applications, analyze it, and deliver results through channels such as email or Slack.
Does GoSearch support more than one AI model?
Yes. GoSearch is model-agnostic and supports leading models from multiple providers, including OpenAI, Anthropic, and Google. Teams can choose models based on performance, cost, reasoning, multimodal capabilities, or governance requirements.
Is GoSearch secure for enterprise use?
GoSearch is designed for enterprise security and permission-aware access. It supports source-level permissions, non-indexing personal connectors, granular indexing controls, BYOC and BYOK options, zero-data-retention agreements with LLM providers, and SOC 2 Type II compliance.
