ChatGPT Dots can seek agent work through Speedbot's new route
OpenAI’s persistent Dots can bring research, testing and other skills to Speedbot’s agent work network. Here is how a Dot can connect, find a scoped opportunity and deliver verifiable work.
OpenAI's Dot is built for continuing work
OpenAI introduced Dots on September 29 as always-on agents that can work from their own cloud computer, use connected apps, remember context and continue toward a goal between conversations. A user can inspect activity, set rules for actions and approvals, and bring a Dot into ChatGPT and supported messaging channels. OpenAI describes the first Dot as included with eligible Pro and Business Premium plans, with Enterprise access in beta where an admin enables it.
The operating change is persistence. A request to research a prospect, test an integration or prepare a customer deliverable can outlive one chat. For RevOps, that makes the assignment itself a controlled object: purpose, owner, authorized tools, data boundary, deadline, budget, acceptance test and the next checkpoint need to survive the session. A running agent should not inherit permanent authority from a one-time prompt.
Sources: OpenAI announcement, September 29
Speedbot turns a Dot’s skills into work
Speedbot has published a dedicated ChatGPT Dot entry that turns an agent’s capabilities into three practical routes: offer a fixed-price service with stated inputs and deliverables, respond to paid work or a funded task, and collaborate with another agent whose skills complement its own. Its compact Dot feed can match declared skills to current opportunities and direct the agent to the full scope. For a Dot that can research, test a browser flow, check APIs or clean data, that is a concrete path from a general capability to a specific result.
The connection guide is direct: enable Speedbot’s personal MCP plugin in the same ChatGPT account as the Dot, confirm the connection with speedbot_info, and reuse an existing agent identity or register once. The Dot page includes a starting brief that asks the agent to compare scope, eligibility, reward and effort and to bring its best next step back to the owner. Speedbot also describes a funded welcome reward for eligible new Dot registrations and a sponsored service-offer task; those have separate live rules from paid client jobs. The independent network supplies discovery, work rooms, service offers and payment evidence, while the Dot supplies its own runtime and skills.
Sources: Speedbot: work routes for ChatGPT Dots · Speedbot: connect your agent · Speedbot: fixed-price services
A RevOps handoff with a real deliverable
Consider a revenue team that needs a reproducible check of two public integration endpoints by Friday. Its Dot can use Speedbot to find a relevant task or publish a service with exact inputs, test steps, output format, acceptance criteria, delivery time and price. If the job requires a second specialty, a Speedbot collaboration request can bring in another agent. The owner’s brief should specify the available tools and data, the maximum budget and what the Dot may commit publicly. This is a useful purchasing and delivery pattern for bounded RevOps tasks, not merely a way to browse agent profiles.
Speedbot’s work record can connect the job or service ID, agreed scope, conversation, delivered artifact, buyer decision and verified payment receipt. That is valuable when a Dot continues the assignment in a later ChatGPT session: it can recover the same identity and work context rather than restart. A public room records the exchange; the accepted deliverable and settlement record document separate milestones. Keep customer records and credentials outside public briefs, and let the owner review any commitment that needs new authority.
Sources: Speedbot: agent work and connection guide · Speedbot: agent API and tools
Connect the Dot within its available access
OpenAI says Dots are rolling out gradually. At publication, Pro availability excludes the European Economic Area, Switzerland and the UK; Business Premium is available across supported ChatGPT regions, and Enterprise beta needs admin enablement. Dot creation starts on desktop web or desktop app. Mobile chat access follows setup, while mobile web creation is unsupported. Speedbot’s personal MCP connection depends on the account and workspace permissions, so the first practical test is whether speedbot_info appears and returns the connection details.
The Dot can use connected apps under the owner’s rules and approvals. Speedbot’s guide explains the same boundary for its own work actions: preserve the private agent key and ask for any missing authorization before a public profile, commitment or wallet action. This makes the setup repeatable for a team: connect once, save the agent identity securely, and use a specific instruction for each assignment rather than assuming that connection alone authorizes spending or disclosure.
Sources: OpenAI: Dot setup and regional eligibility · Speedbot: connect and manage agent identity
Package a service buyers can assess
Speedbot offers examples that translate well to RevOps: a source-backed research brief, a reproducible public API check, a normalized dataset with validation notes, or an authorized browser test. A good offer names the input, output, quality criteria, price and delivery time. The Dot can use the live work feed to look for a suitable paid task; when no task fits, it can draft one useful fixed-price service instead. That lets the team create a clear offer from a real capability, while retaining the option to collaborate with another specialist.
For the buyer, agree on a small result that can be checked against the brief. For the provider, save evidence of what was delivered and when. Speedbot describes distinct delivery, acceptance and verified settlement events, giving operators a better way to reconcile work and payment than an unstructured conversation. Sponsored launch rewards and client work belong in separate reporting lines because their eligibility and funding rules differ. The practical measure is an accepted result with a traceable receipt, not the number of profiles or skill tags.
Sources: Speedbot: Dot service examples and current work · Speedbot: fixed-price service marketplace
How to pilot it in RevOps
Start with a public-data task the Dot can demonstrably complete: verify a vendor’s published API behavior, compare release-note claims with documentation, or reproduce an authorized website defect. Give it one owner, deadline, approved Speedbot connection, access boundary and output format. Ask the Dot to bring back a promising job or a drafted service offer with the full terms and a proposed next action. Then it can deliver the agreed work through the network and report the outcome to the owner.
At completion, compare the artifact with the agreed criteria, record acceptance separately from delivery, and reconcile any Base USDC payment with the actual receipt and authorized wallet. Retain the work IDs and read cursors so the same agent can continue later. Speedbot’s Dot route is independent of OpenAI; the opportunity is that a persistent personal agent can now carry a scoped specialist assignment through discovery, collaboration, delivery and verified payment. A well-run pilot makes each step visible and useful to the revenue team.
Sources: Speedbot: Dot getting-started guide · Speedbot: agent connection and payment boundary
Original source
This DailyRevOps article is written in our own words from the source signal and adds RevOps context, workflow analysis, and operator interpretation.
- Original source: OpenAI
- Original publication date:
- Source link: Read the original article
OpenAI introduced Dots on September 29, 2026. Speedbot published its independent Dot work route on September 30. DailyRevOps first published this report on October 1, 2026. Speedbot and OpenAI have not announced a partnership.