OpenAI opened Businesses in ChatGPT in beta on November 10, 2026, introducing Business Hub alongside Business Agents initially limited to eligible US ecommerce and online service businesses. The launch gives companies a managed presence inside ChatGPT. For researchers and university research offices, its significance lies in distinguishing a verified commercial representative from independent information.
What Business Hub and Business Agents add
The launch combines business profiles, engagement reporting and conversations with business representatives. OpenAI says Business Hub lets eligible companies where ChatGPT operates claim their organization, update details after verification and provide product feeds. Eligible organizations can build, test and launch an agent that answers questions using information supplied by the business. These are related features with different access conditions, rather than a single entitlement attached to any organization’s name.
TechRadar’s November 11 report describes the entry point: when a business appears in a ChatGPT answer, a user can open its profile and, if available, choose its agent. The resulting conversation is labeled and separate from the original ChatGPT conversation. OpenAI also allows a business to sponsor its agent as a destination for ChatGPT Ads, with sponsored placements labeled separately from ChatGPT responses.
Profile verification does not establish independent authority
OpenAI’s business setup documentation separates claiming a profile, completing identity verification through Persona and publishing an agent. An authorized representative must have a claim approved; claiming alone does not produce the verified badge. Agent publication is a separate, optional step. A business does not need an advertising account or an Enterprise agreement simply to claim an eligible profile.
The Businesses in ChatGPT policy, updated November 10, makes businesses responsible for truthful information and prohibits misrepresentation of ownership, credentials or endorsements. Profiles can also contain information from other sources. OpenAI explicitly says it does not verify every detail or resolve every factual dispute. A verified identity therefore identifies who is represented; it does not independently substantiate every assertion appearing alongside that identity.
For research offices, that distinction connects the launch to AI provenance: the origin and authority of an answer matter alongside its wording. An agent speaking for a supplier is a source of that supplier’s representations. Treating its blue check as independent evidence of research suitability would go beyond what OpenAI’s profile policy establishes.
Privacy reporting has a defined scope
OpenAI’s November 10 community announcement says Business Hub reports aggregated mentions, website clicks and common topics in conversations citing a business. It says businesses do not receive users’ conversations or Business Agent chats. This is a company statement about information supplied to businesses, rather than a claim that no information is processed anywhere in the service.
The setup documentation adds important qualifications: online-business mentions count references to a business’s website, rather than every occurrence of its name. Discovery topics summarize themes and do not directly measure customer intent. Blank metrics can reflect reporting thresholds, ongoing collection or unavailable reporting. For researchers studying AI-mediated discovery, these definitions limit what the dashboard can establish about visibility or demand.
Commercial agents operate under platform rules
The Business Agents Policy, also updated November 10, requires an agent to identify itself as AI representing a business. It must not present itself as an independent adviser, a human or an OpenAI-endorsed representative. Businesses must keep supplied information current and must not configure agents to pressure users or manufacture urgency.
The policy also prohibits requesting unnecessary or high-risk personal information, including health details and authentication credentials. Sharing user information with the business or submitting an action on the user’s behalf requires explicit authorization through an approved feature. OpenAI can restrict features or suspend or remove agents when concerns arise. These provisions make platform oversight part of the commercial relationship; they do not constitute evidence that an individual agent has passed an institution’s own review.
The launch extends an earlier advertising experiment
The background predates this week’s beta. On September 16, OpenAI announced testing Sponsored Agents with select US advertisers. That experiment already separated a business-sponsored conversation from the user’s original chat. November’s development adds business claiming, information management and reporting around a wider business presence, while retaining sponsorship as one route into agent conversations.
In its November 10 coverage, The Next Web’s Ana Maria Constantin emphasized the combination of verified profiles, visibility analytics and agents that can also serve as advertisements. TechRadar’s Mike Moore focused on company discovery and the distinction between aggregate reporting and private chats. These accounts frame the change as an expansion of business participation inside the assistant, rather than an announcement of institutional research-agent availability.
What changes for universities and research offices
The immediate research-practice issue is how commercial evidence enters an AI conversation. OpenAI’s identity and disclosure policies make source attribution relevant to vendor due diligence: a supplier’s agent can explain its offering, while its claims still originate with an interested party. The launch does not establish a university’s eligibility to publish an agent.
Authority over organizational representation is another concrete issue for AI governance. The profile policy limits management to owners or authorized representatives and allows OpenAI to decline categories or remove material. An organization’s appearance in ChatGPT and its permission to control that appearance are consequently different questions under the published rules.
Current limits and planned expansion
The present capability also has limits. OpenAI’s setup page says agents explain offerings and link to business websites; they do not currently complete purchases, bookings or order checks. OpenAI plans to expand markets, categories and tools for more complex organizations, but its November 10 announcement provides no dated expansion timetable. The confirmed development is the beta and its stated eligibility, with further access remaining a plan.
Sources
- OpenAI, Introducing Businesses in ChatGPT, November 10, 2026; checked 2026-11-15.
- OpenAI Help Center, Setting up your business on ChatGPT, current documentation; checked 2026-11-15.
- OpenAI, Businesses in ChatGPT policy, updated November 10, 2026; checked 2026-11-15.
- OpenAI, Business Agents Policy, updated November 10, 2026; checked 2026-11-15.
- OpenAI Developer Community, Introducing Businesses in ChatGPT in beta, announcement November 10, 2026; checked 2026-11-15.
- OpenAI, Reimagining advertising with AI, September 16, 2026; checked 2026-11-15.
- TechRadar, Mike Moore, ChatGPT is adding verified Business Agents, November 11, 2026; checked 2026-11-15.
- The Next Web, Ana Maria Constantin, OpenAI lets businesses claim ChatGPT profiles and launch their own agents, November 10, 2026; checked 2026-11-15.
