OpenAI Business Hub brings company profiles to ChatGPT

Key Takeaways

• OpenAI’s beta Business Hub lets companies verify ChatGPT profiles, share product feeds and view aggregated engagement data; it does not expose individual user conversations.
• Eligible U.S. e-commerce and online-service businesses can build ChatGPT agents, and sponsored agents may appear in separately labeled ChatGPT Ads. OpenAI says ChatGPT has 1.2 billion weekly users and plans broader availability.
• J. Crew’s Anthropic-powered Style Advisor offers outfit suggestions based on shopper needs, events and locations; it launches on desktop, with a mobile version planned.
• Invoca found that 70% of surveyed U.S. B2C marketing leaders’ companies use or test AI agents, but only 27% of marketers using them measure revenue, bookings or conversions.
• Nearly three in four marketers surveyed by Invoca cannot track an AI-agent lead from its first ad click to a sale; only 19% reliably return conversation outcomes to ad platforms.
• Ohio State and Microsoft’s five-year health-care partnership targets moving 70%–80% of clinical and research workloads to Azure and using Dragon Copilot to reduce administrative work.
• Telefónica Tech recommends automated testing, sandboxing and human review for AI-generated code; GitHub Copilot Code Review can flag potential errors, security risks and maintenance issues.
• Cisco is bringing Claude Mythos into AI Defense’s adaptive red teaming to test customer-registered AI targets; the capability is planned first for select enterprises.
• Proton plans more transparency on internal AI use, including open-sourcing some tools and publishing security and AI-policy details, while denying it tracked employees’ AI-generated code.
• Oracle CFO Hilary Maxson says cloud demand remains strong; Howard Marks urges investors to assess AI’s long-term risks and opportunities despite uncertainty.

OpenAI adds business profiles and agents to ChatGPT

OpenAI launched Business Hub on Tuesday, giving companies a place to create business profiles in ChatGPT. Users can ask questions about a business and its products, and eligible companies can build specialized agents to answer questions. Businesses must verify their profiles, though OpenAI has not shared the full process. The move follows OpenAI’s effort to make ChatGPT a shopping hub, which it shut down in March, and brings the service closer to Google’s business discovery features.

OpenAI gives businesses tools to connect with ChatGPT users

OpenAI’s Business Hub lets companies claim and verify profiles, update business details, and share product feeds with pricing and availability. Shopify merchants can connect their stores through Business Hub or Shopify’s administration platform, and eligible products already appear in ChatGPT through Shopify Catalog. Companies can see aggregated engagement data, such as mentions in ChatGPT responses, website clicks, and common conversation topics, but cannot view individual chats. Eligible e-commerce and online service businesses in the United States can build agents that answer questions about their products and services. OpenAI plans to expand the program to more markets and business types.

OpenAI launches Business Hub profiles, AI agents and analytics

OpenAI launched Business Hub, which lets companies create verified profiles, offer specialized AI agents, and track engagement on ChatGPT. The feature is in beta and is expected to roll out to businesses later this year. Companies can share details about their products and services and view metrics such as profile views, business mentions, and website clicks. OpenAI says the hub can help businesses reach people who use AI to find products, compare services, and make buying decisions. The company says the profiles also give businesses more control over how their information appears in ChatGPT results.

OpenAI adds verified business profiles and AI agents to ChatGPT

OpenAI is launching its beta Business Hub, where companies can claim and manage profiles, add product feeds with pricing and availability, and see engagement insights. OpenAI said 1.2 billion people use ChatGPT each week. Eligible U.S. e-commerce and online service businesses can build agents that answer follow-up questions when users visit a company profile. Businesses may sponsor agents for ChatGPT Ads, with sponsored placements labeled separately from ChatGPT responses. Insights include business mentions, website clicks, and common discovery topics, and OpenAI says the data is aggregated and does not reveal user conversations. The company plans to expand Business Hub and agents to more markets and business categories.

Invoca finds most marketers cannot track AI agent leads to sales

Invoca’s Broken Funnel Report, released on November 10, 2026, found that nearly three in four U.S. consumer-brand marketers using AI agents cannot track a lead from its first ad click to a sale. The survey included 451 marketing leaders. Only 19% of marketers running agents can reliably send conversation outcomes back to ad platforms, and 26% can connect an agent conversation to the campaign or ad that led to it. Invoca, a Santa Barbara company that sells call tracking, conversation analytics, and AI agents to advertisers, also released separate report editions for home services and healthcare.

Invoca finds most marketers do not measure AI agents by sales

A survey of 451 U.S. business to consumer marketing leaders found that 70% of their companies use AI agents or are testing them. But only 27% of marketers using the tools measure them by revenue, bookings or conversions. Marketers also struggle to link AI conversations to ads and track leads after they go to local branches. Invoca reported that 84% of marketers say their companies respond to web leads within an hour, while only 36% of consumers said they got a response that quickly. The survey ran from Sept. 14 to Oct. 5, 2026.

J. Crew launches AI stylist for personalized shopping advice

J. Crew introduced Style Advisor, an AI shopping tool powered by Anthropic models and trained on outfit pairings from the company’s photo shoots and its styling principles. The tool asks shoppers questions about what they need and uses details such as the event and location to suggest outfits. It will launch on J. Crew’s desktop website, with a mobile app version planned for later. The company says Style Advisor may eventually use shoppers’ past purchases to make more personal recommendations. J. Crew also monitors the tool’s cost and is expanding internal AI training and discussion programs.

Study warns AI reasoning can make systems seem more capable than they are

The article examines how people may mistake the fluent language of generative AI for real understanding. It also warns that AI systems that explain their reasoning can seem like intentional minds with special knowledge. These impressions may lead users to defer to AI and assume it can handle more than the narrow tasks it performs well. The authors argue that this shift can reduce human agency.

Ohio State and Microsoft begin five year health care technology partnership

Ohio State and Microsoft announced a five year partnership to expand AI and cloud technology at the Ohio State University Wexner Medical Center. The plan includes developing smart hospital technology, moving about 70% to 80% of clinical and research workloads to Microsoft Azure, and linking data with Microsoft Fabric and OneLake. The partners will use AI tools such as Dragon Copilot to help reduce administrative work, while keeping clinicians responsible for patient care. They are also exploring AI supported tumor board workflows and hope to develop solutions for communities across Ohio, including rural areas. Early smart hospital projects have raised temperature control from 88% to 98%.

AI coding needs stronger tests and security checks, Telefónica Tech expert says

AI software development is moving beyond code generation to support planning, programming, testing and review, but developers still need ways to check that code is secure and reliable. Pablo López of Telefónica Tech describes how specification driven development and automated tests can check whether code meets requirements. AI can help create unit tests and test scenarios, while sandbox environments can help teams check complex behavior before release. GitHub Copilot Code Review can flag possible errors, security risks and maintenance problems as an addition to human review. The article also points to the OWASP Top 10 as a source of cybersecurity practices for automated reviews.

Proton pledges greater transparency about its AI practices

Proton says it will be more open about how it uses AI inside the company after users raised concerns about its AI strategy. It plans to release and open-source some internal tools, including more advanced versions of Lumo, and publish details about its code reviews, security reviews, and AI policies. Proton also denied installing laptop tracking software to measure employees’ AI-generated code and said it has added security reviews to address threats from AI systems scanning its open-source code. Security experts said the commitments could help rebuild trust, but stressed that the details and follow-through will matter.

Cisco adds Claude Mythos to its AI Defense security platform

Cisco has added Anthropic’s Claude Mythos to its AI Defense platform to help enterprise customers find weaknesses in AI agents before attackers can exploit them. The model helps the platform’s adaptive red teaming engine plan tests and generate attack ideas, including possible multi-step attack paths that standard testing may miss. Cisco says the system works only on a customer-registered target, follows objectives set by a human, and records each action for audit. Standard and adaptive red teaming are already available, while adaptive red teaming with Claude Mythos is coming soon to select enterprises, with wider access planned for qualified customers when available.

Oracle CFO says cloud demand remains strong

Oracle Chief Financial Officer Hilary Maxson said demand for cloud infrastructure remains strong as the company works to expand its data center capacity. Oracle shares rose on Wednesday morning after CoreWeave’s earnings helped support confidence in AI stocks. In an interview with Investor’s Business Daily, Maxson said she believes Oracle’s recent investor day forecast should help answer questions about the company’s growth targets.

Howard Marks urges investors to weigh AI’s long-term risks and opportunities

In his latest memo, investor Howard Marks says people looking at artificial intelligence should ask important questions before making long-term investment decisions. He stresses that investors should consider those questions even when clear answers are not available. Marks also warns against letting uncertainty stop them from acting in the fast-changing AI field. The memo offers a way to assess the sector’s long-term prospects.

Sources

• OpenAI rolls out new business profiles, making ChatGPT just a little bit more like Google
• OpenAI Gives Businesses New Tools to Reach ChatGPT Users
• OpenAI launches ChatGPT Business Hub with verified profiles, AI agents and business analytics
• OpenAI is launching verified business profiles and AI agents inside ChatGPT
• 74% of marketers can't track if AI agent leads become sales, Invoca finds
• AI Agent Sales Performance Lags, Invoca Research Finds
• J. Crew debuts new AI shopping tool that’s designed to think like a stylist
• When Reasoning looks like Hidden Understanding: How Chain-of-Thought Fosters Anthropomorphism, Perceptions of Epistemic Superiority, and Cognitive Deference
• Ohio State and Microsoft announced a five-year partnership, expanding AI use in health care
• AI-powered software development: how to generate secure and reliable code
• Following Heat From Users, Proton Promises to Be More Transparent About AI Strategy
• Cisco boosts its AI Defense platform with Anthropic’s Claude Mythos
• Oracle CFO: Cloud Demand Strong As Company Touts Growth Targets
• Howard Marks: Intelligent long-term investment questions rise with artificial intelligence