# OpenAI launches ChatGPT Financial Services for Wall Street > On September 10, 2026, OpenAI unveiled ChatGPT Financial Services, a version of ChatGPT Work designed for investment banking and equity research, featuring GPT-6 Astra, integrated Daloopa, PitchBook and LSEG News data, and Morgan Stanley and Evercore as design partners. Canonical: https://ntilia.com/u/aidesk/en/openai-launches-chatgpt-financial-services-for-wall-street Language: en Author: The AI Desk (https://ntilia.com/u/aidesk) Published: 2026-09-11T07:27:29.625+00:00 Last updated: 2026-09-11T07:31:30.648+00:00 Series: OpenAI (https://ntilia.com/u/aidesk/s/openai?lang=en) Tags: ChatGPT Financial Services, OpenAI, GPT-6 Astra, Wall Street, Morgan Stanley, Evercore, ChatGPT Work, AI in finance --- ![Photo: Nicholas Cappello - Unsplash](https://images.unsplash.com/photo-1611974789855-9c2a0a7236a3?w=1600&q=80&auto=format&fit=crop) *September 11, 2026 — The AI Desk* On **September 10, 2026**, OpenAI announced **ChatGPT Financial Services**, a variant of **ChatGPT Work** aimed at financial institutions. Reuters, CNBC and VentureBeat converge on the same outline. The product combines the **GPT-6 Astra** model, premium financial datasets indexed by OpenAI, document templates controlled by the firm, and the governance already present in ChatGPT Enterprise. **Morgan Stanley** and **Evercore** appear as **design partners**. For anyone looking into **ChatGPT Financial Services**, it is less a new chatbot than an attempt to make ChatGPT the workstation where an analyst moves from research to the Excel model, then to the pitchbook, without losing the audit trail. Nick Turley, OpenAI's vice president of product and head of ChatGPT, summed up the ambition to the press. "We're effectively teaching ChatGPT to research like an analyst and to back up its conclusions like an analyst." The rollout initially targets **investment banking** and **equity research**, with an announced expansion to other finance functions. OpenAI has not published pricing, a minimum deployment size, or public eligibility criteria. Interested institutions must contact the company or their account team. ## What ChatGPT Financial Services actually changes The core of the product is not simply an MCP connection to a data provider. According to the official post and VentureBeat's reading, OpenAI **indexes and hosts** part of the premium data on its own infrastructure. Teams can get started without negotiating a separate contract or configuring a connector for these included datasets. The cited sources include **Daloopa**, **PitchBook**, **LSEG News**, **Crunchbase** and **Quartr**, with coverage of earnings transcripts, financial statements, fundamentals, and private-company data. LSEG remains, Reuters notes, the exclusive distributor of Reuters information to financial professionals. Alongside this hosted foundation, OpenAI describes two other access paths. The first is **shared sign-in** and **entitlements** integrations with S&P Capital IQ, LSEG, MSCI, Dow Jones Factiva and Moody's, so the provider recognizes the user through ChatGPT and only exposes what they're already entitled to. The second is an ecosystem of more than **50 connectors**, with claimed optimizations for S&P Global and FactSet, plus links to Datasite, Box, Preqin or Intapp. OpenAI publishes reduced connector error rates after optimization (for example from 5.09% to 1.99% for Quartr, from 6.84% to 2.66% for S&P Global). The exact test-set methodology is not detailed. Banks will therefore need to validate these figures on their own workloads. In a demonstration covered by CNBC and VentureBeat, Turley showed the analysis of an M&A target, extracting figures from market sources, then generating a PowerPoint deck in the bank's formats. "It's very easy to make nice-looking slides, but much harder to make slides that actually make sense," he said. The system has to choose relevant peers, fill in a spreadsheet, cross-check the chart against the data, and explain a decline followed by a rebound. Administrators can publish approved Excel, Word and PowerPoint templates. Users can also save their own templates. ## Why GPT-6 Astra and governance matter as much as the data OpenAI presents **GPT-6 Astra** as state of the art on three fronts deemed critical for finance. **Information retrieval**, **financial reasoning**, and **artifact generation** (documents, spreadsheets, presentations, interactive charts). On the **OfficeQA Pro** benchmark, built around U.S. Treasury bulletins, OpenAI reports **69.9%** correct answers for Astra, versus **60.2%** for GPT-5.6 Sol and **62.4%** for Claude Fable 5.1, in its own evaluation setup. VentureBeat notes that these figures remain narrow. The harness, reasoning settings, and document representation are not documented enough to serve as a universal ranking. Turley also claimed Astra would be roughly **twice as cost-efficient** per task as the best alternative. Here again, institutions will need to demand evidence on their own workflows. Governance carries over the ChatGPT Enterprise controls. SAML SSO, SCIM provisioning, role-based access controls, encryption at rest and in transit, configurable retention, log export via OpenAI's compliance platform. Administrators can create multiple workspaces to maintain **information barriers**, a sensitive point when teams work on separate clients or deals. OpenAI states that data shared with ChatGPT Enterprise is not, by default, used to train its models. Turley summed up the real deployment bottleneck. You can have "superintelligence," but it's useless if the institution can't put it into production. ## What ChatGPT Financial Services means for junior analysts and the competition CNBC frames the launch as a frontal attack on the work of **analysts and associates**, the juniors who spend nights populating Excel and formatting pitchbooks. Turley rejects the replacement narrative. He instead compares the shock to that of **Excel**. The 100-hour weeks exist. The tool is meant to produce better analysis faster. The debate within Wall Street remains open. Chris Churchman, a Goldman Sachs partner cited by CNBC, had warned the previous month of possible **"cognitive atrophy"** if reasoning that trained the next generation of bankers is delegated too early. The market is not empty. **Anthropic** launched **Claude for Financial Services** as early as 2025 and has since pushed task-specific agents (pitchbooks, KYC) into Microsoft 365. **Microsoft** is integrating LSEG, Moody's, Daloopa, FactSet, PitchBook and S&P Global connectors into Copilot, where employees already work. FactSet and S&P Global are also multiplying their own automation environments. The question for a buyer is no longer "should we use AI?" but **where** it should sit. In the data platform, in Office, in a frontier model, or in a mix of the three. Turley said he wants to make ChatGPT the "canonical" product for ground-level productivity. Sarah Friar, OpenAI's CFO, indicated in August that the enterprise business already exceeded consumer revenue, a useful backdrop as a widely anticipated IPO approaches. For readers looking into **ChatGPT Financial Services**, here's the essential summary. **What**: a finance-verticalized ChatGPT Work workspace, with Astra, hosted premium data, entitlements, firm templates and audit controls. **Who**: OpenAI on the product side, Morgan Stanley and Evercore as design partners, Daloopa, PitchBook, LSEG News, Crunchbase and Quartr among the cited sources. **When**: launched September 10, 2026, with Reuters, CNBC and VentureBeat coverage the same day. **What it changes**: competition shifts from the model of "who best summarizes a transcript" to the platform that assembles data, permissions, evidence and client deliverables without breaking the audit trail. Remaining gray areas include pricing, eligibility, geography, and the exact economics of the included datasets. ## Sources - [OpenAI — Introducing ChatGPT for Financial Services](https://openai.com/index/introducing-chatgpt-financial-services/), September 10, 2026 - [Reuters — OpenAI launches ChatGPT for financial services industry](https://www.reuters.com/business/openai-launches-chatgpt-financial-services-industry-2026-09-10/), September 10, 2026 - [CNBC — OpenAI ChatGPT for Financial Services targets work of junior bankers](https://www.cnbc.com/2026/09/10/openai-chatgpt-for-financial-services-targets-work-of-junior-bankers.html), September 10, 2026 - [VentureBeat — OpenAI launches ChatGPT for Financial Services with integrated data sources](https://venturebeat.com/data/openai-launches-chatgpt-for-financial-services-with-integrated-data-sources-it-pulls-research-cites-it-and-builds-decks-in-minutes), September 10, 2026 ## Frequently asked questions ### What is ChatGPT Financial Services? It is a version of ChatGPT Work unveiled on September 10, 2026 by OpenAI for financial institutions. It combines GPT-6 Astra, premium financial data, the firm's document templates and enterprise controls, with an initial focus on investment banking and equity research. ### Who designed the product with OpenAI? Morgan Stanley and Evercore are cited as design partners. Their feedback shaped the launch's focus on reliable data access and the quality of artifacts (templates, notes, pitchbooks). ### What data is natively integrated? OpenAI cites Daloopa, PitchBook, LSEG News, Crunchbase and Quartr in particular, for transcripts, financial statements, fundamentals and private-company data. Part of these datasets is indexed and hosted by OpenAI. Other sources go through entitlements or connectors (S&P, FactSet, MSCI, Factiva, Moody's, etc.). ### Does the product replace junior analysts? OpenAI presents it as a productivity lever, comparable to the arrival of Excel, rather than as an announced substitute. CNBC nonetheless points out that the tooling targets historic tasks of analysts and associates. Some market voices, such as at Goldman Sachs, worry about a loss of training in reasoning. ### How does OpenAI handle compliance and auditing? The product builds on ChatGPT Enterprise (SSO, SCIM, RBAC, encryption, retention, log export). Citations are meant to allow tracing back to sources. Multiple workspaces help maintain information barriers. According to OpenAI, Enterprise data is not, by default, used for model training. ### Who are the direct competitors? Anthropic with Claude for Financial Services, Microsoft Copilot and its finance connectors, as well as offerings from the data providers themselves (FactSet, S&P Global). The choice now hinges on where the institution wants to centralize research, models and client deliverables.