The platform combines GPT-6 Astra with institutional financial data from PitchBook, Daloopa, and LSEG News. Future OpenAI models will be added as they are released.
Financial Data Built In
ChatGPT can work directly with:
- PitchBook — companies, investors, funding rounds, valuations, and deals.
- Daloopa — financial statements, company fundamentals, and KPIs.
- LSEG — market data, financial news, and earnings transcripts.
Instead of relying only on model knowledge, ChatGPT retrieves current financial data and uses it as context for analysis.
From a developer perspective, the workflow is essentially:
query → data retrieval → financial reasoning → output → source citations
What It Can Do
The product is designed to automate common analyst workflows:
- company and industry research;
- financial statement analysis;
- earnings transcript analysis;
- comparable-company analysis;
- financial modeling;
- research reports;
- client presentations and other materials.
The key feature is granular citations. Figures and claims generated by ChatGPT can be traced back to their underlying sources.
For example, if the model uses a company's revenue, EBITDA margin, or guidance in an analysis, the analyst can verify where that number came from instead of treating the LLM output as a black box.
One AI Layer Across Financial Data
Firms can also connect existing financial data subscriptions and internal tools.
This effectively makes ChatGPT a reasoning layer over multiple financial databases:
PitchBook + Daloopa + LSEG + internal data → GPT-6 Astra → research/model/output
That could reduce the amount of manual work currently spent searching databases, extracting numbers, moving data into spreadsheets, and assembling research.
OpenAI's Nick Turley describes the goal simply:
We’re effectively teaching ChatGPT to research like an analyst and back up its conclusions like an analyst.
The important part isn't another finance chatbot. It's the architecture: LLM reasoning combined with specialized financial retrieval and source-level verification.
