Corpus Agentis
The field book to agent ecosystems
The field book to agent ecosystems
Catalogue · Finance

Finance agents

Agents built for financial work: investment research, financial analysis and document review across filings, transcripts and market data. A dense, regulated domain where the numbers have to be right, so the credible ones ground every claim in a source and augment analysts rather than act alone. Click a card for its full spec.

3 agents Live data · click a specimen for its anatomy
Field notes

Dense inputs, exact outputs, provenance as the product

Financial agents concentrate where the work is procedural and the record is already structured: reconciliation, reporting, research summarisation and compliance checks. Trading and advice remain thinly represented.

Show more

The reason is accountability rather than capability. A firm can automate a control and still name the human who owns it; automating a recommendation moves regulated responsibility. Expect the category to keep expanding through the back office first and to reach the client-facing layer last.

From the corpus, curated by Brandon Chaplin
Common questions
What do AI agents do in finance?

Research and analysis at volume: reading filings, transcripts and reports, pulling out the relevant figures and summarising them with a source attached. The gain is covering far more material than an analyst can, not making the decision at the end.

Are AI agents accurate with numbers?

Not reliably on their own. Models are weak at arithmetic and can misread a figure from a table, so credible finance tools pull numbers from a structured data source or a calculator rather than generating them. Check any figure against the document it came from.

Can AI agents trade stocks automatically?

Technically yes, and rule-based algorithmic trading has done a version of this for decades. Letting a language model trade unsupervised is different: it is hard to audit and hard to explain to a regulator, so firms keep a named human accountable for the decision.

What data do finance AI agents use?

Public filings, earnings transcripts, market data feeds and a firm’s own internal documents, usually fetched at query time rather than trained in. Access to the right data, and the discipline of citing it, separates these tools more than the model underneath does.

Next in the learning path