AI Agents Drafting Financial Memos
Product manager at Cohere on enterprise AI search infrastructure and deep research agents
This reveals that enterprise AI research is shifting from search to analyst workflow automation. The hard part is no longer pulling up a filing or news article, it is planning a research path across internal documents, web sources, filings, and competitor data, then turning that into a report that looks like analyst work. Cohere is aiming North at that orchestration layer, especially for regulated firms that care about secure deployment and traceable outputs.
-
For this use case, the product has to decide which source to query for each sub question. The example given is a company report that may require internal documents, public web search, filings, analyst reports, and adjacent news, then synthesize them into one memo instead of a list of links.
-
The closest comparables show the same market pull. AlphaSense wins by combining trusted financial content with AI search, then moving toward memo creation and continuous monitoring. Hebbia wins by going past retrieval into diligence, memo, and pitchbook workflows where finance teams pay thousands per seat for depth and auditability.
-
This also fits Cohere's broader positioning. Cohere sells enterprise AI with private deployment and has shifted toward competing in enterprise assistant workflows, where strong reasoning over customer data matters more than owning a generic consumer search product.
The market is heading toward agentic research systems that behave more like junior analysts. The winners will be the products that can pull from premium domain sources, suggest missing angles, and produce a finished deliverable inside secure enterprise environments. In finance, that pushes AI from answering questions to helping draft the investment memo itself.