Customer story
Itochu International is one of Japan's largest general trading houses, and its analysts spend a great deal of their time doing something that is not analysis: assembling the material. Filings from one place, market research from another, the firm's own agreements and reference documents from a third, in two languages, before the actual thinking starts. Itochu Insight is the system built with NeuralSeek to compress that assembly step — and it runs inside Itochu's own Azure environment, which was a condition rather than a preference.
What was built
Each of these is a configuration of the platform rather than bespoke software, which is why it was built in the timeframe it was.
Keyword and vector retrieval run together over regulatory filings, live web research and Itochu's own documents, so a question about a counterparty draws on the public record and the firm's own history of dealing with them in the same answer — with the passage each claim came from cited back.
Rather than a single large prompt, the work is decomposed across a set of narrow agents coordinated by orchestrators — one driving the research path, another driving report production. Each agent does a small, checkable job, which is what makes the intermediate steps reviewable rather than a single opaque generation.
Reported financials are extracted from filings into a structured, year-over-year view sitting alongside the narrative, with each figure traceable to the filing it came from. The point is not the chart; it is that the number in the summary and the number in the table are the same number, from a document you can open.
Output is generated as finished slide decks and PDFs in the formats an analyst actually circulates, in several standard shapes — a summary, a longitudinal view, a risk view, a peer comparison. The last mile of an analyst's day is producing a document, and a chat window does not do that.
A single research paper took one of our analysts three weeks. Two and a half of those weeks weren't analysis. We built something here with the right partner and the right platform. What used to take three weeks now takes 13 minutes, and I signed off.
The deployment characteristics a similar organisation would want to know before asking anything else.
Inside Itochu's existing Azure tenant
Microsoft Entra single sign-on, with validated session tokens
Per-tenant document separation between business units
SEC filings, live web research, and Itochu's own private documents
English and Japanese, from one governed source rather than two corpora
CPP Associates
Itochu Insight is running in production inside Itochu's Azure environment. The quote above is the one Itochu has published with us and is the only outcome claim on this page — deliberately, because it is the only one with a named person's name attached to it. If you want to know whether a comparable engagement is possible in your organisation, the useful conversation is about your sources and your deployment constraints rather than about someone else's timeline.
More stories
What each one built and where it runs.
Home improvement · HR & call center
Retail & commercial banking
M&A advisory
Universal banking · State bank
Pension fund administration
Fintech · Digital payments
Medical education · Pharma & biotech
Higher education · Student services
Telecom · Investor relations
Software · HR operations
Consumer tech · Internal HR
Story 1 of 11: Great Day Improvements
Talk to a NeuralSeek expert about how it fits your stack, where your data has to live, and the governance your auditors already expect.