AI-native financial research

From scattered
information to
clearer research.

We’re developing the concept for a financial research terminal that brings company documents, market context and natural-language analysis into one workspace.

Explore the concept

Early-stage project · Not yet publicly available

JUV GPD / RESEARCH WORKSPACECONCEPT

A question worth investigating

What changed in this company’s latest annual report?

Proposed research workflow

Read the filings
Compare periods
Review evidence
Financial statementsExtract metrics
Management discussionIdentify changes
Risk disclosuresTrace sources

Illustrative interface concept. No live market data, generated research or connected AI service is shown.

The vision

A workspace built around the question.

Financial research often involves moving between filings, news and notes. Our proposed terminal is intended to help analysts and independent researchers explore those sources together, while keeping the underlying evidence visible.

01 / DOCUMENTS

Explore company filings

Planned tools for asking questions about annual reports, financial statements and other company documents, with references to the source material.

02 / COMPARISONS

Investigate what changed

A proposed workflow for comparing companies and reporting periods, extracting relevant figures and highlighting differences for human review.

03 / CONTEXT

Connect the wider picture

Planned research summaries that bring together documents and market news, distinguish observations from interpretation, and surface unanswered questions.

Planned use of Claude

AI-assisted analysis.
Evidence stays in view.

We plan to evaluate Claude for document understanding, structured information extraction and research summaries grounded in retrieved sources.

Our proposed approach combines document retrieval with language-model analysis. Any eventual integration will need testing for accuracy, citation quality and the handling of uncertainty.

  • Trace conclusions to sourcesKeep document references alongside research outputs.
  • Keep numbers verifiableSeparate extracted values from calculations and model interpretation.
  • Make uncertainty visibleFlag missing evidence and conflicting information.
  • Keep researchers in controlSupport investigation and human review rather than automate investment decisions.

Where we are

An early-stage idea.

Concept development

JUV GPD is currently being defined as a financial research product. The workflows on this page describe our intended direction; they are not available product features.

The next step is to build and evaluate a focused document-analysis prototype before expanding the terminal. No public product launch is announced at this stage.