Dhancept

Institutional Equity Research Council

This is the centerpiece of Dhancept's analysis engine: an agentic hierarchy that replicates a professional investment committee.

The 11-Stage Research Council

Every Deep Report runs the same 11 stages, which you can watch live on the progress screen.

  1. Skill Selection: Picks the research skills (for example India-specific listing and surveillance checks) that apply to your query.
  2. Supervisor Agent (Scope): Receives your query, determines the asset class, gathers historical prices, and builds the research outline.
  3. Specialist Desks (Analysis, six stages): Six specialist agents review the asset in parallel:
    • Macro Analyst: Evaluates inflation, currency trends, and interest rates.
    • Fundamental Analyst: Reviews balance sheets, free cash flows, and margins.
    • Valuation Strategist: Calculates DCF, PE ratios, and relative sector multiple targets.
    • Technical Chartist: Screens moving averages, RSI divergence, and volume distributions.
    • Sentiment Aggregator: Evaluates news sentiment, social discussion trends, and analyst targets.
    • Risk Auditor: Focuses entirely on headwinds, regulatory risks, and worst-case scenarios.
  4. Portfolio Manager (Challenge): Synthesizes the desk reports and challenges assumptions to identify logical gaps. If desks disagree, an investment-committee debate of up to 3 rounds runs inside this stage.
  5. Chairman (Synthesis): Compiles the final research view and its confidence.
  6. Independent LLM Judge (Quality): Grades the report for completeness, consistency, reasoning and evidence.

Research Council Interface

AI Research Council

AI Research Council pipeline showing the active research stages and generated report for TCS.NS.

📂 GRADING FILES

You can drag and drop annual reports (PDF, Word, or Excel) directly into the Council interface before executing. The AI agents will ground their research directly on your files rather than relying solely on external APIs.