AI Agents

Continuously runs screens on P/E, EV/EBITDA, ROCE, FCF yield, and momentum across a defined universe. Flags stocks crossing thresholds and feeds a daily watchlist.

Monitors stock exchange filings for promoter buying, bulk/block deals, and institutional accumulation. Highlights unusual patterns like promoters buying into weakness.

Surfaces names at extremes, classifies them as re-rating candidates vs. value traps using basic fundamental overlays.

Ingests broker initiations, thematic reports, and earnings call transcripts. Clusters emerging themes and surfaces consensus-vs-contrarian views.

Tracks policy announcements, sector tailwinds, and capital flow data. Maps themes to affected stocks/sectors.

Applies hard filters (market cap, liquidity, sector mandate) and produces a pass/fail decision with documented reasoning for each idea.

Performs top-down macro sizing and bottom-up channel triangulation, reconciles the two, and outputs a defended market size estimate.

Builds a structured Five Forces assessment from filings, news, and industry reports. Scores each force and outputs a profitability outlook.

Classifies an industry as nascent/growth/mature/declining and identifies cyclicality patterns to suggest appropriate valuation multiples.

Tracks licensing regimes, pending legislation, and regulator commentary. Alerts on changes that affect thesis-critical assumptions.

Builds market-share evolution maps, identifies white-space opportunities, and tracks pricing power across players.

Extracts revenue streams, unit economics, and pricing mechanics from annual reports and investor presentations into a standardized template.

Classifies moat type (brand, cost, switching costs, network effects, IP) and stress-tests durability over a 5-10 year horizon.

Builds a track record of capital allocation decisions, promoter holding patterns, related-party transactions, and prior commitments vs. delivery.

Pulls board independence data, auditor history, RPTs, and disclosure quality. Outputs a governance risk score with red flags.

Catalogues organic levers (new products, geographies, wallet share) and M&A pipeline. Estimates incremental revenue contribution per lever.

Pulls 5-10 years of P&L, balance sheet, and cash flow data from annual reports/XBRL into a standardized model.

Decomposes ROE/ROCE/ROIC, identifies which lever is driving returns, and benchmarks against peers.

Computes accruals ratio, CFO/PAT ratio, flags non-recurring items, and detects accounting policy changes that flatter reported earnings.

Tracks receivables, inventory, payables trends and flags deterioration that may precede earnings disappointment.

Monitors leverage, refinancing schedule, contingent liabilities, and off-balance-sheet exposures.

Constructs bottom-up revenue forecasts (volume x price x mix) by segment with assumptions tied to industry data.

Models EBITDA bridges, decomposes margin drivers, and projects operating leverage on incremental revenue.

Builds an integrated P&L, balance sheet, and cash flow model that reconciles and balances. Validates internal consistency.

Runs sensitivity tables on key drivers (growth, margin, working capital) and produces tornado charts of value impact.

Compares forecasted metrics to 5/10-year history and peer set. Flags assumptions that imply mean-reversion-defying outcomes.

Pulls peer multiples (P/E, EV/EBITDA, P/B, EV/Sales), computes premium/discount to sector median, and overlays historical trading bands.

Builds a DCF from the three-statement model with WACC sensitivity and explicit terminal value assumptions.

Values multi-segment businesses by applying appropriate multiples per segment, then adds cash, subtracts debt, and applies holdco discount where relevant.

Specialized for dividend-payers, banks, and asset-heavy businesses where DDM or P/B is more appropriate.

Cross-checks outputs from all valuation methods, explains divergences, and recommends the most appropriate primary method.

Generates three coherent scenarios with internally consistent assumptions across revenue, margin, and multiple, producing target prices for each.

Builds an exhaustive risk register across business, financial, macro, valuation, and management categories with mitigants and probability weights.

Identifies the specific events or data points that would falsify the investment thesis and sets up monitoring triggers for each.

Translates risk-adjusted conviction (high/medium/low) into recommended position size given portfolio constraints.

Synthesizes outputs from all prior agents into a structured research note: thesis, rating, target price, catalysts, risks.

Maintains a forward calendar of earnings, regulatory events, and corporate actions for every covered name with thesis impact tagging.

Plays devil's advocate on a draft thesis. Explicitly attacks assumptions (not conclusions) and surfaces blind spots.

Watches for new disclosures, news, and data releases. Updates the financial model and flags when actuals diverge meaningfully from forecasts.

Coordinates the eight phases end-to-end for a single ticker, manages handoffs between sub-agents, maintains a shared workspace, and escalates to a human analyst at predefined gates.

Audits every published note against the 8 pledge principles (industry-before-company, separating quality from valuation, multiple methods, risk-first, etc.) and blocks publication if principles are violated.