How AlphaPod Works
A 9-Stage Research Pipeline
Before You See the Thesis.
Every stock runs through 9 specialized research stages — each one looking for what the last one missed. Accounting red flags, growth deceleration, hidden tail risks, unsupported claims. If it's there, one of the stages catches it.
Try It FreeSynthetic demo · Real runs use your watchlist and the full evidence graph.
The Process
What Happens When You Hit “Research”
Every thesis follows this pipeline. Each step feeds into the next — nothing is skipped.
Collect Evidence
Financial + alternative data feed financials, market data, filings, transcripts, developer signals, and crowding indicators into the research graph
Run the 9-Stage Pipeline
9 specialized research stages — Thesis, Quality, Growth, Doctrine, Risk, Portfolio, Publish, Falsify, Verify — each looking for what the last one missed
Verify Every Claim
Contradiction detection against the knowledge graph, evidence gap analysis, and cross-stage consistency checks
Deliver the Verdict
Conviction-scored thesis with risk bounds, thesis triggers, portfolio fit, and a full audit trail
Coverage Universe
600+ U.S. IT & Communications Stocks
From mega-cap platforms to emerging growth names — research any stock in the technology and communications universe.
Software & Cloud
MSFT, CRM, NOW, ADBE
Semiconductors
NVDA, AMD, AVGO, INTC
Internet & Media
GOOG, META, NFLX, SNAP
Hardware & Devices
AAPL, DELL, HPQ, WDC
Cybersecurity
CRWD, PANW, ZS, FTNT
Fintech
SQ, PYPL, COIN, AFRM
Telecom
T, VZ, TMUS, LUMN
IT Services
ACN, IBM, CTSH, EPAM
The 9-Stage Research Pipeline
Each Stage Looks for What the Last One Missed
Every analysis passes through 9 specialized research stages. Each stage is designed to challenge the thesis from a different angle — so blind spots don't survive.
Live pipeline view from a real Deep Research run
Junior Analyst
What it catches: Stale data, missing filings, and gaps in the numbers you'd build your thesis on.
Gathers a broad foundation of financial, market, and alternative data — financial statements, filings, analyst estimates, and real-time news — and frames the initial thesis with fundamental factor analysis.
Senior Quality Analyst
What it catches: Accounting red flags, earnings manipulation, and balance sheet risks that surface-level metrics miss.
Evaluates earnings quality, balance sheet resilience, and accounting red flags. Applies forensic accounting heuristics to detect issues that surface-level metrics miss.
Senior Growth Analyst
What it catches: Decelerating growth disguised by one-time revenue boosts and unsustainable TAM assumptions.
Analyzes revenue drivers, market share trends, product pipeline, and TAM expansion. Assesses whether growth is sustainable or decelerating.
Doctrine Analyst
What it catches: Poor capital allocation, management misalignment, and valuation assumptions that don't hold up.
Applies structured investment doctrine — valuation discipline, capital allocation quality, management alignment, and institutional ownership patterns.
Risk Analyst
What it catches: Tail risks, macro sensitivity, and downside scenarios your conviction doesn't account for.
Conducts scenario analysis, identifies tail risks, assesses macro sensitivity, and produces conviction-weighted risk/reward framing for the thesis.
Portfolio Manager
What it catches: Position sizes that don't match your conviction and portfolio-level concentration risks.
Synthesizes all prior analyses into a conviction-weighted recommendation with position sizing, risk/reward framing, and portfolio-level context.
Publishing Analyst
What it catches: Reports padded with model artifacts, restated thesis text, and detail an investor doesn't need to act.
Synthesizes the prior six analyst outputs into a polished investor-ready report following the AlphaPod thesis template. Drops what's not load-bearing, sequences the narrative, and surfaces the highest-conviction findings up front.
Falsify Analyst
What it catches: Forward-looking claims that don't survive a structured falsification attempt against the evidence graph.
Investigator stage. Takes every load-bearing forward claim from the published thesis and attempts to falsify it against the knowledge graph — prior contradictions, missing corroboration, stale evidence — before the report is finalized.
Verification Analyst
What it catches: Unsupported claims, cross-stage contradictions, and conclusions the evidence doesn't back.
Independently verifies claims against source evidence, flags unsupported assertions, checks for cross-stage contradictions, and produces a final grounding quality score.
Data Foundation
Multi-Source Data Layer. Nothing Left Out.
Every research report draws from a broad foundation of financial + alternative data — financials, market intelligence, filings, transcripts, and quantitative signals — so the analysis starts with the full picture, not just what's easy to find.
Financial Statements
- Income Statement
- Balance Sheet
- Cash Flow
- Ratios & Metrics
Market Intelligence
- Real-time Prices
- Analyst Estimates
- Institutional Holdings
- Insider Transactions
Alternative Data
- News Sentiment
- SEC Filings
- Earnings Transcripts
- Patent Activity
Quantitative Signals
- Technical Indicators
- Factor Scores
- Momentum Metrics
- Volatility Analysis
Evidence Architecture
Graph-Backed, Not Source-Exposing
Every claim in a thesis is tied back to evidence in our proprietary knowledge graph — for cross-source corroboration, contradiction detection, and audit-grade traceability. The graph is the differentiator, so the product surfaces categories of evidence and graph-level signals, not the underlying providers, raw URLs, or parser internals.
Sanitized Evidence Categories
Filings, transcripts, market data, developer signals, AI-infrastructure intel, retail crowding, labor signals, and prior thesis memory — surfaced as labeled categories, not raw provider feeds.
No Source Exposure
Provider names, raw URLs, document IDs, parser names, model identifiers, prompt text, raw edge weights, and pipeline internals never leave the platform. What you see is the graph-level conclusion and the evidence category — never the cookbook.
Audit-Grade Traceability
Every load-bearing claim links to graph evidence with a confidence label, contradiction count, and cross-source corroboration count — enough to defend the thesis without revealing how the graph was built.
Three visibility modes, one consistent contract
Public demos show graph categories and aggregate grounding scores only. Customer accounts see the same plus per-claim evidence count, contradiction count, and category labels. Internal admin tooling — used only by the research team — is the one place the underlying graph is fully visible. The redaction layer enforces this at the API boundary, not just the UI.
Research Quality
Research That Gets Better Over Time
Learns from Its Mistakes
Every thesis is tracked against actual market outcomes. Each analyst is recalibrated based on where it was right and where it was wrong — not just final P&L.
Catches Its Own Contradictions
Cross-references claims across all 9 stages to flag internal contradictions, unsupported assertions, and evidence gaps before you see the report.
Adjusts for Market Conditions
Three-level confidence calibration — overall market, sector-specific accuracy, and individual stock sensitivity — so conviction reflects real-world conditions, not just historical averages.
Watchlist Monitor
The Agentic Screener for Public-Equity Investors
A three-tier screening surface that answers the four questions a portfolio manager asks every day: why now, what changed, the supporting evidence, and the suggested next step. Cheap structured screen runs across the entire universe; deeper evidence overlays scope to the names you actually cover.
Tier 1 — Universe Scan
Cheap, daily, structured.
Estimate revisions, valuation extremes, and pre-earnings setups across ~600–800 names. Fast, low-cost structured screening — designed to surface the top 50 names worth a closer look in under a minute.
Tier 2 — Watchlist Deep Monitor
Evidence overlays where it matters.
Same screens scoped to the names you cover, with Developer Intel and AI Infrastructure overlays attached per result. Cost stays bounded because the deep work only runs on your watchlist.
Tier 3 — Event Boost
Pre-earnings setups, corroborated.
Names reporting in the next 14 days, joined against estimate revisions and Developer Intel signal. One click generates a pre-earnings brief; another runs Deep Research.
Bundled into every paid tier
Free gets a top-5 teaser. Starter ($79) unlocks the full universe scan plus a 10-name watchlist deep monitor. Pro ($179) auto-runs all three tiers daily, with email alerts on new high-conviction signals and a 50-name watchlist. Elite ($399) takes the watchlist to 999 names, supports multiple watchlists per analyst, custom alert thresholds, and per-watchlist email routing.Email-only by design — institutional record-retention and workflow fit.
Premium Intelligence
Lead Time on the Trends That Drive Price Targets
Four specialty intelligence streams that surface vendor adoption, AI infrastructure buildout, retail crowding, and labor cost pressure — months before they appear in earnings commentary or sell-side notes.
Developer Intel
Vendor adoption signals before earnings.
When a public company starts being adopted by AI-native developer ecosystems, that can become an early research signal — surfacing months before management mentions the customer in earnings commentary. Developer Intel surfaces those signals as conservative, evidence-graded ticker cards with cross-source corroboration.
Conservative labels by design. No buy/sell calls.
AI Infrastructure Intel
The AI buildout, mapped to public equities.
Datacenter projects, capex tracking, power-market readiness, hyperscaler economics, and cloud cost benchmarking — focused on the public-company beneficiaries of the AI infrastructure cycle. Sized for analysts who need to model what gets built where, for whom, and at what margin.
Includes an exposure map across the major hyperscalers and chip vendors.
Retail Intel
Trade against the crowd, not into it.
Retail crowding signals across the universe — top crowded names, security-level crowding history, variance decomposition, and a heatmap of where retail positioning has shifted. For active traders and L/S desks who need a contra signal on overcrowded longs.
Real-time alerts available at Elite tier.
Labor Intel
Read labor cost pressure before margin guidance.
Labor factor snapshots, hiring intensity, skilled-visa filings, wage analysis by sector, and the transcripts- derived signals that surface margin pressure. Used by staffing analysts, macro analysts, and single-name fundamental analysts tracking hiring as a leading indicator.
Advanced signals (hiring intensity, margin pressure) at Elite tier.
The same conservative discipline as Deep Research
Every premium intelligence product follows the same rule: surface signals that hold up under cross-source corroboration, label confidence honestly, and never pretend a single data point is a thesis. Buyers who want to skip Deep Research and use one or two intelligence streams alone can buy them as standalone subscriptions. Buyers who want everything choose Elite.
Run Your First Deep Research — Free
Pick any stock. See every risk, contradiction, and blind spot the 9-stage pipeline catches. Full report, full audit trail, no credit card required.
Start Free Trial