Capabilities

Coverage

What it does

1. Earnings & guidance

• Pull guidance from any earnings call - revenue, EPS, margins, opex, mix
• Benchmark management's guide against Street consensus, reported actuals and peer results
• Line up the management commitments call-over-call & see where promises have tightened, softened, or slipped
• Earnings calendar - see who's reporting in any date window with BMO/AMC timing and consensus EPS/revenue
• Surprise history - last N quarters of actual vs estimate per ticker with surprise %

2. Filings diagnostics

• Footnote-level semantic search across SEC filings (10-K, 10-Q, 8-K, 4) and earnings transcripts
• Surface non-headline items the news misses (eg. accounting oddities, liability tails)
• Working-capital diagnostics - inventory days, DSO, DPO, cash conversion cycle
• Reconcile balance-sheet movements against management commentary and flag divergence

3. Events, news & sentiment

• Typed news stories with versioned threads — earnings, corporate, macro, legal, analyst, sector
• Custom-window event chronologies across regulatory, corporate, product, legal, PT
• Quantify what's priced into a catalyst, and isolate what isn't
• Aggregated social chatter by ticker, plus the individual posts driving it

4. Analyst Estimates, ownership & market data

• 5-year forward consensus - revenue, EBITDA, EPS, FCF
• Rating and price target changes by analyst firms
• Ownership - major, institutional, mutual-fund, insider transactions
• Real-time prices and full options chains
• Historical dividends and splits

Methodology — the discipline the system enforces

• Citations on every claim: Numbers, quotes, and assertions all link back to a source
• No hallucinations: If something isn't in the corpus, the system says so. It doesn't invent figures, dates or quotes.
• Data gaps are flagged, not filled: No interpolation, no estimation around missing data

In practice

Two real questions, two real outputs. The conclusions are quoted directly from what the system returned.

"Pull every FSD and Robotaxi commitment from the last 4 Tesla earnings calls, map it against what's actually shipped, and tell me if the slippage pattern is bad enough that the next miss breaks the stock."

It pulled four sequential transcripts. Found a one-year-per-call slippage cadence on autonomy deployment. Separated the hardware commitments that landed on time (Cybercab SOP, April 2026 — the one clean win) from the autonomy ones that keep slipping. Layered in the stock reactions: −3.3% on Q4 '25, −3.3% again on Q1 '26.

The verdict:

"Bulls have demonstrably moved their goalposts along with Musk's. You won't get a 'breaks the stock' moment on yet another Robotaxi city-count slip. You'll get it on Cybercab execution — that's what they're still pricing."

"Tesla's working capital has been weird for a few quarters. Walk me through inventory days, AR, and deferred revenue, and tell me if the trend is consistent with the demand story management has been telling."

It pulled five quarters. Computed inventory days and DSO. Saw the Q3 2025 signature instantly — inventory days collapsed from 71 to 49, then snapped back to 74 — and tied it directly to the federal EV tax credit expiring 30 Sept 2025. Output a chart with revenue, inventory days, and DSO on one canvas so the divergence reads in a glance. Flagged deferred revenue as a data gap rather than guessing.

The verdict:

"The Q3 drawdown was clearly policy-driven, not product-driven. A +$2B inventory build into Q1 2026 on falling revenue is the opposite of supply-constrained — not the demand-strength story management is telling on the auto side."