Not one AI's opinion.
A process.
When you ask Greenwick about a stock — or when a strategy asks on its own — the question runs through a pipeline of specialist agents that research independently, argue both sides, and hand their conclusion to deterministic code for the final say. Here's the whole path, honestly.
-
Market data comes in
Live quotes, daily and intraday bars, and earnings context from your broker's feed — or your own market-data subscription if you bring one. Every analysis starts from the same numbers you can see on the chart.
-
Five analysts research independently
Each covers one discipline and writes its own report — no groupthink, no shared draft. The technical analyst reads price and volume structure; the fundamental analyst reads the business; the macro analyst frames rates, the dollar, and sector rotation; the news and sentiment analysts read what's being said and how the market feels about it.
TechnicalFundamentalMacroNewsSentiment -
The bull and the bear argue it out
Two researcher agents each take a side. The bull builds the strongest honest case for the trade from the five reports; the bear builds the strongest honest case against it. Greenwick deliberately pays an agent to talk it out of every trade — the step a single chatbot answer never has.
Bull researcherBear researcher -
A portfolio manager makes the call
It weighs the debate on the strength of the arguments — rhetoric is discounted, evidence wins — and issues one decision: buy, sell, hold, or no trade, with size, confidence, and the two or three points that decided it. Hold is a real answer, not a failure.
-
The risk engine gets the last word
Everything so far was judgment. This step is arithmetic: position-size caps, exposure limits, a daily-loss lockout, a confidence floor, no averaging down, session cutoffs. It's plain code with no AI inside and no way around it — the AI proposes; the risk engine disposes. A refused trade shows you exactly which rule refused it.
-
You approve — or your rules do
With approval rules on, Greenwick writes the full ticket — symbol, size, entry, stop, take-profit, reasoning — and waits for your click. You can edit it before approving. If you've enabled a strategy to act on its own, it acts only inside the same guardrails, and the big STOP button halts everything in under two seconds.
-
Open positions are watched, mechanically
Safety exits check every position on a fixed cadence — stops, take-profits, time limits, end-of-day rules — and a reconciliation loop treats your broker's records as the truth, every minute. These monitors are deterministic and keep running even for strategies you've switched off, and even if your subscription lapses with a position open.
-
Closed trades get a postmortem
After a round trip completes, a postmortem agent writes up what worked and what didn't, and the distilled lessons are injected into future analysis runs. The desk's memory lives in your local database — switch AI models any time and the experience carries over.
Postmortem analyst
The same pipeline, three ways to run it
Ask about any symbol
Type a ticker in Ideas and run the full analysis on demand. "Analysis only" mode runs the whole pipeline and stops before any order — research with zero execution risk.
Scheduled briefings
A morning scan before the open, a midday check, an evening review with the day's real numbers, and a weekly report with honest benchmark comparisons — all on a schedule you can move in plain local time.
Strategies, if you turn them on
Plain-English strategy cards — momentum, scalp, earnings — run the pipeline on their own inside your guardrails, at the Sentinel tier. A cold streak automatically dials a strategy down, then pauses it and tells you.
Why a team beats a takeaway
You could paste a ticker into any chatbot. Here's what that skips.
A single answer blends analysis, optimism, and style into one confident paragraph. There's no adversarial check, no position sizing, no stop, no memory of its last ten calls — and no way to know which part was insight and which was fluency.
Disagreement is built in: independent reports, then a paid devil's advocate, then a decision-maker who must justify the call, then hard-coded limits, then your approval. Every step is logged and auditable — you can open any run and read exactly why the desk did what it did.
See the pipeline run for real
Watch a live analysis — five reports, the debate, the decision, the risk check — in the demo.