How to read an AI council stock signal
By DANA · Product education · Published 8 October 2026
Start with the decision date, then inspect the individual votes and the method behind them. A council signal is an algorithmic research output. Agreement describes the council; it does not establish that a stock will rise or that a trade suits anyone.
Check the date before the direction
A signal belongs to a particular data snapshot and decision cycle. DANA shows the decision date beside the ticker. A historical BUY or SELL label describes that cycle, not an instruction to act at the current price. Prices and company information can change after the snapshot.
A real example: agreement with abstentions
The public DANA signal snapshot retrieved on 8 October 2026 recorded CRWD with a BUY consensus dated 7 October 2026: 17 BUY votes, zero SELL votes and four HOLD votes among 21 agent entries. The displayed confidence was 100%. That example shows why the confidence label must not be read as “all agents agree” or “100% chance of profit”: four agents did not take a direction. This is a dated illustration of the interface, not a current view on CRWD.
Inspect methods, not just the final count
Open the council page to see the agents and their specialties, then inspect the ticker page for its recorded votes. Multiple models can share data or assumptions, so adding votes does not create independent evidence automatically. A strong consensus can still be wrong. HOLD entries and disagreement are useful context rather than errors to hide.
Separate three different questions
The vote breakdown answers what the agents recorded. Historical evaluation asks how similar outputs behaved in a particular test. Suitability asks whether a real decision fits an individual’s circumstances. DANA exposes research outputs and hypothetical evaluation; it does not answer the suitability question. Backtested results are hypothetical and past performance does not guarantee future results.
A repeatable reading checklist
Record the ticker and date; read the direction alongside HOLD and opposing votes; check the agent methods; compare the output with factual market information; and review the model and simulation limitations. Revisit the dated record rather than assuming today’s result is the same. No signal removes uncertainty or the risk of loss.