5 Relay AI Conversations That Could Save Your Trading Account

Relay launched three weeks ago with a long list of what it can do. Most traders are still using it for the obvious stuff: “how am I doing today?” and “show me my trades.”

That’s a start. But the most useful conversations you can have with Relay are the ones that surface something you didn’t know to look for.

Here are five worth having before your next trading session.

1. Ask it to run a bot health check before the open

Before your session begins, ask:

“Run a health check on all my active bots.”

Relay compares each bot’s recent performance against its historical baseline and flags anything degraded or idle. A bot underperforming its baseline by a meaningful margin is worth pausing before it takes new trades today.

This takes about 30 seconds. It catches issues that could cost real money if you let them run unexamined. A bot that performed well for three months and started slipping last week will show up here.

You can follow up with:

“Show me the last 30 trading days for the bots you flagged and highlight the drawdown periods.”

2. Ask about your worst time of day

You probably have a time window when you shouldn’t be trading. Most traders do.

Ask Relay:

“Show me my P&L and win rate by hour of day.”

The Smart Schedule Analyzer builds a heatmap from your trade history. When the data shows you’re consistently losing in the 3–4 PM window, you have a real reason to tighten your entry window or stop trading before the close.

This is one of those things that’s obvious in retrospect and invisible in the moment. Relay makes it visible.

Follow-ups that work well:

• “Filter that by my QQQ and TSLA bots only.”
• “Re-run this for the last 30 trading days instead of the full history.”

3. Run a what-if on your worst trade from last week

Pick your biggest loss last week and ask Relay to simulate an alternate exit.

“Replay my worst trade from last week with a 20% trailing stop instead of a fixed take-profit.”

Relay shows you what would have happened: the bar-by-bar path, entry, peak move, and the exit price under the alternate parameter. You see whether a different setting would have changed the outcome, and by exactly how much.

This isn’t about guilt. It’s about data. If a different parameter consistently produces better results across multiple trades, that’s a tuning signal worth acting on.

You can keep going with:

• “Now simulate that same trade with a 10% wider stop and no take-profit.”
• “Apply the same trailing-stop logic to all my losing trades from last week and summarize the impact on P&L.”

4. Check your bots for conflicts

If you run more than one bot on the same symbol, ask Relay:

“Check my bots for any symbol overlap or directional conflicts.”

The Cross-Bot Conflict Detector finds bots with overlapping symbols and entry windows that may be competing for fills or stacking directional exposure. Two bots trading SPY in the same window, both going long, double your exposure even if each one has its own position sizing rules.

Most traders set bots up one at a time and never step back to look at how they interact. Relay does it in one question.

Useful follow-ups:

• “Show me the last 20 trades where overlapping bots were in the market at the same time.”
• “Highlight any periods where my net exposure on a single symbol exceeded my max risk rules.”

5. Ask it to score a recent losing trade

Don’t ask Relay why you lost. Ask it to grade how you exited.

“Score my TSLA trade from Tuesday.”

The Trade Scoring feature grades your exit from A through F based on how much of the available move you captured (MFE/MAE analysis). An A means you captured most of the move before reversal. An F means you held too long or exited before the move played out.

The grade tells you whether your system is executing well, separate from whether the trade made money. A well-executed losing trade and a poorly-executed winning trade are different problems, and they require different fixes.

Follow-ups that help you improve:

• “Show me all trades from last week that scored a C or worse.”
• “Group my trade scores by bot and rank them from best to worst.”

How to get more out of these conversations

Relay maintains context within a single conversation. You don’t need to start a new chat for each question. Ask for the bot health check, then say “dig deeper into the one that’s underperforming” and it picks up from there.

The more specific you are about time periods and bot names, the more targeted the answers:

• Good: “Last month.”
• Better: “The last 30 trading days on my QQQ bot.”
• Best: “The last 30 trading days on my QQQ mean-reversion bot during the first 90 minutes after the open.”

If you haven’t tried the Assistant page yet (accessible from the main navigation), it gives you a full conversational workspace with history and quick-start templates. The widget works for quick questions. The Assistant page is where to go when you want to work through something in depth.

Use these five conversations as a pre-session checklist. They take a few minutes, and they can surface issues that would otherwise only show up as unexplained red on your P&L.

{
  "preSessionChecklist": [
    "Run a health check on all my active bots",
    "Show me my P&L and win rate by hour of day",
    "Replay my worst trade from last week with an alternate exit rule",
    "Check my bots for any symbol overlap or directional conflicts",
    "Score my last 10 losing trades and group the results by bot"
  ]
}