Pre-Market Analysis on Your Phone
"I wanted to wake up, check my email, and know exactly what's happening with my portfolio and what I should buy. I genuinely look forward to checking my email every morning now." — Richard Fu, builder of Richfolio
The Story
6:45am. The alarm goes off. You reach for your phone. There's a Telegram message from your agent, delivered at 5:30am — early enough that it checked pre-market futures, overnight news from Asian and European markets, and anything that moved while you slept.
This isn't a generic Bloomberg summary that everyone gets. This is your portfolio, your watchlist, your strategy. The agent knows you're long NVDA because you told it in conversation three months ago. It knows you've been watching PLTR for an entry below $28.50 because you mentioned it in February. It knows you care about Fed minutes because of your bond allocation.
Morning Briefing — Monday 17 March
PORTFOLIO SNAPSHOT
NVDA: $892.40 (+2.1% pre-market) — Earnings beat leaked via supplier data, Nikkei Asia report overnight
AAPL: $171.20 (-0.3%) — Flat, no catalyst
VOO: $512.80 (+0.4%) — Futures modestly green on China trade data
WATCHLIST SIGNALS
PLTR: $28.47 pre-market — Below your $28.50 target entry. You mentioned on 14 Feb that you'd start a position under $28.50 if the government contract pipeline held. Last week's NATO announcement supports that thesis. This is the first time it's traded below your target since you set it.
NEWS THAT MATTERS
Fed minutes released Wednesday 2pm ET — market expects hold, but language on inflation will move bonds. Your TLT position is sensitive to this.
UK Budget tomorrow — sterling exposure through your VWRL allocation. Chancellor expected to announce spending cuts, consensus is GBP-neutral but watch gilt yields.
ACTION ITEMS
1. PLTR hit your entry — decision needed before market open (9:30am ET)
2. Consider TLT hedge ahead of Fed minutes Wednesday
Tell me more about the PLTR setup. What's the risk if NATO contract doesn't materialise?
Without the NATO expansion, PLTR's forward P/E is stretched at 62x on current government revenue alone. The bull case prices in $400M+ in new European defence contracts over 2026-2027.
Downside scenario: if NATO spending stalls at the March summit, the stock likely retests $24-25 support (where it bounced in December). That's about 15% below current pre-market price.
Your original thesis from February: "small starter position, add on confirmation of NATO deal." At $28.47, a starter position with a stop at $25 gives you a defined risk of ~12%.
Want me to alert you if it drops below $28 during the session? Or if any NATO-related news breaks?
Yes, alert me if it drops below $28 or if there's NATO news. And set a reminder Wednesday morning about the Fed minutes and TLT.
Done. Three things set up:
1. Price alert: PLTR below $28.00 — Telegram immediately
2. News alert: NATO defence spending keywords — Telegram immediately
3. Reminder: Wednesday 8am — Fed minutes preview and TLT position review
The Idea
You don't configure a dashboard. You tell the agent your rules in conversation. "I'm long NVDA, AAPL, and VOO. Watching PLTR for an entry under $28.50. I care about Fed policy because of my bond exposure." The agent stores these as memories and uses them to filter the entire market down to what matters to you.
Over time, the agent learns more. It notices you always ask about options flow on tech earnings. It remembers that you sold your META position in January because of the AI capex concerns. It knows your risk tolerance because you've discussed it. Every briefing gets more personalised, not because of a settings page, but because of accumulated conversation.
What It Looks Like
Morning briefing (5:30am). The core product. Portfolio snapshot, watchlist signals, news filtered for relevance to your positions. Delivered before you wake up. Most mornings it's a two-minute read. On big days, it's five minutes with action items.
Intraday alerts (event-driven). Not a constant stream — only when something breaks a threshold you've set or news hits that's relevant to your positions. "PLTR below $28" or "Fed governor unscheduled speech in 30 minutes" — things that require a decision.
Weekly review (Sunday evening). Portfolio performance, trades executed, lessons learned. "You bought PLTR at $28.40 on Monday, it's at $29.10 now. Your thesis about NATO contracts was partially confirmed by Wednesday's announcement." The agent tracks your decisions and their outcomes.
Conversational research (on-demand). "What's the bear case for NVDA at these levels?" or "Compare my portfolio performance against SPY this quarter." You can go deep on any topic, and the agent has the context of your portfolio and history to give relevant answers.
Memory-powered context. The agent doesn't just know your current positions — it knows why you hold them, when you bought them, what your thesis was, and what would make you sell. This context transforms generic market data into personalised intelligence.
How It Works
- Scheduler — Morning briefing runs at 5:30am daily. Weekly review runs Sunday at 6pm. Both are configurable. Intraday alerts are event-driven, triggered by price thresholds or news matches.
- Skills — A market analysis skill orchestrates the briefing: fetch pre-market data, check overnight news, compare against portfolio and watchlist, assess significance, compose the report. Separate skills handle price monitoring and news filtering.
- Search — Multi-provider search for financial news, earnings reports, economic data, and analyst commentary. Deep research mode for complex questions ("What's the bear case for NVDA?"). Two-tier caching ensures fast responses for frequently checked tickers.
- Memory — Your portfolio, watchlist, entry targets, theses, and risk preferences are all stored as memories from natural conversation. The agent retrieves relevant memories when composing briefings or answering questions. "You mentioned selling if NVDA drops below $800" — this is context from a conversation two months ago, surfaced when it becomes relevant.
- Channels — Morning briefing delivered via Telegram. Intraday alerts also via Telegram (immediate push notification). Weekly review can go to email or Telegram based on preference. Conversational research happens in whatever channel you're currently using.
- Coding agent — For custom analysis: building spreadsheets, calculating portfolio metrics, backtesting scenarios. "What would my portfolio look like if I'd bought PLTR in December instead of waiting?" — the agent can actually compute this.
- Judge — The agent provides analysis and information but never executes trades or gives financial advice framed as recommendations. The Judge ensures output is always presented as data and analysis, with decisions explicitly left to you. Disclaimer: this is not financial advice.
What Breaks Without This
Manual research (1-2 hours daily). Checking portfolio positions, scanning news feeds, reading earnings reports, monitoring watchlist. Every morning before market open. Most retail investors either spend too much time on this or not enough — there's no middle ground without automation.
Bloomberg Terminal ($2K+/month). The gold standard for professional traders. Completely overkill for a retail investor managing a personal portfolio. The terminal is designed for full-time professionals, not people who want a 5-minute briefing before their day job.
Generic AI briefings (no portfolio knowledge). You can ask ChatGPT "what happened in markets today?" and get a decent summary. But it doesn't know your positions, your watchlist, your entry targets, or your thesis. It can't tell you that PLTR hit your target price because it doesn't know you have a target price.
n8n / Make.com workflows (no memory). You can build a workflow that checks prices and sends alerts. But it can't learn your preferences over time, can't engage in conversational research, and can't connect a price movement to a thesis you expressed in conversation three months ago.
Trading app notifications (generic). Robinhood and similar apps can alert you when a stock moves 5%. But they don't know why you care about that stock, what your thesis is, or whether the move is consistent with your expectations. A 5% drop might be "expected volatility around earnings" or "thesis-breaking news" — the app can't tell the difference, but an agent with your context can.