Skip to main content
This guide walks through building a Playbook from scratch — the investment strategy for a specific portfolio. A Playbook defines how that portfolio is run: its investment universe, risk rules, position sizing, screening criteria, and buy/sell discipline. Your Agent applies it automatically to every analysis it runs for that portfolio, so you do not restate the strategy each conversation. Your Agent’s own character and tone are separate — they live in its Profile, not the Playbook.

​
Why it matters

Without a Playbook, a portfolio has no standing strategy and you restate its universe, risk rules, and sizing every session. With one, every analysis your Agent runs for that portfolio respects them automatically, without you restating them. The investment is small and the payoff compounds across every future conversation about that portfolio.

​
What goes in a Playbook

A complete Playbook covers:
  • Investment universe — the markets and asset types you focus on
  • Watchlists — the names you track
  • Risk preferences — what you avoid, what you tolerate
  • Position sizing rules — concentration limits and sizing logic
  • Screening criteria — the factors that define a candidate for this portfolio
  • Buy/sell discipline — entry and exit rules, when to trim or add
  • Preferred research style — depth, framework, what to emphasize
  • Reporting format — how you want output structured

​
Step 1: Start with your universe and style

The fastest way to build a Playbook is to describe the portfolio’s strategy in a few sentences: 
Add to my Playbook: I focus on US large-cap technology and software. I prefer profitable companies with strong free cash flow, durable growth, and reasonable valuation. I avoid highly leveraged businesses and unprofitable speculation.

​
Step 2: Add your risk and sizing rules

Add to my Playbook: Maximum 10% per single position. At least 8 holdings. Flag any position that grows beyond 15%. Avoid concentration in a single end market.

​
Step 3: Define how you want research done

This shapes the output of every analysis: 
Add to my Playbook: For single-stock research, prioritize business quality, revenue durability, margin trend, free cash flow, valuation, and management commentary. Use technical indicators only as supporting context. Always separate facts from assumptions and include the risks and what would invalidate the thesis.

​
Step 4: Add your watchlist

Add to my Playbook: My core watchlist is AAPL, MSFT, NVDA, TSM, [others].

​
Step 5: Review and refine

Read the Playbook back and check it reflects how you actually invest: 
Show me my current Playbook. Is anything contradictory, vague, or missing?

​
Keeping it current

A Playbook is living. Review it whenever your strategy changes, your risk tolerance shifts, or your universe expands. A stale Playbook quietly steers the Agent toward assumptions you have already abandoned, which is worse than having no Playbook at all.

​
What to keep out

Temporary tasks do not belong in the Playbook. “Analyze AAPL today” is a one-off for chat, or a scheduled task if it recurs. The test: if you would want it applied six months from now, it belongs in the Playbook; if it expires this week, it does not.

​
Common mistakes

  • Too vague. “Invest in good companies” gives the Agent nothing to act on. Be specific about what “good” means to you.
  • Contradictions. Rules that conflict confuse the Agent. Review for consistency.
  • Letting it go stale. The most common failure is building a Playbook once and never updating it as the strategy evolves.