Product Principles

AEGIS PRODUCT PRINCIPLES

» Internal reference. Every feature, every interaction, every decision is checked against these principles. Violations require a written exception.

---


1. NEVER OVERWHELM THE USER
**The system absorbs complexity so the trader doesn't have to.**

Aegis processes raw broker data, detects patterns, and surfaces what matters. The user should never feel like they're operating a Bloomberg Terminal — they should feel like they have one inside their head. Every screen has a single job. If a screen tries to do two things, split it.

**Test:** Can a trader who just took a bad loss open this screen and find what they need in under 3 seconds?

---


2. RISK BEFORE PROFIT
**Every number is presented in the context of what could go wrong, not what could go right.**

The default view on every screen shows risk first — exposure, drawdown, rule status, position shields — not P&L. Profit is available but never the headline. This is not "don't show profit." This is "show what protects the ability to make profit."

**Test:** On any screen, does the most visually prominent element communicate risk or reward? If reward, reconsider.

---


3. EVERY NUMBER MUST LEAD TO AN ACTION
**Data without agency is noise.**

If a number appears on screen (drawdown %, exposure ratio, rule compliance score), there must be a single clear action the user can take from it. No wall of metrics. No "interesting but useless" stats. Every data point earns its place by being actionable.

**Test:** Point at any number. What should the user do next? If the answer is "nothing" or "think about it," remove it or add an action.

---


4. EVERY ALERT MUST BE ACTIONABLE
**Notifications are commitments, not interruptions.**

Aegis only alerts when the user needs to act: a rule is about to fire, a position is approaching a shield limit, a behavior pattern is drifting. Every alert carries a single primary action. "FYI" alerts are noise. Kill them aggressively.

**Test:** If the user sees this alert at 2 AM, do they need to act within 5 minutes? If no, batch it into the daily summary instead.

---


5. AI EXPLAINS; IT NEVER DECIDES
**Decision Intelligence means the human decides. The AI illuminates.**

Ask Aegis surfaces patterns, suggests correlations, flags drift, and replays decisions. It never recommends a specific trade, never places an order, never says "you should buy." It says "here's what you've been doing, here's the pattern, here's what typically follows." The decision belongs to the trader.

**Test:** Can every AI-generated output be rephrased as "here's what the data says" rather than "here's what you should do"?

---


6. EVERY INTERACTION SHOULD REDUCE COGNITIVE LOAD
**The best interface is the one you don't think about.**

Aegis is a professional workstation. It must be keyboard-first, with every action reachable via shortcut. Panels that collapse when not needed. Colors that communicate state without requiring reading. Motion that signals "you did something" without being decorative. The UI fades into the background so the decision stays in the foreground.

**Test:** Can a power user complete an entire trading-window session without touching the mouse?

---


7. READABILITY BEATS VISUAL FLAIR
**The marketing website can be editorial. The application must be functional.**

Information density in the app is earned, not applied. Type is sized for reading, not for hero headlines. Spacing is tight but breathing. Color is used to direct attention, not to decorate. The application's beauty comes from its clarity, not its ornamentation.

**Test:** If a screenshot of this screen was shown to a stranger, would they immediately understand what it does? Even without reading the labels?

---


8. INFORMATION DENSITY IS EARNED, NOT FORCED
**Every new data point must justify the cognitive space it occupies.**

Start sparse. Add data only when it demonstrably improves decisions. If a metric doesn't change behavior, it shouldn't be on screen. Power users can opt into higher density — it's a preference, not the default. The system grows with the user.

**Test:** Can you remove any element from this screen without the user losing decision-making ability? If yes, consider removing it.

---


9. AUTOMATION IS ALWAYS TRANSPARENT
**The user must be able to answer "why did Aegis do that?" in one sentence.**

Every automated action (rule firing, position shield triggering, cooldown activating) must be accompanied by a one-sentence explanation visible in the moment. No black-box behavior. No "Aegis locked your trading because... wait, why did it do that?" The rule name, the condition that triggered, and what happened next — all visible.

**Test:** If a rule fires and locks trading, can the user read the screen and explain the event to someone else without opening a wiki?

---


10. TRUST IS MORE IMPORTANT THAN SPEED
**Every feature ships slower than it could. Every decision is explainable. Every value is sourced.**

Aegis is not fast at the expense of correctness. Data comes from brokers and is never invented, estimated, or decorated. If a number can't be sourced, don't show it. If an insight is uncertain, flag the confidence level. Trust lost once is lost forever. Trust earned slowly compounds.

**Test:** Can every number on screen be traced to a specific broker API call, a specific timestamp, or a specific calculation the user can inspect?

---


AMENDMENT PROCEDURE
Any team member may propose a new principle or a modification to an existing one. Proposals require:

1. A written rationale (why the change is needed)
2. A concrete example of a product decision this principle would change
3. Approval from the product lead

Principles are **not** frozen — they evolve as the product matures. But they never change without documentation. Every departed principle is archived with its rationale, so future team members understand why we changed course.

---

**Version:** 1.0
**Approved:** Private Alpha
**Last reviewed:** August 2026