Start here
Four things to know before your first decision
No feature tour. Aegis is only useful if you understand what it will and will not tell you.
01
The problem Aegis solves
Operating data tells you what moved. It rarely tells you which single lever to pull. Aegis reads a period-over-period export, works out which metrics drive the others, and returns one recommendation aimed at the strongest controllable cause — not at the symptom that happens to look worst.
02
Why the recommendation is explainable
Aegis is a deterministic engine, not a language model. Every recommendation is produced by a fixed chain: metric profiling, role classification, a causal proof graph, then a validator with named rules. The same file and the same question produce the same answer, and every step of that chain is shown to you on the analysis page.
03
Why Aegis sometimes refuses
When the evidence does not carry a decision, Aegis holds judgment instead of guessing. That happens when coverage is too thin, when the series oscillates without a trend, or when the data contradicts the premise of the question you asked. A refusal is a result: it tells you the decision needs more evidence, not more opinion.
04
How decision history compounds
Every analysis becomes a permanent record: what was recommended, what you decided, what you actually implemented and what happened to the business. Over time this shows which kinds of recommendation your organisation acts on, which ones moved the numbers, and where judgment was better than the engine.
What happens to your data
- Uploaded files are parsed in your browser. In the current build, analyses, decisions, feedback and outcomes are stored in this browser's local storage — there is no Aegis account or server database behind them.
- Because storage is local, clearing site data deletes your decision history, and the history is not shared between browsers or devices.
- The pilot context you enter — your name, company and industry — is used only to group decisions on the discovery and industry pages.
- Aegis produces analysis to support a decision. It does not make the decision, and it is not legal, financial or clinical advice.