FAQ & troubleshooting

Why is my problem ranked low?

Ranking is a composite of frequency, severity, revenue impact, and strategic fit — with a floor that caps problems scoring very low on severity or strategic fit. A problem with many mentions but little business impact will rank below a less-frequent but more strategic one. Click the problem to see its scores and the evidence behind them.

What is staleness?

Evidence loses influence as it ages (roughly a six-month half-life). Evidence tied to successful outcomes decays slower; evidence tied to failed predictions decays faster. This keeps your rankings current and rewards evidence that proves useful.

Do I need my own API key?

You can use platform credits, or bring your own OpenAI or Anthropic key (stored encrypted). If an agent or chat says it needs a key, add one in Settings → Billing. Keys are only decrypted at the moment they’re used.

Can I undo what the AI did?

Yes. Chat proposals are previews you approve. Agent-created nodes are reversible via undo, and rejected review items stay on the canvas (dimmed) rather than being deleted. High-stakes actions like synthesis ask for confirmation first.

Do I have to follow the phases in order?

No. Phases are a guide. Synloom gives gentle hints (like “no evidence added yet”) but never blocks you from working wherever you need to.

Nothing happens when I run synthesis

Synthesis needs evidence to work with. If your workspace has no evidence items, import some first (a dozen is enough to start). Check the Evidence tab to confirm your imports landed.