Get to your first useful read.
The fastest way to understand Kam is to pick one watched spot, ask what changed, and review the Market Shape card.
www.kam-ai.com/docs/start/first-useful-read
Kam should prove value before it asks the user to understand a full dashboard. The first session should answer one real pre-bet question with source context, freshness, caveats, and a next check.
The first job
| Step | What to do |
|---|---|
| Pick one watched spot | Choose a game, team, future, prop, or saved read. Avoid starting from a blank chat box. |
| Ask what changed | Ask what moved, what is stale, where markets disagree, and what should be verified next. |
| Read the Market Shape card | Look for sportsbook context, prediction-market context, freshness state, missing info, and the single next check. |
A useful answer ends with action
- Wait when the number looks gone or source freshness is weak.
- Track when the market gap is interesting but not yet actionable.
- Pass when the read depends on missing or stale context.
- Save when the reason is worth reviewing after the game.
- Verify when one source would materially change the decision.
Understand Kam chat request and response cycles.
Use the same request and response contract as the React Native chat screen: visible prompt, request context, grounded source reads, structured response, and receipt.
Kam is a self-describing product.
Kam publishes a public semantic interface so humans and agents can understand capabilities, request context, receipts, response states, and typed next actions.