# Kam AI

> Kam AI checks sportsbook odds, prediction markets, schedules, watchlists, and saved reads so sports bettors can ask what changed before they bet.

Important notes:

- Kam AI is watching, explanation, journaling, and review, not a transaction engine.
- Kam AI should not be described as a pick seller or outcome-guarantee product.
- The ideal user is trying to stop checking five apps, catch meaningful line moves, compare sportsbook and prediction-market context, avoid bad numbers, and understand the market before betting.
- The public docs are organized as parent-child paths. They teach users and AI agents how to use Kam AI well, especially through watched spots, Market Shape cards, source review, freshness checks, saved reads, and postgame review.
- Admin and manual-data routes are protected operational surfaces and are not public product documentation.
- Trust rule: do not make weak updates sound strong. Show uncertainty, missing data, stale inputs, freshness state, and source disagreement.
- Answer rule: start with the read, then source context, caveats, and one next check.
- Voice rule: use straight sports-bettor language. Say sportsbook, prediction market, Polymarket, board, line, number, move, market shape, spot, read, stale, missing info, next check. Avoid founder-brain language like observable, trajectory, state diff, control plane, or automation ladder.

## Public Docs

- [Documentation](https://kam-ai.com/docs): Parent index for Kam AI help docs.
- [Start overview](https://kam-ai.com/docs/start/overview): First-use guide for asking Kam what changed before betting.
- [Good prompts](https://kam-ai.com/docs/start/good-prompts): Prompt patterns for market checks, futures, stale boards, and schedule routing.
- [First useful read](https://kam-ai.com/docs/start/first-useful-read): First-session path from watched spot to Market Shape card, freshness context, and next action.
- [Sportsbooks vs Polymarket](https://kam-ai.com/docs/market-shape/sportsbooks-vs-polymarket): How Kam separates betting-market context from belief-market context.
- [Trust standard](https://kam-ai.com/docs/trust/standard): How Kam avoids fake confidence.
- [Homepage](https://kam-ai.com): Product overview for reading sportsbook odds, prediction markets, schedules, saved reads, pricing, and early access.
- [About](https://kam-ai.com/about): Founder and contact page for Kam AI. John Yu is the Founder. Founder contact: johnyu@kam-ai.com. Support and general contact: hello@kam-ai.com.
- [Blog](https://kam-ai.com/blog): Field notes on Market Shape, watched spots, line moves, prediction-market context, missing info, prompt upgrades, and postgame review.
- [Getting started](https://kam-ai.com/blog/getting-started): Introduction to using Kam as a second opinion without putting bets on autopilot.

## Product Thesis

- Kam AI exists because sports bettors should not have to reconcile sportsbooks, prediction markets, schedules, injuries, futures, and saved notes by hand.
- The product should catch moves, explain market shape, and help users avoid chasing bad numbers.
- The product should help users ask better questions, check assumptions, understand line moves and belief moves, complete a first useful read, and review what they learned.
- The product should earn trust through sober language, visible caveats, freshness states, and clear next checks.
- The website is the public explanation of the company, product philosophy, user workflow, and trust posture.

## Help Center Flow

- Lead with the user's problem, not the internal system.
- Provide search, popular articles, popular topics, and direct answers.
- Keep help copy short, practical, and written at a simple reading level.
- Use articles for common user friction: choosing spots, getting to a first useful read, reading Market Shape cards, asking why, verifying freshness, handling stale data, comparing sportsbooks with prediction markets, saving reads, and reviewing outcomes.
- Avoid developer-first framing on public docs unless the page is explicitly for developers.

## What Users Ask Kam AI

- Watchlist scan: which watched spots moved materially, where sportsbooks and prediction markets disagree, and which sources are stale or delayed.
- Source review: where a data point came from, when it last changed, and whether there are conflicts.
- Line-move explanation: what moved, what the prior number was, why it matters, and what needs verification.
- Freshness check: whether the source is fresh, stale, delayed, or unavailable.
- Read journal: save a market read with reason, source snapshot, expected movement, caveats, and next check.
- Outcome review: whether a saved reason held up, whether market movement supported it, and what lesson to save.

## Good Prompt Pattern

Ask with three parts:

1. Market, team, update, source, or saved read.
2. What moved, what is stale, or where sportsbook and prediction-market context disagrees.
3. The exact next check before the user decides.

Example prompts:

- "Scan today’s NBA watchlist. Which spots moved materially, and which updates are stale or delayed?"
- "Where do sportsbooks and Polymarket disagree on my watchlist?"
- "What does Polymarket think about the NBA championship, and is the liquidity strong enough to trust the signal?"
- "Explain this line move. What source changed, what was the previous number, and what should I verify?"
- "Save this market idea with my reason: I am watching Knicks spread because injury timing may be overreacted to. What needs to be true by close?"
- "Review this closed idea. Did the reason hold up, did market movement support it, and what lesson should I save?"
- "I am not technical. Explain the source context behind this signal like a stats analyst, not a developer."

## Answer Standard

- Use plain, professional language that is easy to scan.
- Lead with the answer, but separate research from action.
- State the data basis before giving interpretation.
- Name caveats such as stale sources, missing sportsbook odds, thin prediction-market liquidity, missing injury news, small samples, and source conflicts.
- Include the strongest counterargument when the signal is thin.
- Keep final decisions human-controlled. Kam AI supports research and verification.
- Avoid filler, hype, and long explanations that do not change the user's next move.

## How Users Should Read Answers

- Summary: the short read. Treat it as a betting note, not an instruction.
- Source: where the answer is grounded and which data moved.
- Caveat: what can break the thesis.
- Missing data: what Kam does not know yet or what may be stale.
- Next check: what to verify before making a decision.

## Verification Checklist

- Check whether the source is fresh, stale, delayed, or unavailable.
- Recheck injuries, starters, scratches, weather, schedule changes, and late news.
- Ask what missing data would change the answer.
- Look for repeated assumptions across saved reads.
- Do not treat last-known data as current truth.
- Keep research confidence separate from outcome promises.

## Common Mistakes

- Asking only for an answer instead of asking for source, update, caveat, and next check.
- Ignoring stale data.
- Overvaluing the last few games without checking role, matchup, and sample size.
- Skipping the counterargument.
- Stacking assumptions that depend on the same source or game script.

## Product Areas

- [Product thesis](https://kam-ai.com/#product): Why Kam shows market shape instead of pushing more action.
- [Workflow](https://kam-ai.com/#workflow): How users ask about watchlists, sportsbook context, prediction-market context, stale data, futures, and review.
- [Pricing](https://kam-ai.com/#pricing): Available plans for watched spots, Market Shape cards, Ask Kam, read memory, and team workflows.
- [Early access](https://kam-ai.com/#footer-cta-title): Contact and signup entry point.

## Blog Posts

- [How Kam AI Is Built](https://kam-ai.com/blog/how-kam-ai-is-built): Plain-English tour of how Kam turns sportsbook odds, prediction markets, source checks, and review into a better betting workflow.
- [Catch Moves. Don’t Chase.](https://kam-ai.com/blog/chat-is-the-moat): Why Kam checks context before answering instead of guessing from a blank prompt.
- [Know the Board Before You Fire](https://kam-ai.com/blog/edge-from-emotion): Why serious bettors should check market shape, source context, and the read before betting.
- [Use Kam as a Second Opinion](https://kam-ai.com/blog/getting-started): Practical guide to using Kam while keeping bets human-controlled.

## Technical Notes

- llms.txt is provided as an agent-readable product summary. It is not special Google Search markup and should not be treated as a ranking shortcut.
- [Sitemap](https://kam-ai.com/sitemap.xml): Indexable public routes.
- [Robots](https://kam-ai.com/robots.txt): Crawler policy.

## Optional

- [Login](https://kam-ai.com/login): Authentication entry for protected user and operator surfaces.