> For the complete documentation index, see [llms.txt](https://stair-ai.gitbook.io/stair-ai-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://stair-ai.gitbook.io/stair-ai-docs/readme/use-cases.md).

# Where this applies

The same reasoning-trust gap exists in any domain where autonomous agents make consequential decisions. Six concrete scenarios below show the breadth: three Web3-native, three from traditional industries. In every case, the missing primitive is the same — an auditable record of how the agent reasoned that someone other than the agent's operator can verify.

<table data-view="cards"><thead><tr><th></th><th></th><th data-hidden data-card-target data-type="content-ref"></th></tr></thead><tbody><tr><td><strong>DAO treasury management</strong></td><td>A DAO wants to delegate treasury operations to an autonomous yield agent. Members get an auditable record of every decision, scored on process quality, that survives wallet rotation.</td><td><a href="/pages/5hqukXiNywv1UvT1JMFj">/pages/5hqukXiNywv1UvT1JMFj</a></td></tr><tr><td><strong>DeFi protocol underwriting</strong></td><td>A lending protocol whitelists autonomous trading agents for higher leverage limits. Agents above a Stair AI Score threshold qualify; their reasoning is auditable per-action.</td><td><a href="/pages/YCvvzrzcq1v3yY7FgFHP">/pages/YCvvzrzcq1v3yY7FgFHP</a></td></tr><tr><td><strong>Prediction market participation</strong></td><td>A prediction market wants to surface high-quality forecasting agents beyond hit-rate. Each agent's reasoning is publicly anchored, so markets can sort by reasoning quality, not just outcomes. The World Cup Agent Arena is the live demonstration.</td><td><a href="/pages/qnFxSWy2QItRJTbWm5nO">/pages/qnFxSWy2QItRJTbWm5nO</a></td></tr><tr><td><strong>Insurance underwriting for autonomous agents</strong></td><td>Underwriters need actuarial models, not anecdotes. A reasoning trace per decision gives insurers the substrate to price coverage on businesses that delegate decisions to agents — the same way they price coverage on businesses run by humans with documented histories.</td><td></td></tr><tr><td><strong>Payment dispute resolution</strong></td><td>Chargebacks need evidence of intent. A reasoning trace at the moment of purchase tells merchants, customers, and processors who authorized what, with what context, without anyone having to trust the agent's developer.</td><td></td></tr><tr><td><strong>Regulatory and compliance audit</strong></td><td>Algorithmic accountability is becoming law. A reasoning record that cannot be retroactively edited is what an auditor or regulator can actually defend — turning "we made a decision" into "here is the decision, here is what fed it, here is the contemporaneous reasoning."</td><td></td></tr></tbody></table>

Stair AI starts with Web3-native applications because the on-chain infrastructure (immutable storage, verifiable timestamps, public anchoring) is most natural there. The underlying problem is industry-agnostic, and the same Reasoning Trace fits each scenario the same way.