# Best practices

> Proven techniques for accurate, on-brand, high-resolution agents: structure your data, pin high-stakes answers, set guardrails, and iterate from real conversations.

A great agent is mostly a function of great data and clear guardrails. These practices come from what consistently separates agents that resolve 70%+ of tickets from ones that frustrate customers.

## Structure your knowledge for retrieval

- **One topic per source.** Split a sprawling FAQ into focused sources. Retrieval works better on tight, single-topic chunks than on one giant page.
- **Lead with the answer.** Write support content answer-first; the model surfaces the first relevant lines.
- **Remove noise.** Navigation, cookie banners, and marketing fluff dilute retrieval. Point the crawler at content pages, not your homepage.
- **Keep it current.** Re-crawl when policies change. Stale data is the most common cause of wrong answers.

## Pin high-stakes answers with Q&A

Anything involving money, eligibility, or legal commitments should be a **Q&A pair**, not left to paraphrase. Q&A sources take priority over crawled text, so the exact wording you approve is what customers get.

> **DON'T LET THE MODEL IMPROVISE POLICY:** A model paraphrasing "30-day returns" can drift into "no-questions-asked refunds." Pin the precise policy as Q&A and the agent quotes it verbatim.

## Set clear guardrails in the prompt

Your system prompt should tell the agent what to do when it doesn't know, and what it must never do:

```text
Answer ONLY from the provided knowledge. If something isn't covered,
say you'll connect the customer with the team — never invent policy,
prices, order details, or shipping dates. Keep replies short and friendly.
```

## Choose the right model

Faster, cheaper models handle the bulk of routine questions well. Reserve a larger model for agents that reason over complex policies or long context. See [Models & model choice](/docs/agents/models) for the trade-offs and per-model credit costs.

## Iterate from real conversations

1. **Watch the logs** — Review real conversations weekly. The questions customers ask reveal the gaps in your data.
2. **Fix the data, not just the prompt** — Most wrong answers are a missing or unclear source. Add a Q&A pair and move on.
3. **Track resolution rate** — Use Analytics to measure how often the agent resolves without a human, and where it escalates.

## What's next

- [Response quality](/docs/getting-started/response-quality) — How grounding, citations, and retrieval produce trustworthy answers.
- [Data sources](/docs/agents/data-sources) — Every source type and when to use it.
- [Run a support audit](/docs/audit/overview) — Score your current support transcripts for accuracy and resolution.