> ## Documentation Index
> Fetch the complete documentation index at: https://docs.harpgtm.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Qualification and playbooks

> Write qualification questions once, let your agent answer them per account with evidence, and package sequences as playbooks that require qualification before anyone is enrolled.

Qualification is how Harp decides who is worth contacting. You write the questions; your AI agent answers them for every account, with evidence; and playbooks use the answers to gate enrollment and fill your emails.

## The question library

Open **ICP > Qualification Questions**. Each question has a key (used as a template variable) and the wording your agent answers.

Good questions have a verdict, can be checked in public and ask for evidence. See [How to write qualification questions an AI agent can answer](https://harpgtm.com/learn/qualification-questions-for-ai-agents) and the [question templates](https://harpgtm.com/templates/qualification-questions).

## Answers

Your agent saves answers with `save_answers`: a verdict, the value, the evidence and the source. Every answer keeps its history and its age.

* `research_gaps` lists, per account, the questions that are missing or stale, so the agent refreshes only those.
* `account_brief` shows an account's answers with their evidence and age.

## Answers in your emails

Answers resolve as template variables keyed by the question key, and override variables with the same name. In a playbook campaign, an email whose variable has no answer yet is **held** and retried later, never sent with a gap.

## Playbooks

Open a sequence under **Sequences** and use its playbook settings:

* **Use as playbook**: campaigns can be started from this sequence.
* **Require qualification**: contacts are enrolled through your agent only if their account passes qualification.
* **Require copy review**: every email waits for review before it sends. See [Copy review](https://docs.harpgtm.com/guides/copy-review.md).
* **Qualification questions**: extra questions this playbook depends on.
* **Copy notes**: lasting rules for drafts in this playbook.

Your agent lists playbooks and starts campaigns from them with `playbooks`, and enrolls contacts with `enroll`. Skipped contacts come back with a reason.

> [!NOTE]
>
> **Require qualification** applies to enrollment through your agent (`enroll`). Enrolling from the web app does not check it.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.