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

# Analyze Your Onboarding with AI

> Five prompts that turn live onboarding data into rankings, funnels, checklist friction reports and survey readouts — no dashboards, no exports.

Building your onboarding is [half the loop](/build-onboarding-with-ai) — the
other half is knowing what works. The MCP server exposes the same per-type
analytics the dashboard uses (flow funnels, per-task checklist data,
per-question survey breakdowns, session history), so an AI assistant can read
the numbers **and** explain them: rank content, find the drop-off step, cluster
survey verbatims into themes.

The video below runs five analysis prompts against live data:

<iframe className="w-full aspect-video rounded-xl" src="https://www.youtube.com/embed/dLy3KTpuFqM" title="Analyze your onboarding with AI — Usertour MCP demo" allowFullScreen />

<Tip>
  Analysis needs **no write access**. Connect the assistant with **Read-only**
  on the OAuth consent (or a read-scoped token) and it can do everything on
  this page but change nothing — the grant enforces what the prompts request.
  See [read-only by default](/api-reference-v2/mcp#read-only-by-default-prompt-injection-safety).
</Tip>

These prompts deliberately ask for more than any product tracks and instruct
the assistant to say which metrics are unavailable — so you get an honest
report of what the data supports, not invented numbers. Run them as-is, or trim
them to the metrics you care about.

## 1 — Portfolio overview: everything live, in one table

```text theme={null}
Use the Usertour MCP to retrieve performance data for all active flows,
checklists, announcements, surveys, and Resource Center content from the
last 30 days.

Present the results in a concise table with:

* Content name and type
* Number of users reached
* Starts or impressions
* Completion rate
* Dismissal rate
* Average completion time, if available
* Primary CTA conversion rate
* Trend compared with the previous 30 days

Highlight the strongest and weakest experiences. Clearly state the
reporting period, sample size, and any metrics that are unavailable.
Do not change any content.
```

## 2 — Effectiveness ranking: top three, bottom three

```text theme={null}
Use the Usertour MCP to analyze all onboarding content from the last
30 days and rank it by effectiveness.

Evaluate each experience using:

* Reach
* Engagement rate
* Completion rate
* Drop-off rate
* CTA conversion
* Contribution to activation, when activation events are available

Identify the top three and bottom three experiences. Explain the likely
reasons behind their performance using only available data. Separate
confirmed findings from hypotheses.
```

## 3 — Flow funnel: find the drop-off step

Replace `[FLOW NAME]` with one of your flows:

```text theme={null}
Use the Usertour MCP to analyze the flow named [FLOW NAME] over the
last 30 days.

Build a step-by-step funnel showing:

* Users who entered each step
* Step-view rate
* Next-step conversion rate
* Drop-off rate
* Average time spent on each step
* Final completion rate

Identify the step with the largest meaningful drop-off. Recommend
specific changes to the copy, CTA, targeting, or step structure, but
do not modify the flow.
```

## 4 — Checklist friction: which tasks stall adoption

```text theme={null}
Use the Usertour MCP to analyze the User List onboarding checklist from
the last 30 days.

For every checklist item, report:

* Number of users who started it
* Number and percentage who completed it
* Manual versus automatic completions, if available
* Median time to completion
* Related flow launch and completion rates
* The most common completion order

Identify the checklist items that contribute most to activation and the
items that create the most friction. Recommend whether any items should
be reordered, rewritten, merged, or removed. Do not make changes.
```

## 5 — Survey readout: from responses to opportunities

```text theme={null}
Use the Usertour MCP to analyze responses from the product feedback
survey.

Provide:

* Total invitations, starts, submissions, and completion rate
* Average and median rating for each rated feature
* Rating distribution
* Overall satisfaction score
* Most frequently praised features
* Most common usability problems
* Recurring feature requests
* Major themes from open-ended responses
* Representative short excerpts, with personal information removed

Group feedback into positive, neutral, and negative themes. Do not
invent sentiment scores or conclusions when the sample size is too
small.

Finish with the five most important product opportunities, ranked by
user impact and frequency.
```

## Where to next

<CardGroup cols={2}>
  <Card title="Build your onboarding with AI" icon="sparkles" href="/build-onboarding-with-ai">
    The other half of the loop — four prompts that ship a complete onboarding
    experience.
  </Card>

  <Card title="MCP server reference" icon="robot" href="/api-reference-v2/mcp">
    Connect any client, see every tool, understand scopes and read-only mode.
  </Card>
</CardGroup>