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

# Churn & Renewal Risk

> A standing watch for accounts showing early signs of churn, or of expansion. The quiet account is the one that does not renew.

Churn is rarely a surprise to the data and usually a surprise to the team. The signals were there: the champion stopped replying in March, the executive sponsor left in April, support volume climbed and then went silent, nobody from your side had a real conversation in nine weeks. None of it triggered anything because nobody was watching for the absence of things. Churn and renewal risk watches continuously, and flags accounts drifting in either direction, because the same signals that precede churn also precede expansion when they run the other way.

## Data Sources

| Category         | Required | Recommended |
| :--------------- | :------- | :---------- |
| CRM              | ✓        |             |
| Calls            | ✓        |             |
| Email + Calendar |          | ✓           |
| Support tickets  |          | ✓           |
| Messaging        |          | ✓           |

**Starter** (required only): renewal dates, account ownership, and what the calls say about health and satisfaction.

**Full** (required + recommended): the same, plus the signals that actually move first. Email engagement falling off, meeting cadence stretching, support volume rising or going quiet, and shared Slack channels going dead. Silence is the earliest indicator, and it only exists in email, calendar, and messaging.

See the [Context Graph](/docs/unifying-gtm-data/customer-context-graph) for how each source feeds the watch.

### Choose your path

| Path                       | Best for                                                                                 | Output is delivered                   |
| :------------------------- | :--------------------------------------------------------------------------------------- | :------------------------------------ |
| Build in Clearskies        | Teams who want a standing watch, since the whole point is not having to remember to look | To Slack DMs or channels              |
| Build in Claude or ChatGPT | Digging into one account, or reading the whole book before a QBR or a board meeting      | In the Claude or ChatGPT conversation |

## Build in Clearskies

This is the use case where the workflow path matters most. An analysis you have to remember to run does not catch the account nobody was thinking about.

### The Workflow

In the Clearskies app, open the workflow builder and describe what you want in plain language:

```text theme={null}
Every Monday at 7am, find the accounts with a renewal in the next 180
days. For each one, check engagement: when we last had a two-way exchange,
whether meeting cadence has stretched, whether known contacts have gone
quiet or left, and what support volume looks like. Flag accounts where
engagement has dropped materially, and separately flag accounts where it
has increased and expansion language has come up. Post the flagged
accounts to the #customer-success channel.
```

The app configures the schedule, the accounts in scope, and the destination.

Scheduled workflows have no meeting context, so the accounts come from a record lookup rather than a calendar. Record lookups cap at 100 records per step, so scope by renewal window or segment rather than sweeping the whole book at once. A 180-day renewal window is usually enough to keep it under the cap and still leave time to act.

Flag only what changed. An account that has been quiet for a year and always will be is not a signal; an account that replied within a day for eight months and has not replied in three weeks is. Set the thresholds so the channel gets a handful of accounts a week rather than the whole list.

### Deploy to Slack

Once published, only the accounts that trip a threshold arrive:

```text theme={null}
Renewal Risk — week of Aug 10
Accounts renewing in 180 days: 62 · Flagged: 4 at risk, 2 expanding

At risk

Meridian Systems · $140K · Renews Nov 30 · Owner: Priya
Engagement dropped sharply. Two-way exchanges ran weekly through June,
last one was Jul 21, 20 days ago. Four outbound touches since, no reply.
The monthly check-in has been declined twice. Shared Slack channel has had
no customer message since Jul 9. Support tickets stopped entirely in July
after running 3 to 5 a month, which is usually disengagement rather than
satisfaction.
Changed since last week: yes, this is the first week it has tripped.

Coastal Group · $88K · Renews Oct 15 · Owner: Priya
Executive sponsor left the company Jul 28, confirmed on a call. No
replacement relationship established. The remaining contact has never been
on a call alone with us.
Changed since last week: no, third consecutive week flagged.

Two other accounts flagged, both on a single signal (meeting cadence
stretching). Lower confidence.

Expanding

Ardent Labs · $65K · Renews Jan 31 · Owner: Daniel
Engagement up. Two new stakeholders joined calls in the last month, both
from a team we had not worked with. On the Aug 6 call they asked about
pricing for additional seats.

Northbrook · $110K · Renews Dec 31
A second business unit was named on the Aug 11 call as wanting the same
setup.

Worth acting on
Meridian is the one that changed this week and it is the largest. Coastal
has been flagged three weeks running with nothing on the record showing a
response, which is a process problem rather than a detection one.
```

### Customize the rollout

* Set the renewal window: 180 days catches it early, 90 keeps the list short
* Define your own thresholds for days since a two-way exchange, cadence change, and support-volume shift
* Watch both directions, or only risk, depending on who reads the channel
* Track whether a flag is new this week or repeating, so a persistent flag is visible as an unanswered one
* Route it to CS, to the account owner's DM, or to a leadership channel
* Chain it with [product feature requests](/docs/use-cases/product-feature-requests) so gaps raised by at-risk accounts read as retention risk rather than roadmap input
* Chain it with [handoffs](/docs/use-cases/handoffs) so a departing sponsor triggers a fresh account brief for whoever picks it up

## Build in Claude or ChatGPT

For the account in front of you, and for reading the whole book before a QBR or a board conversation.

### The Prompt

Open Claude or ChatGPT with the Clearskies MCP connected and paste this (or modify to your needs).

```text theme={null}
Using the Clearskies Context Graph, assess renewal risk across the
accounts renewing in the next two quarters.

For each account, tell me:
- When we last had a real two-way exchange, and with whom
- Whether engagement is rising, flat, or falling against its own history
- Whether any known contact has gone quiet or left the company
- What support activity looks like, including if it stopped
- Anything said on a call that bears on renewal or expansion

Rank by risk. Then separately list the accounts where engagement is rising
and expansion has come up.

Compare each account against its own baseline, not against the others. A
quiet account that was always quiet is not the same as an account that
just went quiet.
```

The last instruction is what separates a useful list from a list of your least chatty customers.

### Make it yours

* Add `Just Meridian Systems, in detail, with a timeline` when one account needs the full picture.
* Add `What would you do about the top three` to get to action rather than diagnosis.
* Add `Which of these look like the accounts we lost last year, and what did those have in common` when you have enough churn history to pattern-match against.
* Add `Draft a re-engagement email to the champion` when the answer is that somebody needs to reach out today.

### The Skill (Claude)

The `clearskies:account-brief` skill in the Clearskies plugin for Claude generates a full account brief from the Context Graph, across CRM, calls, email, calendar, and Slack. It is the right follow-up once an account is flagged and you need the whole story before deciding what to do.

<Tip>
  Want this packaged for your team? [**Reach out**](mailto:support@clearskies.cc) and we'll help you customize the skill to your workflow.
</Tip>

## The Plugin

**Ready to make this part of your team's workflow?** We'll set up the Clearskies plugin with you. Clearskies workflows, Claude skills, all tailored to how your team works.

[**Book 15 minutes →**](https://calendly.com/pouyan/new-meeting)

## Next steps

1. [Sign in to Clearskies](https://app.clearskies.cc/)
2. Connect your data sources
3. Get your [Clearskies MCP server](/docs/building-with-clearskies/mcp-server) and try with [Claude](/docs/building-with-clearskies/mcp-clients/claude) or [ChatGPT](/docs/building-with-clearskies/mcp-clients/chatgpt)