Running Your Renewal Process with an AI Agent

You can run your entire SaaS renewal process through an AI agent by connecting your billing, product usage, and support data to Outlit, then using Claude Code to query that context before every renewal touchpoint. This guide covers five workflows you can set up today: renewal risk scoring, pre-renewal briefs, expansion opportunity flagging, at-risk escalation, and post-renewal health checks. Each workflow includes a ready-to-use prompt template and a complete CLAUDE.md you can drop into a project.

Renewals break down because the context is scattered

The CSM running your renewals is supposed to know everything about the account: whether they're using the product, whether support has been clean, whether billing is healthy, whether they've grown into a bigger plan, and whether anything went wrong in the last quarter. That context exists. It's in Stripe (billing and subscription data), PostHog or SDK events (feature adoption and usage trends), Pylon (support history), Slack (internal account discussions), and Fireflies or Granola (call records from the last QBR).

Nobody pulls all of it before a renewal conversation. The CSM checks Stripe for the renewal date, maybe skims the last support ticket, and goes in with whatever they remember. If the account has been healthy, that's usually fine. If the account has been quietly disengaging for three months, the CSM finds out during the renewal call, not before it.

The problem gets worse at scale. A CS team managing 200 accounts can't do deep renewal prep for every one. Prep happens for the enterprise accounts and the rest get a templated email. The accounts that churn are usually the ones nobody looked at closely enough.

An agent connected to Outlit changes the economics of renewal prep. Outlit unifies billing, usage, support, and conversation data into a single account profile. The agent queries that profile 90 days before renewal and every week after. It scores risk, flags expansion signals, and produces a brief before every renewal conversation. The same depth of prep that used to be reserved for top accounts now happens for every account on the book.

By the end of this guide, you'll have a CLAUDE.md and five prompt templates you can copy into a Claude Code project and run today.

What the agent is actually doing

The agent is not replacing your CS team's judgment. It is pulling together context that already exists across your tools and presenting it in a form that makes renewal conversations more informed. Outlit connects to Stripe for billing and subscription data, PostHog or SDK events for product usage, Pylon for support history, and Slack for internal account context. Where call records are connected (Fireflies or Granola), the agent can also reference what was discussed in the last QBR or check-in.

When the agent queries Outlit for an account approaching renewal, it receives a unified profile: billing status, usage trends, support ticket history, and conversation context. It reasons over that profile to produce the specific output each workflow requires, whether that's a risk score, a prep brief, or an escalation alert.

How the renewal agent runs

Every workflow follows the same pattern. The agent reads its instructions from the CLAUDE.md, queries Outlit for the relevant accounts, reasons over the unified profile, and produces a structured output for the CS team to review.

# Every workflow follows this pattern
1. Agent runs (manual, scheduled via cron, or triggered by a wrapper)
2. Reads CLAUDE.md (role, scope, workflow instructions)
3. Queries Outlit (account ID, renewal window, or full book)
4. Receives unified profile (billing, usage, support, conversations)
5. Reasons over context (what's relevant to this renewal?)
6. Produces output (risk score, brief, escalation alert)
7. Surfaces to human for review and action

Five workflows, explained

How to set this up

You need Outlit with your billing and usage sources connected, the Outlit CLI authenticated, and Claude Code with the Outlit skill installed. Authenticate the CLI with outlit auth login; if it asks for a key, create one in Settings > API Keys. Remote MCP clients use your workspace MCP URL from Settings > CLI & MCP and complete OAuth in the client.

Step 0: Install the Outlit CLI and authenticate. Run the CLI login flow and let the CLI store the credential.

curl -fsSL https://outlit.ai/install.sh | bash
outlit --version
outlit auth login

Step 1: Connect your data sources in Outlit. Log into Outlit, go to Integrations, and connect Stripe (required for renewal dates and billing data). Then add PostHog or SDK events for product usage, and Pylon for support history. If your team records calls, connect Fireflies or Granola for QBR and check-in context. Each source you add makes the renewal workflows more accurate.

Step 2: Install Claude Code. You'll need Node.js 18 or higher.

npm install -g @anthropic-ai/claude-code
# Verify install
claude --version

Step 3: Install the Outlit skill for Claude Code. This gives Claude Code the Outlit workflow guidance and CLI commands it needs.

outlit setup claude-code
# Verify auth and agent setup
outlit doctor

Step 4: Create your project and add the CLAUDE.md below. Make a new directory for your renewal agent, drop in the CLAUDE.md, and you're ready to run.

Step 5: Run the renewal risk scoring first. Pick the weekly risk scoring workflow as your starting point. It covers all accounts in the renewal window at once, so you'll see immediately whether the data sources are connected correctly and the output quality matches your expectations.

Step 6: Schedule recurring workflows. Use cron or your preferred scheduler to run the risk scoring weekly. The pre-renewal brief runs on demand before each conversation. The escalation and post-renewal workflows can be triggered by a wrapper script that monitors account status changes in Outlit.

What changes when this is running

Every account approaching renewal gets the same depth of analysis regardless of size. The risk scoring catches accounts that are quietly disengaging weeks before the renewal conversation, not during it. Your CS team stops being surprised by churn they could have seen coming.

Pre-renewal briefs become a habit because the output is immediately useful and saves more time than it takes to read. The CSM walks into the conversation knowing what was said on the last call, what the support history looks like, and what the usage trend has been. The customer notices the difference.

Expansion opportunities stop being left on the table at renewal. The agent identifies accounts that have grown into a bigger plan before the CSM has to guess. Renewal conversations that could have been flat become upgrade conversations, backed by data the customer already generated.

At-risk escalation means the CS manager hears about problems before the renewal window closes, not after. The account that went dark three weeks before renewal gets attention from a senior person with enough context to act. The escalation is specific enough to be useful: not "this account is at risk" but "their VP Eng asked about self-hosted, was told no, and usage dropped 62% since."

Post-renewal health checks catch the accounts that renewed reluctantly. A team that renewed but immediately disengaged is telling you something. The agent catches commitments that were made during the renewal conversation and never followed through, giving the CSM a concrete action before the next quarter.

Get started with the CLAUDE.md

Copy the CLAUDE.md below into your project to get started. It contains the agent's role, instructions for querying Outlit, and all five workflow definitions. Below it are individual prompt templates you can run as-is, with variables to swap out per account.

# Renewal Process Agent
 
## Role
You are a renewal operations agent. Your job is to monitor accounts approaching renewal, assess risk, identify expansion opportunities, and produce briefs and alerts that help CS teams run informed renewal conversations. You do not contact customers directly. All outputs are internal. The CSM or CS manager decides what to share externally.
 
## How to use Outlit
Before every response, use the installed Outlit skill and CLI to fetch context for the relevant account or account list. Do not rely on prior conversation context or general knowledge about a company. Fetch current data on every request.
 
Outlit returns unified account profiles including: billing data (Stripe), product usage (PostHog or SDK events), support history (Pylon), internal context (Slack), and call records (Fireflies or Granola) where connected.
 
If a specific data source is missing or not connected for an account, say so explicitly. Do not invent context or assume health based on incomplete data.
 
## Output principles
- Be specific. Reference actual usage numbers, actual ticket IDs, actual billing states. Generic summaries waste time.
- Be brief. A risk score should fit in a Slack message. A renewal brief should fit on one page.
- Distinguish signals from noise. A single missed login is not a signal. A champion's login frequency dropping from daily to zero over three weeks is.
- Cite your sources. Every claim should reference whether it came from Stripe, PostHog, Pylon, Slack, or a call record.
- Never contact a customer or prospect directly in any form.
 
## Workflow instructions
 
### Renewal risk scoring (weekly)
1. Query Outlit for all accounts with renewals in the next 90 days
2. For each account, pull: billing status and payment history (Stripe), usage trend for the last 90 days (PostHog or SDK events), open and recently resolved support tickets (Pylon), any internal Slack context
3. Score each account as high, medium, or low risk
4. For high and medium risk, cite the specific signals driving the score
5. Produce a prioritized list sorted by risk level for team review
6. Flag any account where data sources are missing as "unable to score, needs manual review"
 
### Pre-renewal brief (on demand)
1. Query Outlit for the account: billing, usage, support, Slack, call recordings
2. Produce a brief with five sections: account snapshot (plan, tenure, billing health), usage (trend, feature adoption, key users), support (open tickets, recent support history), relationship (last call notes, stakeholder changes, concerns raised), watch for (what could come up in the renewal conversation)
3. Keep it under 300 words. Produce the brief for the CSM to review.
 
### Expansion opportunity at renewal (triggered)
1. Run this when an account enters the 90-day renewal window
2. Query Outlit for usage data, current plan limits, and feature adoption
3. Identify signals: usage growth exceeding plan limits, new team adoption, feature usage suggesting a higher tier
4. Produce an expansion brief with: signal summary, usage detail, opportunity sizing, one suggested talking point grounded in their actual usage
5. Produce the brief for the CSM and AE to review
 
### At-risk renewal escalation (triggered)
1. Run this when an account within 60 days of renewal shows: usage drop >30% in 30 days, multiple unresolved support tickets, champion inactive >14 days, billing failure, or explicit cancellation mention in support or call recording
2. Query Outlit for full account context
3. Produce an escalation alert with: trigger signal, supporting signals from other sources, call context if available, recommended action
4. Produce the escalation alert for the CS manager to review, not just the CSM. This bypasses the normal weekly report.
 
### Post-renewal health check (30 days after)
1. Run this 30 days after an account renews
2. Query Outlit for usage comparison: 30 days pre-renewal vs 30 days post-renewal
3. Check whether any commitments made during the renewal conversation (from call recordings or Slack) have been followed up on
4. Produce a health check with: renewal outcome, usage comparison, commitments check, flag if anything needs attention
5. Produce the health check for review. If usage dropped >15% post-renewal, escalate to the CS manager.
 
## What you should never do
- Contact a customer or prospect directly in any form
- Score an account as low risk when data sources are missing or incomplete
- Share escalation alerts with the CSM before the manager has reviewed them
- Skip querying Outlit and rely on your own knowledge of a company
- Invent usage numbers, ticket counts, or billing details

Prompt templates

Variables in gold are yours to swap out. Everything else runs as written.

Run the weekly renewal risk scoring.
 
Query Outlit for all accounts renewing in the next [90 / 60 / 30] days.
 
For each account, pull billing status, usage trend, open support tickets, and any Slack context. Score each as high, medium, or low risk. For high and medium, cite the specific signals.
 
Output a prioritized list sorted by risk level. Produce the report for [CS manager name] to review.
 
Do not action anything without my confirmation.

Prepare a pre-renewal brief for [account name].
 
Renewal date is [date]. CSM on the account: [CSM name].
 
Query Outlit for everything we have: billing, usage, support, Slack, call recordings. Produce a brief covering: account snapshot, usage trend, support history, relationship context, and what to watch for in the renewal conversation.
 
Produce the brief for [CSM name] to review. Keep it under 300 words.
 
Do not action anything without my confirmation.

[Account name] is entering their renewal window (renews [date]).
 
Query Outlit for their usage data, current plan, and feature adoption. Check whether usage has grown beyond their current plan limits or if new teams are adopting the product.
 
If there is an expansion opportunity, produce a brief with: the signal, usage detail, opportunity sizing, and one talking point for the renewal conversation.
 
Produce the brief for [CSM name] and [AE name] to review if applicable.
 
Do not action anything without my confirmation.

Run an at-risk check on [account name]. They renew in [X] days.
 
Query Outlit for: usage trend (last 30 days vs prior 30 days), open support tickets, champion activity, billing status, and any mentions of cancellation or alternatives in call recordings or support tickets.
 
If any escalation signals are present (usage drop >30%, unresolved tickets, champion inactive, cancellation mentions), produce an escalation alert with the trigger, supporting signals, and recommended action.
 
Produce the alert for [CS manager name] to review. Do not share with the CSM until the manager reviews.
 
Do not action anything without my confirmation.

[Account name] renewed [X] days ago on [date].
 
Query Outlit for usage comparison: 30 days before renewal vs 30 days after. Check whether any commitments made during the renewal conversation have been followed up on.
 
Produce a health check with: renewal outcome, usage comparison, commitments status, and a flag if anything needs attention.
 
Produce the health check for [CSM name] to review. If usage dropped more than 15%, escalate to [CS manager name].
 
Do not action anything without my confirmation.

Frequently asked questions

How do I automate my SaaS renewal process with AI agents?

Connect your billing data (Stripe) and product usage data (PostHog or SDK events) to Outlit, install and authenticate the Outlit CLI, install Claude Code, and run outlit setup claude-code. Then write a CLAUDE.md that defines your renewal workflows: risk scoring, pre-renewal briefs, expansion signals, at-risk escalation, and post-renewal health checks. The agent queries Outlit for unified account context before every output and produces results for your CS team to review.

What data sources does the renewal agent need?

At minimum, billing data from Stripe for renewal dates and payment history, and product usage data from PostHog or SDK events for engagement trends. Support data from Pylon significantly improves risk scoring. Slack context and call records from Fireflies or Granola add depth but are not required to start. The agent will flag when data sources are missing rather than guessing.

How far in advance should the agent start monitoring renewals?

90 days is the standard window for initial risk scoring. This gives the CS team enough time to intervene on at-risk accounts before the renewal conversation. The weekly risk report runs continuously for all accounts in the window. At-risk escalation triggers activate at 60 days for accounts showing danger signals, giving the team enough runway to course-correct.

Can I use this without Claude Code?

Yes. For coding agents, Outlit supports outlit setup claude-code, outlit setup codex, outlit setup gemini, outlit setup opencode, and other documented setup aliases. For remote MCP clients such as Cursor or VS Code, use the workspace MCP URL from Outlit and complete OAuth in the client. Custom scripts can use the API with an Outlit API key.

What if an account does not have usage data connected?

The agent will flag the account as "unable to score, needs manual review" rather than guessing at health. Incomplete data is better than fabricated confidence. Start by connecting Stripe (you need renewal dates regardless) and add usage and support sources as you are ready. Each source you add makes the risk scoring more accurate.

How does the post-renewal health check work?

Thirty days after an account renews, the agent compares usage from the 30 days before renewal to the 30 days after. It also checks whether commitments made during the renewal conversation (captured in call recordings or Slack) were followed through. Accounts where usage dropped post-renewal or commitments went unmet get flagged, because they may not renew next time.