Renewal Risk Agent
Names the reason a renewal is at risk, not a score
- 8hard rules
- Things it must never do, enforced in every run
- 5handover rules
- Conditions where it stops and gives the work to a person
- 7output fields
- A structured result, not a paragraph you have to re-read
- ≤7steps per run
- Bounded work: at most 5 tool calls, enforced by the runtime
Read-only by design. This agent has no path to sending, spending, publishing or committing anything. Where that limit is the product, removing it would remove the reason to trust the output.
Overview
By the time a customer says they are not renewing, the decision is old. It was made when the champion moved on, when usage quietly dropped to a third, or when the fourth ticket about the same thing sat unanswered for a week. The renewal conversation is where the decision surfaces, not where it happens.
Health scores were supposed to solve this and mostly do not, because they collapse everything into a number. A customer at 72 tells an account manager nothing to act on, and the number is usually dominated by whatever is easiest to measure — logins — which is a poor proxy for value. A customer can log in daily and be miserable.
This agent reports reasons instead. Usage of the features they actually bought it for rather than usage overall. Whether the person who signed still works there. Support history that repeats. Whether they were ever onboarded onto what they pay for. Silence from an account that used to talk. Each is a different problem with a different fix, and a score hides which one you have.
Every at-risk account comes with the specific evidence, the time left before the notice window closes, and the one thing most likely to change the outcome. It contacts no customer — a renewal conversation is the account manager's, and it should be informed rather than automated.
Benefits
Saves attempted while there is time
Not in the last fortnight, when it is a discount negotiation.
The right save for the right problem
Onboarding, product gap and neglect need different answers.
Honest forecasting
Including which renewals you should not be counting.
What it does
A reason, never a score
"72" is not actionable. "They stopped using the one feature they bought it for in April" is.
Usage of what they bought, not usage overall
A customer can log in every day and never touch the thing that justified the purchase.
Tracks the champion
The person who signed. When they leave, the renewal usually leaves with them, months early.
Reads support history as a pattern
Four tickets on one theme is a product problem the account has already decided about.
Counts the days you have left
Against the notice window, not the renewal date. That is when the choice actually closes.
Contacts nobody
An automated renewal email to an unhappy customer is how a save becomes a loss.
What this changes
What this changes about your week
- It takes over
- Working out which renewals are genuinely at risk, and why.
- Done today by
- An account manager, from memory, usually too late.
- How often
- Quarterly, and the notice window decides whether you had a choice.
- Typical manual effort
- Every account needs usage, tickets, contacts and contract read together.
Why Sales in particular. Health scores collapse five different problems into one number. '72' tells nobody what to do, and the score is usually dominated by logins — which a miserable customer produces daily.
Runs unattended
Started by you or by an event, and it finishes on its own. Nothing waits for someone to be at a desk.
The same standard every time
The two-hundredth item is held to the bar the first one was. Consistency is the part people cannot sustain.
It cannot act on its own
Renewal Risk Agent has no path to sending, spending or committing. That limit is why its output is safe to act on.
How it runs
This agent runs server-side through the PROMIVO runtime. Each run is logged step by step and every tool call is permission-checked before it executes.
- 1InputValidated against the agent input schema
- 2Modelclaude-opus-5 with automatic fallback
- 3Toolsextract_structured_data
- 4PermissionsRuns READ actions automatically; anything higher waits for approval
- 5OutputStructured result matching the output schema
- 6LoggingTokens, duration and every step recorded
Demo
Demo dataIllustrative sample output, abridged.
Input
{
"accounts": [],
"asAtDate": "2026-09-02",
"currency": "USD"
}Output
{
"atRisk": [
{
"risk": "stopped-using-what-they-bought",
"account": "Meridian Group",
"evidence": "They bought this for automated reporting. Report runs went from 40 a week to 3 in April and have not recovered. Overall logins are unchanged, which is why nothing flagged.",
"annualValue": 96000,
"daysToActBy": 24,
"oneThingThatWouldHelp": "Find out what changed in April. A single call to the operations lead, not a renewal conversation."
},
{
"risk": "never-onboarded",
"account": "Cobalt Interiors",
"evidence": "Paying for three modules since signing 11 months ago. Two have never been opened by anyone at the account.",
"annualValue": 31000,
"daysToActBy": 51,
"oneThingThatWouldHelp": "An onboarding session on what they already pay for. Costs an hour."
}
],
"healthy": [
"Northwind: usage up, champion in role, no repeated tickets. No action needed and it is worth saying so."
],
"escalate": true,
"disclaimer": "An assessment from your own records. No customer has been contacted, no discount is proposed, and no account is predicted to churn — these are named risks with the evidence behind them.",
"valueAtRisk": 127000,
"escalationReason": "A 96,000 account is at risk with 24 days before its notice window closes, and one account has never been onboarded onto two modules it pays for.",
"signalsUnavailable": [
"Feature-level usage is missing for 6 accounts; those were assessed on overall usage, which is weaker."
]
}Requirements
Integrations
No integrations required.
Before you start
- Accounts due for renewal, with value, renewal date and notice period
- Usage data, ideally by feature
- Support ticket history
- Contact roles and status
- Requires the All-Access plan or higher
Use cases
Quarterly renewal review
Every account due, ranked by risk and value.
A new account manager
The real state of a book they just inherited.
Board reporting
Retention risk with reasons rather than a chart.
After a product change
Which accounts depended on what changed.
Pricing
All-Access
Billed monthly through your PROMIVO subscription. Cancel at any time.
Runs consume your plan allowance for agent executions and tokens. See plan limits
FAQ
Will it email customers?
No. An automated renewal email to a customer who is already unhappy is how a save becomes a loss. It briefs the account manager, who has the relationship.
Is this a health score?
Deliberately not. It reports named risks with evidence. A score is easier to display and almost impossible to act on, because it never says which of five different problems you have.
What if we do not track feature-level usage?
It uses what you have and states what it could not assess. Overall usage is a weaker signal and it will say so rather than presenting it as the same thing.
Reviews
Reviews open once customers have run this agent.
Need this agent to work differently?
Tell us what to change and our team will scope a customised version for your business.
Customize this agent