Outcome Agents
Outcome Agents turn a revenue goal — protect renewals, or grow expansion, for one product — into a program FunnelStory runs and measures. An agent studies your past renewals to find which product adoption separated accounts that renewed (or expanded) from accounts that didn't, finds the accounts renewing soon that sit on the wrong side of that line, and tracks whether they move across it. Use it when you want a data-backed answer to "which adoption gap should we close before this quarter's renewals, and for which accounts?"
FunnelStory enables the Outcome Agents page per workspace. When it's on, Outcome Agents appears in the left sidebar. Creating agents, generating cohorts, activating a cohort, and attaching a flow all require permission to edit workspace configuration.
How it works
An agent moves through three stages:
- Draft — you pick a template, a product, and the windows to learn from and act on, then generate cohorts.
- Cohort Analysis — FunnelStory proposes cohort pairs. Each pair is an At-Risk Cohort (accounts renewing in your target window whose adoption matches a group that historically renewed poorly) and a High Retention Cohort (the next adoption level up, which historically renewed meaningfully better among comparable accounts).
- Activated — you activate the pairs worth running. FunnelStory locks the at-risk roster and checks it daily to see which accounts reach the target adoption, which churn, and which are still at risk.
The only template today is Feature Adoption.
What the analysis learns from
The agent labels each account's most recent renewal decision in the Training Window as a success or not. For a Renewal target, success means the account renewed. For an Expansion target, success means it renewed with ARR more than 10% above its previous amount, so an ordinary price increase doesn't count as expansion.
Candidate signals come from the product you chose:
| Source | What's measured |
|---|---|
| Activity models that record activity for the product | Activity within the lookback window, and all-time activity |
| Account metrics scoped to the product | The metric's 30-day and 90-day average |
Account metric IDs are shown as readable names on cohort cards, the same way account-metric filters show them. Usage in the 30 days before each renewal decision is ignored, because activity that close to a renewal reflects a decision already being made rather than predicting it.
FunnelStory keeps only the signals that show a real difference in outcomes, compares accounts only against peers that share the traits you grouped by, and requires at least 20 accounts in a group before it measures that group. A pair is proposed only when the higher adoption level renews at least 10 points better and the gap holds up statistically. If no pair clears that bar, the analysis returns no cohorts rather than a weak one.
Readiness
Before you can generate cohorts, FunnelStory checks whether the product has enough history. It needs roughly 100 labeled renewal decisions in the training window and at least two usable signals whose history goes back far enough. When the check fails, the form shows "Not enough data is available to generate cohorts yet." and Generate Cohorts stays disabled.
The pages in this section
| Page | Covers |
|---|---|
| Create an Outcome Agent | The draft form: template, product, windows, outcome target, and how accounts are grouped |
| Read the cohort analysis | Cohort pair cards, needle-mover cohorts, and next best actions |
| Activate and measure | Activating a pair, the Activated tab and report, and attaching a flow |
Related
- Built-in and Outcome Agents — the concept behind these agents
- What's Next — a separate, per-account cohort read on an account's Renewal tab
- Products, Account metrics, and Product activity — the data the analysis uses
- Prevent Churn with a Retention Outcome Agent and Grow Expansion with an Expansion Outcome Agent — end-to-end guides