Churn Risk Agent

When an account health score drops, the agent starts the analysis straight away. This keeps the underlying changes recent, so Customer Success can see what moved before the signal gets harder to trace.

• Trigger: An account health score drops
• Timing: As soon as the drop is recorded

Next, the agent pulls the account's usage and escalation count. These changes often move together, so the analysis needs both to separate lower engagement from a growing support problem.

• Account: The account whose health score dropped
• Fields: Usage and escalation count

The agent then pulls the account's sponsor and champion. Their engagement carries more weight than usage alone, so the analysis can identify when the health score is falling because key people have disengaged.

• Include: The account sponsor
• Include: The account champion

Using the account data and key contacts, the agent lists every factor contributing to the health score drop and states the size of each change. It includes factors that improved as well as those that worsened, so Customer Success can see the full movement behind the score.

From the contributing factors, the agent names the single factor driving the health score drop. If the factors are unrelated, it says so rather than forcing them into one explanation.

Finally, the agent sends the contributing factors and main driver to the account's CS owner and their lead in Slack. Both receive the signal while the cause is still recent, so it does not depend on one person escalating it later.

• Recipients: The account's CS owner and their lead
• Delivery: Slack message
• Includes: The contributing factors and the main driver