Repeatability
Medium
The analytical structure — cohort segmentation, churn risk scoring, profitability ranking, CCC calculation — is repeatable and well-defined. However, the segmentation logic (how to define 'client types' across retainer, project, and hourly) requires upfront judgment that may shift each cycle.
Ambiguity Tolerance
Medium
The quantitative outputs (CCC, margin by client, payment lag) have crisp success criteria. But 'churn risk' and 'most profitable after service delivery cost' require definitions the user hasn't fully specified — cost allocation methodology and churn signals must be agreed before the agent can know when it's done.
Data & Tool Availability
Medium
The user has 24 months of invoice and payment data, which is the core input. However, service delivery cost data (staff time, COGS per client) is not mentioned and is essential for true profitability analysis — if it's missing or unstructured, the agent's output will be incomplete or misleading.
Error Cost
High
Misclassifying a client's churn risk or overstating profitability could lead to real strategic missteps — dropping a recoverable client, over-investing in a money-losing one, or presenting flawed analysis to firm leadership. Errors here are not trivially reversible and carry reputational risk for the CFO.
Human Judgment Required
High
The recommendations layer — what to actually do about at-risk or unprofitable clients — requires relationship context, strategic intent, and qualitative knowledge about each client that no agent can access. Even the segmentation choices embed judgment calls a fractional CFO is paid to make.