By the time attrition shows up in outflow reports, the relationship is usually already lost. TAZI RETAIN predicts client attrition and the deposits or AUM at risk before assets move, using explainable ML coached by your own advisors. Every score carries a named attrition pattern, and every retention action is validated by an expert panel and a simulated focus group of your client population before it reaches an advisor.
Share of actual churner AUM captured working the ranked list vs a random list of the same size — measured against real outcomes.
Of churned AUM captured working the top tenth of the book.
After expert-panel and focus-group validation — with 27% of the book cleared as 'no action needed'.
At a US community bank — $400K+ in deposits retained per year, predicting churn three months ahead.
A companion explanation model turns every score into a business-named pattern — a reason code advisors can say out loud and compliance can read.
AUM × risk instead of a dashboard. Each name arrives with its drivers, so the advisor knows why before picking up the phone.
Risk trajectory, drivers, talking points and a recommended action in one answer — grounded in the client's history and the named pattern behind the risk.
Your SMEs and policy, encoded as a panel with explicit goals and vetoes. Every sign-off is a logged meeting with the reasoning on the record — panels even challenge suspect data before endorsing a score.
Loyalists, switchers and rejectors — personas carrying real properties from your client population react to each retention action before any client sees it.
Advisors just ask: portfolio attrition risk right now, who needs attention this month, what drives risk, help retaining one client. Every answer shows its data checks on screen — auditable, not a black-box score.
Confirm the attrition definition and the business KPI with advisory leadership; provide anonymized client data; agree the advisor action inventory (call, review, offer).
Train the prediction, segmentation and explanation models on your data; advisors name the segments and correct talking-point tone.
The action-proposal agent designs a personalized retention action; the expert panel applies firm policy and the focus group simulates each client persona — suppressing needless manager alerts.
Review ROI, NPV and compliance; go / no-go; production with monthly monitoring and drift detection.
A US financial institution deployed the TAZI RETAIN Agent and saved $1M/mo in NPV over the first 3 months.
Four numbers you already know. See the revenue TAZI can help you protect by catching at-risk customers early and acting before they leave.
Figures in USD, pre-set for a midsize US company (illustrative — change any field). Outreach is assumed absorbed in existing capacity; only the platform cost is netted, matching TAZI's ROI convention.
Share of customers you lose in a year.
Average annual revenue (or margin) per customer.
Advisors / relationship managers who would act on the recommendations. Used only for the platform-cost estimate.
Of the at-risk customers TAZI surfaces, the share you keep once your team acts. Default 35% is the average across TAZI deployments — replace it with your own holdout result.
Estimated annual platform cost: $186K — auto-estimated as $150,000 base + $1,200 per team member / year, used only for the ROI percentage.
Every input maps to data your teams already hold. Here's where to pull it and how to compute it.
| Lever | Where it comes from | How to compute / benchmark |
|---|---|---|
| Number of customers | Billing / CRM — active-subscriber count at the start of the period. | Count of paying customers on day 1 of the 12-month window. |
| Annual churn rate | Billing cancellations / subscription system. | Customers lost in 12 months ÷ customers at start. Sanity check: median subscription churn was ~4.1% in 2023 (Recurly); retail/telecom runs higher. |
| Yearly value of a customer | Finance — recurring revenue, or contribution margin for a profit-based ROI. | Annual recurring revenue ÷ active customers. Use margin, not revenue, if you want net profit protected. |
| Save rate with TAZI | A holdout / control test once live. | Run a control group: (retention of actioned at-risk) − (retention of untouched at-risk). That lift is your true save rate. Start at the 35% TAZI-deployment average, then replace with your measured number. |
| Team size | Org chart — advisors / relationship managers / retention agents. | Head count that would act on TAZI's recommendations. Drives platform-cost estimate only. |
Illustrative model; figures depend on your inputs and are not a performance guarantee.
Schedule a personalized demo and see TAZI RETAIN Agent's validated actions running on your institution's own data.
See RETAIN Agent in ActionNo commitment. No credit card. Just insight.