What is next best action?
How is the “next best action” determined?
The idea comes from the customer service and sales operations of large financial service providers: instead of sending every customer all the campaigns they qualify for, a decision is made for each customer about what makes the most sense right now. That may be an offer – or the decision not to communicate at all today, because a more important message is running or the contact limit has been reached.
There are three approaches. Rule-based: decision logic that the company defines and that is traceable at all times. Predictive: models estimate which action is most likely to work. Hybrid: rules set the permitted framework, and predictions help with selection within that framework.
The more automated the decision and the greater its impact on the customer, the more important transparency and human oversight become. The PBM Campaign Platform covers elements of this without letting a language model decide on sending: prioritisation of competing campaigns, rule-based selection of arguments and content variants, and AI recommendations for argument, send time and channel as suggestions that people accept or reject.
Marlene Hoffmann would qualify for a reinvestment journey and a credit card offer. Because her investment matures on 31 March, reinvestment takes priority; the credit card offer waits. Within the reinvestment journey, the security argument is chosen because of her conservative risk profile.
Distinction
| Term | Difference |
|---|---|
| Product recommendation | Suggests a product without weighing channel, timing or contact limit |
| Campaign | One measure for many; next best action decides between measures for each customer |
| Hyper-personalisation | Shapes the message; next best action chooses which message |
Related pages
-
Platform
Hyper-personalisation -
Platform
AI with approval -
Blog In preparation
[BLOG: Hyper-personalisation in bank advisory: the adviser as personal sender]