As part of my ongoing series on Marketing Decision Management products I got an update from NICE about their real-time solutions. NICE was founded in 1986 and has 25,000 customers across 150 countries with nearly $1B in revenue. NICE’s focus is on helping their clients get closer to their customers. This involves knowing the customer (text and voice analytics for instance), engaging employees (workforce analytics for instance), acting in real-time (call center or web interfaces) and closing the loop for continuous improvement. NICE solutions are focused around a customer interaction hub that delivers real-time and cross-channel interactions. The solutions focus on fraud, workforce optimization, call volume reduction, sales optimization, customer retention, VoC, compliance and handle time optimization. Customer touch points supported include contact centers, back office, web, mobile, IVR and in-person channels like branches. NICE has recently acquired Causata (reviewed here) to improve their ability to deliver a seamless customer experience across the web and contact center.
The real-time decision management component under the covers offers real-time guidance in business solution-specific designers. Integrated with relevant parts of the NICE portfolio, the engine covers compliance, service optimization, service to sales (cross sell, up-sell, retention etc), plus desktop tagging and triggering of calls for compliance and quality. Each agent desktop has the real-time decisioning engine that is integrated with web, Windows or terminal applications. This is then integrated with the various NICE applications for workforce management, quality scores etc for additional decisioning inputs and personalization of the guidance provided to agents in real-time.
This portfolio of products is designed, at one level, to optimize sales. One element is the Service-to-Sales module designed to help, for instance, any agent or CSR to interact more intelligently with a customer in real-time during service interactions. This is classic next best action infrastructure designed to execute in real-time. The application can do a range of things during the interaction with the customer. It can determine if a service interaction meets certain business criteria and qualifies for up selling or cross selling. It can flag risks (such as a churn risk) on a CSR screen. It can also start short dialogs.
These dialogs walk a CSR through a series of questions and can be abandoned at any point (if the representative feels like they have asked too many questions for instance). The Service-to-Sales solution then calculates the next best action based on various data inputs such as the interaction context (what is being said, if there is a mention of a life changing event for example), who the customer is, the agent profile and so on. The solution uses various techniques including real-time scoring, predictive scores in the database, external scoring services, business rules, customer segmentation, offer value and more, to prioritize, arbitrate and rank order these offers or actions Each offer has its key selling points, rebuttals and scripts to support the CSR in order to successfully upsell, cross sell or retain the customer. A particularly useful feature of these dialogs is that the system pre-prepares a summary of what was done in the call (and why) for inclusion in the call center system.
To set these up the business owners use a designer. The designer allows a series of flows to be configured or created. These involve steps like qualification, eligibility, guided dialogs, next best offer generation, guided selling and fulfillment. Standard flows are provided and can be edited or new flows entirely can be created. Predictive Analytics developed in standard tools can be embedded in the recommendation approach by configuring business rules, accessing a database (to get scores for instance), calling a scoring service or importing PMML-based models to score in real-time.
These flows use offers, created and managed in a pool. Offers are assigned to the flows for which they are suitable, channels for which they work and a customer category. Offers can have validity periods as well as targeting information such as frequency, time since last offer and priority/ranking. Each segment is assigned a score for the offer allowing it to be targeted on specific segments. Selling points, rebuttals and more can be defined, allowing a complete set of information to be defined in an offer-centric way. The offers can be more or less anything, they are not limited to what you would traditionally regard as a marketing offer.
Within the flows are decisioning steps. These select the relevant offers and are controlled using a dashboard. This allows business users to manage the offer presentation approach. They can use a slider to control the number of offers and a dial to balance propensity (customer characteristics that make them likely to accept an offer) and arbitration (company business rules, policies). 7 categories can be given a weighting for the algorithm too – the weight given to business rules, predictive models, segment category, segment offer, agent skills or certification, value and priority can all be varied. This is designed to give business users real control over the offers presented.