Back to the Future in Customer Service

This week’s report is a product evaluation of V-Person, the virtual agent offering of Creative Virtual, a privately held supplier that was founded in 2003 and is based in London, UK with offices in the US, India, the Netherlands, and Australia. The report updates our September 19, 2012 evaluation. Since that time, Creative Virtual has made significant, attractive, and very useful improvements to V-Person. Personalization is the most significant, most attractive, and most useful set of capabilities.

Personalized Customer Service

Creative Virtual takes the familiar rules-based approach to personalization, matching customer profile attributes with attributes of knowledgebase answers. Personalization V-Person is the first virtual agent offering to implement personalization. This is a big step forward for virtual agent technology and a big step for customer service.

As we stated in our report, virtual agents that can deliver personalized customer service are especially attractive for account management scenarios in financial services, government, healthcare, telecommunications, and travel. JP Morgan Chase Bank is the first user of V-Person’s personalization. Here’s a screen shot. Note that Richard Simons is CEO Creative Virtual USA.

jpmc&co copy

Rules-Based Personalization

We first saw this rules-based personalization approach in ecommerce systems back in the late 1990s. Rules executed at runtime matched customer/account profile attributes with product attributes to select and deliver content. Ecommerce systems stored and managed all of this account and product data. The stored and managed web content, too. The challenge for analysts and administrators was to minimize the number of rules processed. Too many times, in their zeal to deliver exactly the right content, they would ask the ecommerce systems to execute dozens of rules for every customer interaction. The time needed to process the rules might slow response time so much (Remember that this is 1990s era processing speed and power.) that customers frequently abandoned leave the site. Analysts learned this lesson and reduced the number of rules. Personalization was still pretty good.

Customer service applications have a different challenge to perform rules-based personalization. While their knowledgebase answers or cases/tickets have plenty of predefined attributes, they do not store and manage customer/account data. The customer/account profile attributes needed for personalization are typically stored and managed in CRM systems, billing systems, or account management systems—systems external to customer service applications. Application integration is required to collect the customer profile attributes that personalization rules use to select content.

Customer service applications, particularly virtual agent applications, have not been so good at application integration. Most are missing high-level integration tools and/or packaged integration facilities. (Note though that IntelliResponse, whose IntelliResponse Virtual Agent is among the leading virtual agent offerings, just announced packaged integration with Salesforce Service Cloud. Major progress there.) Some don’t even expose web services or lower level APIs. That makes customer service application integration, the application integration necessary for personalized customer service, low-level programming work, work for the consultants of customer service application suppliers. That’s not a very attractive, very repeatable, or very cost effective approach.

Collecting Customer/Account Profile Attributes in V-Person

Creative Virtual offers two approaches for collecting customer/account profile data. The first is that low-level programming integration. Python scripting is V-Person’s integration mechanism. The second is cookies. Cookies can store the customer/account profile attributes collected by Python scripts in the first approach or analysts can use V-Person’s facilities to collect these attributes through a (virtual agent) question and (customer) answer dialog and then pass the data to web developers to set the cookies.

When use of virtual agent applications requires logging in to a customer support site, and it should for any account management activity, then online collection of customer profile attributes is a decent approach. Using cookies to store them has some advantages but, in this time of high sensitivity to privacy and security, certainly some disadvantages, too.

Back to the Future for Application Integration, too

Through its personalization capabilities, Creative Virtual has broken new ground and raised the bar for customer service. Props to them. The firm’s customers will be able to deliver a better customer service experience to their customers. That’s a terrific first step.

The next step has to be the application integration improvements to make the implementation of personalized customer service faster, easier, and more extensible. Programming by consultants within professional services engagements with customer service suppliers won’t cut it for the long term. What’s needed is exemplified by IntelliResponse’s new packaged integration with Salesforce Service Cloud—packaged integration with the leading CRM and account management products and higher-level integration mechanisms for integration with custom CRM and account management apps. Like rules-based personalization, packaged integration is widely used and well proven. Application integration standards have been around for years. Customer service application providers should go back to the future again.