Skip to main content
Customer Interaction Management needs EDM

(Posted by guest blogger, James Taylor)

It seems to be customer interaction week (month) on my blogs. Today I finished reading an article that prompted this post - Service-Oriented Customer Interaction Management by Ashutosh Roy of eGain.  There was a lot to like in the article but it seemed to me that a focus on the automation and management of decisions was a glaring omission.
The first line of the article is:

Brands are reinforced every day with thousands of well-executed interactions   with valued clients

Now there's a sentiment with which I have to agree most strongly. Indeed, this reinforces the key premise of enterprise decision management or EDM - that little decisions add up and that a focus on the automation and improvement of these operational decisions is critical to the effective execution of a business. This is particularly true in the case of customer treatment decisions where an approach often called "best next action" is used to assess what will best develop the customer next time there is an interaction and then deliver the selected interaction across whatever channel makes sense.

Ashutosh goes on to make some good points about the value of a more component-based, service-oriented platform. In particular he talks about the problem that monolithic CRM systems tend to capture business process

in UI instead of workflow, knowledge, and reasoning

Again I think this hits the nail on the head. Capturing business process, assuming a broad definition of business process that is more than "that which I can store in a BPMS", in its components is absolutely the right thing to do. The process and flow elements get stored as workflow and activities in a BPMS, the knowledge is stored as reusable and accessible content elements and reasoning... While Ashutosh does not really discuss reasoning explicitly, reasoning should be stored as business rules and analytic models that automate decisions. These decisions should be implemented as decision services (to fit with the other service-oriented components). Reasoning, in a customer interaction sense, must contain rules that are defined by regulation, by policy, by expertise and by the customer's preferences. It must also contain segmentation models so that different kinds of customers can be treated differently and it must use what is known about a customer to make useful predictions. Reasoning, in other words, is implemented using EDM. Regular readers will know that EDM also emphasizes adaptive control - the constant testing of "champion" approaches against potentially more effective "challengers" - to ensure that the way customers are treated is not allowed to become stagnant.

For this new architecture, he identifies 6 requirements:

  • Open BPEL
  • Open knowledge architecture
  • Open, first-class data exchange interfaces
  • Virtualized logical data views across repositories
  • Componentized, lightweight UI widgets that can be deployed anywhere
  • Packaged best-practice templates to bootstrap operation

To which I would add:

  • Externalized, managed decision services
  • A rule architecture that combines customer and company rules
  • Adaptive Control to constantly test and refine new approaches
  • Rapid deployment of executable analytics derived from all the data available

There's lost of posts on Customer Service on this blog and you might enjoy these particularly:

I also posted a series of summary articles about using EDM to improve customer service over on my other blog. Check out these links:

Finally you might enjoy this white paper I wrote called "Smart-Enough Customer Decisions"


Visit my Smart (Enough) Systems Blog(RSS) or my ebizQ blog (RSS). Buy the book or visit the companion wiki.

related posts