With so much of customer service now automated through call centres and online interaction, consumers expect their dealings with businesses to be reliable, predictable and professional – if lacking in any personal touch. But those raised expectations are putting pressure on organisations to raise the bar when dealing face-to-face with customers.

That was a situation that telecoms giant BT wanted to address when trying to streamline the systems that manage the hundreds of thousands of customer home visits made each year by its 28,000 field service engineers.

Customer satisfaction is influenced as much by factors such as traffic delays, fluctuations in demand and the availability of engineers and equipment as it is by the quality of the service itself.

Failure to do so accurately can often mean that customers, who specifically set aside their own time to be available at a given hour of the day, have had that time wasted – a cardinal sin for any business to commit.

Prior to a joint development project with Indian IT services provider Infosys, BT did have some bespoke systems that helped it to make such predictions. But the fragmented nature of these, says Gilbert Owusu, technical group leader at BT’s Intelligent Systems Research Centre, made them brittle and unreliable.

What was needed was a system that could aggregate all the pertinent information, make more accurate predictions and allow those to be worked into the business processes already in place.

The key to that, Owusu and his team decided, was to build an application platform on which the system would sit. And to speed the development, they called in Infosys, not only because it could provide the appropriate development resources, but because it could apply its own patented RADIEN application framework to the project.

“The RADIEN framework abstracts lots of the technical aspects of using J2EE (the Java development platform),” explains Owusu. “Things that would take you days can be done in a matter of hours.”

What Infosys built was a series of application services, including FieldPlan, which enables managers to design resource allocation plans, and FieldSchedule for planning engineers’ appointment schedules according to their location.

But at the heart of the Field Optimisation Solution’s (FOS’s) service-oriented architecture is the FieldForecast system, which predicts demand by applying decision support algorithms to data from BT’s customer relationship management system.

The service-oriented nature of the application framework ensures that employees in different roles – from field engineers and local managers to schedulers – can use an appropriate blend of available services.

BT also recognised that regional managers’ local knowledge was just as important to the success of customer visits as anything they could automate. As such, the FOS allows managers to decide whether to automate reservations according to the prediction algorithms or to override these to cater for other considerations.

Since launching in the summer of 2006, the FOS has brought about a 10-minute reduction in the average journey to an appointment, and hence a significant improvement in productivity.

According to Owusu, more accurate allocation of resources has also made it more likely for issues to be resolved on the first visit, which, along with improved punctuality, has boosted customer satisfaction.

Further reading

Process report BPM success requires a deep understanding of the underlying technology and processes, say early adopters.

Find more stories in the SOA & Development Briefing Room