One of the most frustrating and disheartening experiences has got to be hanging on the phone, queuing endlessly for a customer service agent. In fact, the anonymity of the situation makes it even less tolerable than standing in a physical queue because at least there you can see where you are in the line.
For large businesses such as banks, insurance firms and telecoms companies, which can field thousands of queries a day, customer service can be a competitive advantage or a millstone, depending on how well they do it. A growing number of organisations are now turning to advanced technology to enable them to enhance their customer service levels while cutting costs.
The concept has become known as customer self-service. As the name suggests, it allows customers to get the information they need without necessarily going through customer service agents — although these are on hand if required. The savings that can result from automating calls are substantial. Industry analyst Gartner recently calculated that using the cost of answering a customer query costs US$1.95 per call using an automated system compared to US$4.50 using a customer-service agent – roughly one third the cost.
Of course basic self-service systems have been around for a number of years. But whereas these simply present users with a number of different options (for example press 1 for x, press 2 for y), the new systems are actually interactive in the sense that the ‘agent’ responds to specific questions a customer may have. The latest self-service technology is underpinned by advanced speech-recognition engines developed by major software firms such as Scansoft and Syntellect. The software learns to ‘understand’ a range of different accents and once it picks up keywords used by the caller, it automatically matches the query with one of a number of prerecorded responses held in a database.
Eircom claims to be the first Irish organisation to have implemented this new advanced form of self-service technology. It is used with its customer support call centre, accessed via the 1909 customer service number.
According to Dervilla Mullan, head of automated channels at the telecoms firm, three factors are driving the uptake of speech-recognition services: cost-efficiency, customer satisfaction and demand for self-service systems.
Eircom is using the system to answer routine customer enquiries such as calls from people wanting to know the status of their phone line order. “Once we launched the service we found that roughly a third of the callers ringing with an order status request did not proceed with their call; they had got the update they needed,” says Mullan.
Eircom also plans to introduce automated fault reporting on the system before the summer. Customers can ring up at any time of the day or night to report a fault, which will be logged and then passed on to the next available engineer to deal with. “Whereas previously it would not have been economical to have agents working around the clock, fault reporting can now be done on a 24/7 basis,” notes Mullan.
The system can be used to re-route calls from the consumer call centre to other Eircom service operations. For example, business customers that mistakenly call the 1909 number are automatically re-routed to the Corporate Service Centre, freeing up the consumer centre from having to deal with these calls.
Currently 10pc of the 100,000 to 120,000-plus calls taken by Eircom’s consumer call centre each month are handled by the self-service agent and the company plans to increase this to 15-20pc over the coming months.
Fexco is another well-known business that is using speech-recognition technology. It recently awarded a €550k contract to IT services firm Datapoint to implement a Syntellect speech-recognition system. This will answer more than two and half million calls Fexco handles on behalf of Western Union every year in its two call centres in Co Kerry. The system has been in trial for several months and is due to go into full deployment in May this year.
Initially, it will only be used to answer routine queries such as requests for listing of money-transfer agencies and not for actual money-transfer requests. But, as marketing services manager Shane McElroy points out, automating the routine queries will make a big difference to the efficiency of the call centres.
“A lot of calls tend to be very repetitive. For example, a number of Western Union customers are Filipino nurses who wish to send money back home. They will want to know the exchange rate on a given day and when the rate reaches a level they are happy with, then they’ll send money. They don’t need to talk to live operators for that type of query.”
What the technology does, he says, is to allow resources to be used more efficiently by giving agents more time to answer detailed customer queries.
Fears have been expressed that speech-recognition systems will eventually do away with customer service agents altogether but McElroy disagrees. “This technology won’t knock call centres on the head; you will still need people.”
He expects that as Western Union’s business grows, the call centres will be able to absorb a lot of the extra call volume without needing to recruit as many new agents as if the speech-recognition technology were not being used.
Some challenges still lie ahead in the adoption of more self-service systems within the Irish industry. For example, there is still a degree of public resistance to self-service systems. It is partly a generational issue according to Mullan, who says in general older consumers will look for personal contact whereas younger users will be happy to use an automated system. “We’re finding a lot of customers still want the security of speaking to an agent so, by using certain commands, callers can be transferred directly to an agent.”
The resistance may also be partly due to the inferior quality of some basic service systems where the options were limited and the desired information not always be readily available. The technology underpinning the new systems is much superior, argues McElroy, who says that Fexco would not be using the technology were it not up to standard.
“We tried using it a number of times in the past few years but found it wasn’t good enough. The Western Union operation has to deal with a wide range of difficult accents spoken by people from a large number of countries. Only now have we found the software is up to speed,” he says.
Mullan is adamant that, despite the fears about the technology, speech recognition is here to stay. “It’s similar to the ATM. It was seen as a bit of an enemy at first but trying doing without it now.”
By Brian Skelly