Call Centres Of The Future
Having established itself as the call centre capital of the world, salaries are now increasing in India, attrition rates are very high, and infrastructure costs are rising. So how does the industry stay competitive? Call centres are now embracing technology to stay ahead.
Hi, this is Ruchi. I am calling from Moneymoney bank, we have a pre-approved personal loan of Rs. 40,000 for you, madam…” Does that sound familiar? How many times have you received such a call? Do you immediately hang up; do you humour the agent and carry on conversing; abuse the poor agent… or end up buying the product that she is selling?
A ‘call centre’ is a generic term for help desks, information lines and customer service centres. Call centres have become the central focus of many companies, as these centres stay in direct contact with the firm’s customers. Call centres are of two basic types: those that handle inbound calls, where customers call in for service, and outbound call centres, where agents from contact centres call customers to offer services. Call centres not only offer phone-based support, but also support through online chat, SMS and e-mail Microsoft MCTS Training.
Call centre to the world
India, today, is the call centre to the world. Companies outsource their call centre operations because that way they can get access to skilled and expert staff without having to worry about recruiting them, training them and retaining them. India has both the skilled manpower and the lower costs that make it the preferred destination.
But running call centres is a highly competitive business. The same reasons that make India attractive could, in turn, make some other location attractive in the not so distant future. In India, salaries are increasing, attrition rates are very high, and infrastructure costs are rising. So how does the industry stay competitive? Call centres are embracing technology to stay ahead.
Computers as interviewers
The call centre industry in India is growing at the rate of 40 per cent per annum. However, attrition rates are as high as 30 per cent. This means that lots of people need to be hired. For every one person that you hire, you need to interview ten people. So the cost of hiring one person can be as high as Rs 20,000 if the screening cost is included. The cost of interviewing and hiring a thousand agents is itself a couple of crores. Large call centres hire a few thousand new employees every year, so the cost of interviewing is a major expense for them. As the call centres move to second-tier cities, it is expected that the ratio of hires-to-rejects will be even higher, making the hiring process even more expensive.
Call centres hope to reduce this cost by automating the first level of screening. Software to test human language skills now exists. The basic language skills needed by an agent include fluency, good vocabulary, good grammar, proper pronunciation and good comprehension. An interviewer marks each candidate based on these skills. Today, speech recognition and natural language processing (NLP) technologies are maturing, and can be used to evaluate candidates on each of these parameters. For example, to test pronunciation, the system checks whether the person is laying stress on the correct syllables in a word. Many Indians pronounce available as ‘avlabel’, and a speech recognition system is able to detect the stress on the wrong syllables. Comparisons between automated systems and human interviewers show that they are usually in agreement about a candidate’s worth, thus paving the way for automated screening processes to be put in place. Now it is only a matter of time before such systems are widely adopted. Very soon, your interviewer may be a computer. Kiosks can be set up in remote corners, and candidates can go there and get interviewed at their convenience Microsoft MCITP Certification.
Good agent, bad agent
Once candidates are hired, they are moved to the floor to take live calls from customers. Agents on the floor are typically evaluated based on their call handling time and the call outcome. The average handling time for a call varies from process to process, and can range from a few minutes to an hour. For example, if you have a banking-related query, you would talk with the agent for about five minutes. On the other hand, if the problem is regarding a hardware or software issue on your computer, the call to troubleshoot it could go on for an hour. The call outcome is measured in terms of specific goals or targets being met. For example, banking loan agents would be assessed on how many customers they were able to convince to take personal loans. Thus, agents are evaluated based on how many calls they took and how many of those calls resulted in successful outcomes. To be viable, a call centre must have its agents perform well on both these parameters.
The message, “Please note that this call may be recorded for training and evaluation purposes,” should sound familiar to you? Most calls to a call centre are recorded and stored. Quality analysts review them to prepare agent-wise and general reports. They randomly select a few calls of individual agents, and then prepare reports on their strengths and shortcomings.
Natural language processing systems are capable of going through a large collection of calls and analysing the reasons for success and failure. Many agents, who were obviously fans of Amitabh Bachchan in Kaun Banega Crorepati, started using his famous phrase on foreign customers “Can I lock this for you?” The foreigners were not familiar with this phrase. A natural language system found that agents who simply asked, “Shall I make the booking for you?” made more bookings. Now, finding this kind of correlation between phrases and call outcomes is something that a human reviewing a few calls is unlikely to make out. But an NLP system analysing thousands of calls ‘sees’ this very easily. Such systems can be used to analyse the language and behaviour of ‘good’ agents, and compare them with the ‘bad’ agents to point out good practices that result in satisfied customers. Call centres are adopting natural language systems a lot more to analyse conversations, and to identify good and positive practices.