The right customer service at the right time
Customer services to suit the customer
An increasing number of customer service calls and emails are answered by a robot. Real people are there, ready to step in for the more complex questions, but it's becoming harder to know how many are needed and when.
Customer Services are the unsung heroes in a great many businesses. Whenever a major lapse in service occurs, a quick and effective response from Customer Services can help create a reputation for good service and build loyalty amongst the customers affected.
Robots however are increasingly involved. Simple questions are fine, but the moment they become more complex real people need to step in. But a balance has to be struck. Intervening too early will drive up the need for staff. Intervening too late risks frustrated customers who know they are talking to a machine.
Customer Service resource levels are no longer driven by the volume of customer requests alone. With robots handling many of the more routine, they must now be determined by the number predicted to be too complex for a robot to handle. But these can arrive at any time of the day and via any channel. Getting this right requires insight about patterns in demand created by time of day, channel, type of enquiry, prevailing sentiment expressed, length of call, and the need for further assistance from elsewhere in the business.
The Figuringoutdata solution
Figuringoutdata.com can create a solution that employs the best of data analytics methods to predict how many people are needed and the skills and experience required.
Request characteristics are established from historical records that will separate out those that qualify for a machine response. Key phrases will be determined that will trigger a switch from machine to a real Customer Services Agent.
The history of transactions will be used to detect changes in sentiment. Real-time transactions will be monitored against the threshold to establish when a real Customer Services Agent should intervene.
Patterns by time and type
Historical transactions containing triggers for a real Customer Services Agent to intervene will be mapped against historical volumes dimensioned by features such as channel, time of day, proximity to events such as product updates.
Predicting the resources required
A prediction algorithm will be created that uses the historical patterns determined by categorisation, features and sentiment change to predict the number of Customer Service Agents required and the skills and experience they need to possess.
If you would like to know more then please use the form below, or email us at email@example.com