Many banks are turning to customer experience analytics to rapidly analyse customer conversations and automatically identify those needing a more urgent response.
AI-based technologies can now remove the need for complex models and instead, analyse customer interactions in real-time.
Most complaints are handled by contact centres, with a dedicated team triaging, routing, queueing and then handling the complaints. Complaints can be highly complex, e.g. there might be multiple issues (the average is over two per complaint) and different customers and issues require different treatment and validation.
Being able to accurately handle and predict unhappy customers is a very hard data science problem. It is made even more complex when the nature of complaints changes as new products and services appear and recede, as well as spikes in incidents and trends.
Machine learning models have traditionally been ineffective as they constantly require refreshing because of new and changing signals in the dataset and if the models are becoming out of date quickly then it is hard to reallocate the staff doing the triage manually.
All the while, the speed of the handling has a significant impact on NPS. With all these increasing demands, how do you triage your most valuable interactions? Do you keep adding more staff to your contact centres to triage, classify and handle the complaints?
Automatically identify and fast track customer complaints
With AI interaction analytics, you can send complaints straight to the relevant team for a faster resolution. We’ve helped banks reduce resolution time by up to 3 days which really boosts customer retention.
Complaint handling is traditionally laborious, slow and inaccurate. Much research has identified direct correlations between the speed of handling a complaint and the emerging satisfaction and also the associated compensation costs.
The key aspects to improve customer retention with complaint handling are speed of problem resolution; taking a proactive approach; and the communication of next steps in the process.
When it comes to speed of response, dealing with specific complaints assigned with certain criteria can improve response rates dramatically. However, doing this manually simply involves using more and more case handlers.
Routing complaints automatically and prioritising by issue and category is also difficult due to the nature of complaints i.e. unsolicited, long and sometimes multi-topical. As a result, manual classification is often impossible within an acceptable time frame for the unhappy customer.
By using the latest AI however, it is possible to automatically classify unstructured data such as text and provide an early warning for issues that need resolving fastest.
This can lead to better and quicker outcomes at a much lower cost. Banks can handle complaints through a multitude of channels, whether they come through email, over the phone, logged in an online support portal, or through live chat. When you fully utilise AI analytics it can organise all your complaints into highly accurate clusters and categories making certain that they are dealt with expeditiously.
This effectively means that you can free up significant human resources. Humans no longer have to triage incoming conversations one-by-one, or even checking each AI classification.
They are only checking the records brought to them by the attention of the AI. In a bank, this can be 5% or less of the personnel involved, meaning that 95% of the staff can be redeployed saving cost, but more importantly turnaround time which can drop to less than a quarter of the original time.
A leading FinTech SME was experiencing hyper-growth and looking to improve its FCR (First Contact Resolution rate). Its primary channel was chat as it appealed to a certain demographic who preferred that medium. Within a couple of days, their customer experience analytics had generated a more accurate topic model and emotional intent ‘root causes’ of why customers had to re-contact the company.
After previously trying to improve their model for months without effect, the FinTech business was able to action changes within a couple of weeks to make a massive step change in its FCR. It was able to automatically classify and code their customer interactions, freeing up agents from the lengthy manual ACW (After Call Work) which was not done particularly well and only one tag at a time even if there were multiple issues.
One other benefit was that the business was able to move away from relying on customer satisfaction surveys, which were unreliable, increasingly infrequently completed, and had not proven useful in resolving business issues.
By Dan Somers, CEO of Warwick Analytics