Customer support organizations across industries today, regardless of them being in B2B or B2C space, are going through a rapid transformation. End-user and enterprise customers-alike are demanding a much faster response time than ever, to the point that can only be serviced by an automated process. Reducing turn-around time (TAT) and mean-time-to-resolve (MTTR) along with the total cost of operation (TCO) while maintaining CSAT has been the highest order imperative for support organizations.
Moreover, since companies that are required to provide support to their customers receive feedback & queries in simple text across multiple online & social-media channels, and receive raw-machine logs from their own install-base of equipment, it is inevitable to extract insights from unstructured and semi-structured data and it comes through from each of these channels. Add to this the fact that most software companies have transition their product development methodologies to Agile + DevOps; this seldom enables end-users to face issues with the product sooner than the support organizations are trained to answer them.
Fortunately, with advanced natural language processing (NLP), machine learning & deep learning capabilities increasingly being available from service providers as well as data platform providers, it is getting easier by the day to fire-up intelligent applications in record time. Moreover, since the turn of this decade, Big Data technologies based on the Apache & Hadoop ecosystem have evolved to a point that useful information from these distinct datasets can be extracted either in motion or at rest and rapidly turned into useful actions for support organizations.
That said, the key ingredient to make any support process efficient regardless of the underlying technologies used, is contextualization. And building the right contextualization requires the right approach to extract intent and build correlations thereof.
Cognitive analytics can be termed as a combination of NLP, context & intent extraction and machine learning. Cognitive Product Support can therefore be considered an application of cognitive analytics to product support.
Although it is fair to say that many organizations have begun to invest heavily in revamping their support processes, a majority of legacy as well as modern organizations are still looking to leverage today’s technologies to re-structure their support process and automate in order to bring down the total cost of operations (TCO) to drastically low levels.
Automation strategies can be implemented in online channels as well as offline channels. There are several ways to not just respond to simple queries directly to customers using auto bots (i.e. with no human intervention), but also make information available through intelligent search capabilities built right into the FAQ sections. Now, when none of these help and a call is made or email is sent to an agent for help, automated resolutions can be presented to her instantly based on the context of the problem and a deep analysis conducted on the knowledge repository. As an example, it is generally believed that more than 30% queries for consumer-centric organizations can be resolved through intelligent automation.
A truly cognitive product support solution would make the “right resolutions” for a given customer problem available at the “right time” for immediate consumption and ticket resolution. And this can only be brought to bear if you have all the three tenets of cognitive analytics working for you in real-time.
To know about the far-reaching potential of this revolutionary phase of product support, read our whitepaper.