Business intelligence and web site analytics are playing an increasingly pivotal role in decision making, providing much needed evidence to support operational-critical choices and strategies. However, organisations often ignore one of their biggest resources of genuinely actionable customer intelligence, their virtual assistant.
The chat log files of a virtual assistant are a rich source of data provided by customer interactions, but only if they can be accessed fast enough will they have a positive effect on an organisation’s bottom line. Unfortunately with large data sets that’s rarely the case.
An impressive amount of data can undoubtedly provide solid information, but it is often a case of too little, too late, by the time everything has been analysed. It is no use identifying a key issue weeks or months after it was upsetting customers - the damage to sales, profits and/or reputation is done. Sometimes irreversibly so.
However, by using the natural language technology that powers your virtual assistant to analyse the data, free format customer conversations can be interpreted and then automatically tagged and categorized for further analysis in much shorter timescales.
This provides a deeper knowledge of customers needs and purchasing requirements that can be fed back into an organization’s strategy and enable businesses to react quickly when errors occur. For example, the statistics might show that customer complaints are up, but it is the analysis of free format text that quickly shows that it’s because of a packaging or manufacturing error.
Analysing data in this way also makes it more flexible. The business environment is constantly changing, what is important to measure and know today, is likely to be completely different the following week and in some cases even the following day. By having a key set of data that can be analysed cross-dimensionally, organisations can splice the information easily to retrieve the knowledge they need.
It’s not just customer service that can put this data to good use. Marketing can see the best time for promotions based on when questions are most asked about a service, product management can see features that customers are asking for that aren’t yet available and sales can understand in more detail what made the customer abandon their cart.
Since the customer is no longer just click and pointing through a website, but asking specific questions in their own words, it is possible to have a greater understanding of what the customer is thinking, something that general web statics do not show. Adopting a strategy of combining both statistical and contextual analysis will deliver the type of insightful information that can make a difference to an organisation’s success, particularly in an environment where decisions must be made quickly.
Virtual assistants may have started out life to provide help to overstretched call centres, but they are quickly becoming an organisation’s most knowledge, flexible and valuable employee.
About Andy Peart
Andy Peart is CMSO of Artificial Solutions, an international business that “Makes Technology Think,” believing that people should be able to interact with technology intelligently, using their own voice, in their own terminology and across whatever channel they choose.