I grew up (professionally) in Call Centers.

I did all the jobs there were to be done at one time or another: answered phones, played supervisor, managed people, planned for growth, interfaced with business users – and many more you probably don’t want to know or do (if you ever are asked to upgrade a cabling plan, run fast). Back then tools and systems were almost non-existent; certainly not as advanced as today’s solutions. The most dreaded part of working the front line was always (well, when you get past the bad people who like to yell) doing wrap-up notes.

These are the notes we had to write in the customer record following a call to summarize what happened, what they said, what we said and did, and the result. The idea was that if the customer called back we knew what happened. The advent of recording and the evolution of CRM solutions that pull in information from everywhere and automatically record everything the agent and the consumer do across all channels creates loads of transactional, operational, and even feedback information. The wrap up note is a thing of the past (I know they are still being done in some places, but the number of interactions that demand them is minimal compared to every one of them in the past).

To an extent, that is a shame. We used those notes to  gather intelligence about what customers wanted, what worked and what didn’t, and to understand better what we could do to serve our customers better or – at the very least before we were concerned with customer-centricity – more efficiently. The amount of knowledge we could recover from those notes was quite large, and more than one of my jobs included parsing them for knowledge and insights. Think of them as an early form of “Big Data” (there’s the obligatory mention).

Today we are doing the same in a programmatic, automatic manner. We hope that the amount of information we receive from the multiple systems and solutions we implement and query will give us comparable information to what we were able to obtain before: knowledge, insight, efficiency and better performance.

Alas, the question is – does it do it?

Well, yes and no – and it depends (now you feel like you are reading something written by an analyst, right?).

Let me explain.

We generate a large amount of information — more than we ever did before. This information is valuable to improve processes and experiences, for training and management, and even to spot potential problems based on the recorded paths of execution (what agents did, what they found, what they were missing). We can generate a lot of information on performance for efficiency, and feedback for effectiveness reporting.

But we are lagging in generating and managing knowledge (used to provide better answers to customers) from the collected data and information. Knowing who said what, and when, to whom helps if you are in trouble – but not being able to categorize what was said and easily find it to repeat it in a similar situation is akin to throwing it away after it was done. You have to find ways to extract knowledge from the automatic collection of data. If you want to mine all the transactional and operational data, agent notes, and other knowledge from closed service interactions there are three things you have to do:

1.  Reference Your Taxonomy. Sounds silly, but do you know what constitutes knowledge for your organization? What information should be preserved and structured so that we have the information readily available in case of future need? The rule of thumb we always used was that if two people called with the same problem, it should be recorded. Somewhere along the line that got changed to every-single-thing must be recorded and considered knowledge. This is hardly the case, but there are certain aspects that should be preserved. If you have a proper taxonomy in place, then capturing and recording knowledge is simple: use the taxonomy to categorize what you find out (semantic analysis would be great, tagging is sufficient) and to structure the information you parse. Convert the answers to knowledge-based entries, and you can generate knowledge from your new problems. To maintain existing knowledge, the same process applies, but following the categorization you can find the related content that already exists and improve it.

2.  Leverage Text Analytics. There is a very slim chance that you can identify what you need from the many interactions your organization executes every day. To this end, text analytics helps you analyze the large amount of data and find the elements that would need to be categorized to generate knowledge. Knowing what you are looking for (product names, error codes, knowledge repositories accessed and used, etc.) and analyzing the closed interaction data for those specific words and phrases is the way to find what specific interaction should be parsed, categorized, and processed for knowledge. Text Analytics together with usage logs can help you find the areas in your knowledge repositories that may need some work and improvement – and help you collect more knowledge that you thought you had in those closed records.

3.  Recognize Unlikely Sources. This is the toughest one, but the one likely to reward you more. Traditionally we only recognize knowledge as that which generates in our organization. The thought that we know better because we have the best background and resources is a logical approach. Alas, in the last few years we have seen the emergence of Subject Matter Experts and online communities that have become excellent sources of knowledge – in some cases even better than the organization. And while it may not sound right, recognizing those unlikely sources of knowledge and tapping into them to aggregate and curate the organizational knowledge is one of the tenets of the service resolution. Parsing and embedding those external conversations and community-generated content as part of the resolution of service interactions or even making them part of the knowledge generation workflow the organization uses to grow the knowledge-base is essential to deliver better resolutions in the future.

All of this, of course, needs a goal – and that goal should be to increase customer (and employee) satisfaction. Alas, that is the topic for another day. What do you think? Is this something you are doing? Is this something you could do? Share your story, ask your questions, and debate the notions down below – we’d love to hear from you.