NLP and Artificial Intelligence
Perspectives / Automating first line support using natural language processing
The situation
The rapid advancement in Artificial Intelligence over the last decade has opened the door to reducing the cost and time delays of getting the right information into the right people’s hands. This paper discusses the benefits, and relatively low barrier to adoption, outlining how to improve the customer experience, executing quickly, effectively, and in-expensively when using NLP & Artificial Intelligence within large Financial organisations, with an example demonstrating the value add for topics such as Basel III, FRTB and LIBOR.
Improving the customer experience can do done quickly and at a low cost when using NLP and Artificial Intelligence
The problem
The tried and tested approach for either a front-line support team or a policy response team is to have some form of hotline / shared e-mail address / submission screen which acts as a point of entry for the staff member that is experiencing a problem. Depending upon the size of the teams involved or the volume of queries drives whether this is a ‘free-for-all’ where the first person to see the mail or hear the phone picks up and answers, or whether a dedicated first-line triage team pick it up and either address simple questions themselves or route the question to a specialist.
Sometimes this whole process is captured in the ‘back end’ by a query manager which co-ordinates the routes the query, provides prioritisation for urgent queries, and provides a host of statistics about a number of queries, average duration to close queries, SLA breaches, and so forth. When the problem is not large enough and is dealt with through just a shared mailbox, detailed statistics are limited and transparency of resolution time and effort is almost impossible to both track and proactively manage.
These traditional approaches help the business to achieve its goals and solve the problems encountered, but they do nothing to stop the problem happening in the first place nor do they operate at pace with multiple handoffs, relying on people to read the email adding lengthy pauses between each response and related action.
In addition, whilst the time taken to get a response adds drag to the efficiency of the overall organisation, the unseen risk is the quality of the information being given. How do you audit policy advice or ensure consistency in a fundamentally people reliant process?
Natural language processing to the rescue
The advancements in NLP (Natural Language Processing) have transformed this problem space allowing full automation of the first line support process. Each of the large cloud providers has a knowledge base tool that can be fed the information documents, helping to build standard question and answer knowledge bases, and which have live connections to other systems. Most importantly, from a user perspective, is the almost instantaneous response to questions where the confidence level is above a pre-set level.
Adopting these automated tools changes the management paradigm. The focus shifts away from response times and volumes to stopping the underlying problems happening that caused the problem and the need for the question in the first place.
Alongside curating the knowledgebase, reviewing the questions being asked by ensuring the answers are accurate, the effort can be directed to root cause solutions such as improving training, system fixes, and streamlining processes which all contribute to the overall efficiency of the organisation.
Dispelling the fear and mystery of Artificial Intelligence
The terms relating to anything AI / Machine Learning / Natural Language processing related carry all the baggage of technical mystery and obfuscation. Combine that with cloud based and the wizardry is squared, the scepticism cubed…but deep understanding of how these technologies work is not required and a strong monitoring and feedback framework will give far more insight, control and business value than can be achieved with manual processes alone.
To de-mystify how this is done, at the very least prove the art of the possible, here at NextWave we have documented the steps required for you to build your own chatbot in 30 minutes on the AWS Kendra platform. Click this link to download a set of instructions in PDF format.
Our view
The role of the knowledge management function now needs be curating the index, monitoring of the accuracy of the responses, and constantly tuning the results. This pivot allows pro-active knowledge management and insight to both the questions being asked.
Additionally, the response to questions is immediate so no wait-time for people asking questions, savings on effort managing tickets and, working on the basis that for every person asking a question ten people are thinking it, a significant organisational capability improvement.
The list price-point of $5k a month (before any corporate discounts) make for a compelling business case, but you do need to factor in the knowledge management and referral handling costs on top , rather than jumping to a straight comparison to any existing costs and processes.
Dispelling the fear and mystery of Artificial Intelligence
The terms relating to anything AI / Machine Learning / Natural Language processing related carry all the baggage of technical mystery and obfuscation. Combine that with cloud based and the wizardry is squared, the scepticism cubed…but deep understanding of how these technologies work is not required and a strong monitoring and feedback framework will give far more insight, control and business value than can be achieved with manual processes alone.
To de-mystify how this is done, at the very least prove the art of the possible, here at NextWave we have documented the steps required for you to build your own chatbot in 30 minutes on the AWS Kendra platform. Click this link to download a set of instructions in PDF format.
Our view
The role of the knowledge management function now needs be curating the index, monitoring of the accuracy of the responses, and constantly tuning the results. This pivot allows pro-active knowledge management and insight to both the questions being asked.
Additionally, the response to questions is immediate so no wait-time for people asking questions, savings on effort managing tickets and, working on the basis that for every person asking a question ten people are thinking it, a significant organisational capability improvement.
The list price-point of $5k a month (before any corporate discounts) make for a compelling business case, but you do need to factor in the knowledge management and referral handling costs on top , rather than jumping to a straight comparison to any existing costs and processes.
