Deep Learning AI for CX and XLA
You might be wondering, what do CX and XLA mean? CX stands for Customer eXperience and XLA is what is known as eXperience Level Agreement. CX is a derivation of UX (User eXperience) and XLA is a derivation of SLA (Service Level Agreement). I bring these two terms into focus because the successful AI systems of the future will be those that provide superior CX and employ outsourced capabilities that will be measured by XLA. UX and SLA are still useful terms, however both don’t have enough nuances to address the more important business driven reasons for deploying advanced AI.
A survey by Forrester that looked into the question of process improvement showed the importance of CX improvement:
The Forrester survey claims, “Today it is customer experience, with enterprises expecting to put top priority on digital automation two years hence.”
Lisa Joyce in “Digital CX Now The Ultimate Factor In The Fate of Banking Brands” writes about the importance of the need for better CX for digital natives. The top three hot buttons that technology savvy customers expect from their financial institution are:
Easy to understand (68%)
Minimal effort to use it (62%)
Speed to complete transactions (49%)
I do my banking with a very large banking institution. Honestly, it has always been pathetic and continues to be pathetic despite all the improvements we have in UX design. The real problem is, to be able to do the above three well, you must have an intelligent interface that can anticipate your needs.
How do we create superior Customer eXperience (CX) with Deep Learning AI? Uber has recently published information about “Improving Uber Customer Care with NLP & Machine Learning”. Uber reported in a detailed experience report how they have employed ML methods to improve their CX. Uber describes a problem resolution system where their customers can submit and track issues with their service. This is a UI that every customer does not want to do, but has to do. By making this experience as painless as possible, the overall CX of the service is improved.
Uber system essentially works by reducing the cognitive load for customer agents by automatically suggesting the top 3 solutions. Uber reports that:
this can reduce ticket resolution time by over 10 percent while delivering service with similar or higher levels of customer satisfaction, as measured by customer service surveys.
Uber’s latest work is in leveraging Deep Learning to further improve on their earlier Machine Learning system:
Using deep learning frameworks, we were able to train our models in a multi-task learning fashion, with a single model capable of both identifying the issue type and suggest the best possible solution.
In the “Deep Learning AI Playbook” we describe this very methodology of identifying ‘cognitive load’ within existing processes. These can be classified in four buckets: lack of memory, not enough meaning, information overload and acting fast. Each kind of cognitive load requires a different Deep Learning approach. However, by approaching development in a systematic way, we can discover the areas that have the highest impact to be improved by DL automation. In the Uber case, what is being addressed is the need to act fast in the resolution of a customer problem.
To remain competitive and to reduce costs, the systems need to leverage external systems rather than the naive strategy of building everything in house. However, how do we ensure the same level of CX when using third party systems? This is where the idea of XLA comes into play. Typical agreements with third party systems covered what is known as a Service Level Agreement (SLA). SLA typically measures availability and responsiveness of a service. However, XLA focuses on the quality of the experience as a measure. These involve interaction measures such as customer satisfaction or the ease of access to services.
AI will likely become the coordinator of the CX and thus it needs to have an appropriate way of ensuring that other services that it employs are of an equivalent experience. This therefore demands an agreement between parties that goes beyond the standard SLA. An XLA is an agreement between parties that explicitly documents the expectations in customer experience that an outsource provider will deliver.
As we progress further towards the maturity of AI based technologies, we shall see these two buzzwords (CX and XLA) propping up more frequently in the conversation. This all falls into an umbrella I call “Conversational Cognition”.
