Search

Go Beyond Knowledge Management To Improve Support Experience



For large Enterprises, Data Variety trumps volume while looking for insights. More enterprises are inclined to make Big decisions with data. To do this efficiently, applications need to be intelligent. They should have the ability to:

  1. Accumulate data

  2. Analyze data

  3. Act based on data

The business processes in Enterprise applications so far have been workflow-based. The data is gathered from the workflow to then be utilized to do reporting.


The future of Enterprise applications will be based on

  • Intelligence rather than workflow and

  • Prescriptive rather than reporting


Essentially, Intelligence will drive the workflow and Prescriptive actions will be based on predictions. This is where the top-tier AI systems stop.


There are many tools coming up in the market branding themselves as AI or expert tools. The Customer support market is especially proliferated with Robotic Process Automation (RPA), Chatbots and automation tools. Companies are at a loss on how to evaluate them and know when to use what. Their focus is on Return On Investment (ROI) and the tools don’t specifically cater to the customer needs. This article highlights how to think beyond buzz words and articulates the differences in technology to enable picking the right solution for your support/service teams.


Why Complex Systems Need New Thinking In Customer Support

When products are new, come out as the most simple, why do we need to support innovation? Complexity is increasing exponentially, complex systems fail in complex ways, and complex failures need dynamic and adaptive responses.

  • Complex systems contain mixtures of latent issues: A system’s complexity means that it is impossible for it not to contain multiple flaws.

  • Most flaws are insufficient to cause significant issues. They are regarded as minor factors during operations.

  • The flaws change constantly, due to evolving technology and organizational factors, and even as a result of efforts to resolve existing flaws.

  • Complex systems run as broken: Redundancies in the system, and the ongoing expertise and effort of humans, ensure that the system continues to function, sometimes in degraded mode.

  • Issues have multiple causes, not a single root-cause. Every single flaw is insufficient to cause a major issue. It is the linking of multiple faults that creates the circumstances required for a significant failure.

Innovate Customer Support From Ground Up!

The level of complexity has increased not just because products are complex but also the environment and usage are complex. In this ever-increasing complexity, support teams are focused on three main goals:

  1. Solve today’s problem

  2. Make sure it doesn’t happen again — Reduce customer escalations article talks about aligning customer expectation, empowering front line, prioritize which customers to focus on, communicate

  3. Predict problems before they happen

Whether it is self-service, community, agent assist, or auto-support, Problem prevention and problem elimination are the core. As companies look at ways to do the above, most of them believe incorporating AI will help them. They believe AI to be:

  1. Automate repeated tasks. This is called Automation or Robotic Process Automation (RPA) for short. In this method, you program the repeated tasks and let the computer execute them.

  2. Rules-based and statistics-based tools are an expansion to traditional Business Intelligence (BI) engines.

  3. Machine Learning algorithms that identify anomalies and patterns within data in the enterprise

  4. Identify relevance using Natural Language Processing (NLP)


Ideally, a tool can be called a true AI tool when it utilizes all of the above along with:

  1. Extending NLP to include Content and Intent along with relevance

  2. Extending Machine Learning to provide prescriptive actions

  3. Utilize Tribal knowledge from the users

Tribal Knowledge: Beyond AI in support


Intuition helps only when it is in the area of expertise — Originals by Adam Grant

Value of data as a function of a number of observations in an ML domain, here machine vision. Each vertical line represents the same complexity of a particular problem. Credit: Nicole Immorlica, Microsoft Research


The key to a successful AI journey incorporates utilizing knowledge and learning from humans, getting the feedback loop, and incorporating it into learning. AI is good for complex systems and dimensions of data. Feedback procured from humans on prescriptive actions provided by the tool adds to another dimension. At Ascendo, we believe one of the most important dimensions is every interaction the users are having with the tool. These interactions become critical learning opportunities. A learning engine knows what an agent is looking for, what the data suggests as actions, what the agent decides and how the agent interacts with the tool. It has the ability to not just learn from data, not just learn from feedback from its users but also from user interactions. What we have seen is when we include tribal knowledge from interactions, the prescriptive actions are more optimal than just utilizing data. Essentially, we have expanded the dimensions of decision-making as in the pictures below.



Top 10% of AI tool dimensions: Where AI tools need to be


When someone says AI tool, do ask what do they mean by AI. Following are some leading questions:

1. What role does artificial intelligence play in this tool to help customer support teams?

2. How are relevance and feedback enhancing the Machine Learning dimensions of the tool?

3. Can the tool go beyond these dimensions to identify intent and learn from interactions?