Artificial Intelligence: What every customer experience professional should know

This article is an abridged version of the whitepaper by the same name, published by the CXPA.

The term “AI” has many different definitions. Here is the Oxford definition: 
 
Oxford definition: The theory and development of computer systems able to perform tasks normally requiring human intelligence, such as visual perception, speech recognition, decision-making, and translation between languages.
 
AI has several capabilities as demonstrated in the following chart, based on and adapted from the work of Cellstrat:

How can AI enhance CX?

A recent IBM study highlighted how companies are using AI to enhance CX in three main categories:

  1. Insights. Leveraging AI can help uncover and identify actionable customer insights that drive impactful business decision-making.
  2. Customer Interactions. Integrating AI into various market facing experiences that customers can connect and interact with.
  3. Automation. Deploying AI to efficiently and effectively automate workflow process, saving time and resources.

AI across the Customer Journey

AI is already widely used by organizations throughout the customer journey.

Awareness: Predictive Analytics can be used to introduce product suggestions to buyers or display information to build awareness of related products.

Consideration & Comparison: As websites begin to standardize and integrate data, they are able to facilitate comparisons of products for customers, providing a more informed shopping experience.

Purchasing & Onboarding: Learning about the customer based on data patterns can provide alternatives, upgrades, or other recommendations.

Retention & Support: AI embedded systems can monitor website and in-app activity for distress indicators. This helps to identify the types of issues customers are encountering and respond in real-time through FAQs or virtual service support agents across platforms and devices.

Advocacy & Feedback: Feedback solutions seek to engage customers in a two-way conversation to better gauge their feelings, whether positive or negative. Using Sentiment Analysis, AI can quickly pick up signs of dissatisfaction and take the appropriate action.

Design Considerations

Below are some design considerations originally introduced by Simon Chan, the original founder of PredictionIO.

  1. Is it solving the right problem(s)? There is a big temptation for some businesses to “jump in” to an AI program without first taking the time to evaluate if it is even truly needed.
  2. Who are its users and what are their needs? Chan suggests that to implement AI successfully, you must understand the needs of your users. Consider all potential users of the system, their basic needs, and the typical missions they undertake with current processes to fulfill those needs and add value to the customer’s journey.
  3. What should be “in scope”? Not every user need is going to be met, and many users will have competing needs. After creating a list of needs through detailed research, the financial resources, and technical feasibility will then play a part in narrowing the focus of the final product.
  4. How will you measure success? Define the measures and metrics you will use to define the success of the AI. For example, is it for a bot to be able to handle a certain percentage of web inquiry on its own?
  5. How will you govern the evolution of the AI? Change management will play a huge role as technologies continue to grow exponentially in power over the next few decades. A proper steering committee must be put in place so that the AI is able to adapt and change as the business needs of customers evolve over time.

Implementation Guidance

While there is plenty of support for implementing AI, many organizations either view AI as the newest shiny toy, or the “silver bullet” to solving all their CX woes. They risk going “all in” without formulating an AI strategy. This will inevitably impact alignment to overall organizational goals, which create inconsistent experiences and rework of processes. 
 
Before engaging AI, organizations should ask themselves three key questions:

Recommendations

The overwhelming amount of AI information and options may appear to be quite complex to understand. A clear view of the problem you are trying to solve and ensuring it aligns to the strategic objective is a good place to start. The following best practices are recommended for a successful result.

  1. Conduct research. Read blogs, books and whitepapers, listen to podcasts, attend events such as webinars and conferences. The Appendix in the full version of the whitepaper lists suggested additional resources.
  2. Treat AI as a product. Put your product manager hat on, ensure all appropriate stakeholders are involved in decision-making. Like any product roll out, be sure to ask: what does success look like?, what measures and metrics should be considered to track performance and ROI? and processes need to be in place to ensure a smooth client experience?
  3. Define use cases. Ensure that the use case is properly defined with the appropriate level of detail. You must focus on one complete use case for one type of user at a time, starting with the highest priorities for the business.
  4. Pilot first. Start small with something like a chatbot, or with a subset of customers.
  5. Keep customers top of mind. Do customers want a bot/AI? Review your Voice of Customer surveys and data. Gathering guidance from customers’ needs, expectations, and preferences will give you the business intelligence on how and where to implement AI. Be sure to ask customers specifically about their needs when considering implementation.
  6. Integrate with other channels in the customer journey. A chatbot is another customer channel in the customer journey. Integration will ensure they have a consistent experience across channels.