The AI Revolution in CX Measurement


Share on LinkedIn

Let’s face it, most CX measurement programs could use a revolution – AI or other. Because CX measurement programs are:

  • Crumbling: CX measurement relies too much on surveys. But transactional surveys capture only between 2% and 7.5% of interactions. And survey response rates are declining (not surprising, because firms flood customers with CX surveys that feel like interrogations and don’t lead to change).
  • Limited: Most programs ignore data that could help track CX quality. Like data customers leave in digital interactions or calls. Or data from new types of interactions (e.g., chats or the 3.5 billion images customers share daily on social media).
  • Superficial: Measurement programs aren’t embedded in organizations, say 90% of CX professionals.*
  • Futile: Metrics don’t help employees to improve CX, say 79% of CX professionals.*
  • Incomplete: Firms aren’t measuring how important customers feel about important experiences, think 65% of CX professionals.*

12 AI use cases have the potential to transform CX measurement programs

We identified 12 promising use cases for how AI can revolutionize CX measurement and VoC programs. My new report, The AI Revolution in CX Measurement (Paywall), describes those use cases and what you need to do now to prepare.

12 AI use cases can improve how firms track CX metrics and drive CX action

If you know everything about AI, skip ahead. For a quick primer on AI read on. AI is the umbrella term for a number of technologies that allow computers to mimic human’s ability to sense, think and act. To look beyond the hype, keep in mind that AI:

  • Is not general purpose. It’s trained to succeed in very specific use cases.
  • Is deep learning and more. Deep learning use cases are really amazing (e.g. identify sarcasm in written text). But lots of AI use cases rely on the more “traditional” machine learning.
  • Needs a lot of training. That requires time and the right training data.
  • Relies on humans. Not computers but humans select the data and algorithms, train the models, and apply judgement to make sense of the results.

Use AI to track CX more effectively and efficiently

The use cases for AI in tracking CX stem from AI’s ability to get insights from more types of data – including lots of unstructured data sources. That means we can rely less surveys to understand what customers say or how they feel. We can instead mine text, calls, images and videos. Here are three cool things that we can do as a result (read about more use cases in the report):

  • More sophisticated emotion analysis. AI can mine more types of emotion signals and put them in context. For example, AI determines that a customer on a video is smiling. But he or she is also trying to input a Captcha code into a computer screen, so he or she isn’t happy but frustrated.
  • Turn surveys from interrogations to conversations. We can allow customers to leave video, image, and audio feedback because we can mine it at scale. In the future, firms can use conversational interface solutions to have a feedback dialog with customers.
  • Use CX data to predict survey scores. Instead of surveying customers, algorithms predict how customers would have answered a survey.

Use AI To Generate Insights That Drive CX Actions

AI reduces time-to-insights and drive more action on the insights. Here a selection of use cases from the report:

  • Identify emerging issues across data sources. CX pros can use AI to detect anomalies and emerging issues they didn’t know to monitor or that they would have seen too late.
  • Add narratives to dashboards and query CX data in natural language. Narratives on dashboards (courtesy of natural language generation) help business users know what to pay attention to. And vendors that offer bots help business users to get in touch with data. In the future, we’ll see more business users query data in natural language in conversational-interface applications.
  • Augment employees in real time to improve CX. Using in-stream analytics, firms can analyze interactions as they occur and provide prompts that help employees improve CX in the moment. AI algorithms also help with micro-learning so employees can practice customer-centric behaviors in near real-time.

The CX professional of the future: A critical thinker and master collaborator

Only with critical thinking and collaboration skills will you be able to:

  • Make AI results useful: You’ll need to work with data scientists, business stakeholders, privacy experts, outside vendors and many more. Why? They help you figure out the right AI use cases (given the available data, your firm’s analytics maturity, and budget). They also help you avoid data quality issues and biases.
  • Keep calm and improve CX: AI will put pressure on the beginning of the “funnel” in a CX measurement program: Many more insights will be available. That puts an even higher burden on your ability to sift through them and prioritize them.

For more use cases, details and examples, check out my new report The AI Revolution in CX Measurement.

* Source: 2017 Q3 State Of Customer Experience Maturity Online Survey. N=348 CX professionals, shown is % strongly agree.

Maxie Schmidt, Ph.D.
Maxie Schmidt is a Principal Analyst at Forrester, serving Customer Experience Professionals. She leads Forrester’s research on CX measurement. Her work focuses on questions like how to measure CX, how to tie CX quality to financial outcomes (AKA how to make the case for CX), andinnovation in CX measurement and VoC programs. Follow her on Twitter @maxieschmidt.


Please use comments to add value to the discussion. Maximum one link to an educational blog post or article. We will NOT PUBLISH brief comments like "good post," comments that mainly promote links, or comments with links to companies, products, or services.

Please enter your comment!
Please enter your name here