This article was written by Troy Gedlinske (CDI’s Expert in Residence) along with US based colleagues at CDI Services, a global leader in Conversational AI training, auditing, and consulting services.
This is the second article in a program designed to help customer service leaders develop a realistic, effective AI strategy.
Each article is a summary of a more detailed, hands-on workshop designed exclusively for CS leaders at the VP level and above. To learn more about CDI’s suite of workshop offerings, consulting and auditing services you can contact CDI’s Global Head of Sales in Ben Taylor on [email protected]
Executive Summary
-
Over the last fifty years, Customer Service has been altered beyond recognition by new technology. Although there have been hiccups, customers have responded positively overall, especially to self-service options offering them convenience and control.The success of self-service CS changes how customer-obsessed companies think about CS Quality. Fair or not, many customers view even the most successful live CS interaction as a process and/or product defect. That hurts the customer perspective of Quality—no matter what your internal KPIs say.
-
When properly conceived and implemented, an AI-powered assistant is an efficient, effective self-service option—no queueing or phone calls required. This boosts the customer perception of CS Quality as measured by NPS, HMD, DSI, and other external Quality KPIs.
-
AI-powered agents can also boost your SLAs and other internal Quality KPIs, e.g., by allowing you to expand your Hours of Operation (HOOP) without the excessive cost and complexity. This can have a profound effect on CS Quality.
Self service and the customer perception of Quality
Quality is one of the five pillars of CS. As such, it should be a big part of your AI strategy.
Although there are many different ways to define Quality within your company, customer-obsessed companies recognize a simple truth:
💡 When it comes to Quality, customer perceptions are the only thing that matters.
This has always been true to some extent—even when the only CS options were store visits and snail mail. But over the last fifty years, technology has altered CS beyond recognition: toll-free numbers in the mid-1970s; fax in the 1980s; email in the 1990s. Customer expectations naturally changed with each new technology.
Then, with the Web 2.0 internet in the 2000s, a new class of self-service solutions emerged. From FAQs to customer community forums to YouTube videos, customers now have multiple options to solve their own issues without a live CS interaction. This has been a huge boon for customers and companies, and there’s no going back.
But from the customer perspective, the benefits of self-service solutions come with a catch:
💡 Today, CS solutions that require a live interaction can hurt customer perceptions of Quality.
Fair or not, even a successful CS interaction may be interpreted by the customer as a substantive defect in process and/or product. Even worse, it can signal a misalignment of interests between you and your customers (e.g., when companies require customers to call an agent to cancel a service that they signed up for with just a few taps on their phone).
External vs. internal Quality KPIs
Given the importance of the customer perception of Quality, it helps to organize related KPIs into external and internal perspectives.
As you can see, the external perspective includes two of the most common Quality KPIs: Net Promoter Score (NPS) and How’s My Driving (HMD). One of the advantages of these KPIs is that they are measured via surveys triggered at the end of a customer interaction. This ensures that you’re sourcing your data directly from your customers.
A third KPI is Downstream Impact (DSI), which attempts to quantify how CS Quality changes the behavior of customers in a way that benefits your company. DSI is finicky and often derivative of NPS and HMD. But if you focus on specific drivers like Churn Risk and Churn Rate, DSI offers a neat mechanism for quantifying improvements to your programs and processes.
Compared to the external perspective of Quality, the internal perspective is less important. But less important does not mean unimportant.
For example, one of the most common and useful internal Quality KPIs are Service Level Agreements (SLAs). Although SLAs are (and should be) aligned with your customer and company needs, most companies that accept real-time inbound contacts with human agents typically accept an SLA of 80% to 90% of contacts answered in 60 seconds.
Within the context of Quality, SLAs are particularly interesting because of their relationship to Hours of Operation (HOOP). Like SLAs, HOOP varies depending on the nature and scale of your business. However, there is a conventional wisdom about which companies should offer 24/7 HOOP (e.g., global financial services firms, airlines) and which are fine sticking to normal business hours. This conventional wisdom is relevant to our discussion here because it is being disrupted by AI, a point we return to below.
How AI can drive your external Quality KPIs
When properly conceived and implemented, a customer-facing AI-powered agent provides what most (if not all) your customers want: a direct, efficient way to solve a problem when they want to solve it. No getting on the phone, no waiting for a reply, no time or reason to vent on social media.
For you, the benefit is a direct boost in customer perceptions of Quality: of your company, your products and services, and the CS processes you have in place. This should drive improvements in NPS and HMD. These improvements should also show up in DSI (e.g., by lowering churn).
An AI-powered agent can also boost CS Quality indirectly:
- An AI-powered agent frees you to invest more in your human agents, e.g., to give them the time, training, and tools they need to focus on a narrower set of more complex problems. There is nuance here: by definition, complex problems are harder to solve and therefore more stressful for both customers and agents. But the existence of an effective self-service option can provide a “rising tide” to make solving any kind of problem easier.
- An AI-powered agent can be designed to facilitate a thoughtful, informative warm handoff to its human counterpart when the problem is too complex for self service. This makes the subsequent interaction between the customer and agent more productive.
- Observing how customers interact with AI-powered agents can address common blind spots. For example, some CS Ops teams assume that a self-service interaction is resolved if it doesn’t generate a CS interaction. In fact, customers may abandon a self-service attempt out of frustration, turning instead to Google or Reddit (either to solve their problem or to complain). Unlike interactions with other self-service options, an interaction with an AI-powered agent generates data that can be analyzed.
This last example is fascinating because it hints at the kind of insights that are available once you set up feedback loops between AI-powered agents, human agents, and survey results.
For example, after a customer and human agent solve an issue, you can use feedback from both to develop and test optimized solutions via an AI-powered agent. You can then use these results to improve both AI and human agents. This is much harder when only human agents are available for testing because even high-performing human agents may be inconsistent from interaction to interaction.
Indeed, just improving the consistency of customer experience can boost customer perceptions of Quality. This is where AI-powered agents thrive, because they’re always the same—regardless of geographic location, day of week, or time of day. This eliminates one half of the variability of customer-to-human agent interactions.
How AI can drive your external Quality KPIs
Any success you have using AI-powered agents to improve the customer perception of Quality should improve your internal Quality KPIs as well. This is because most internal KPIs are derivative of external ones.
That said, we’d like to call out one very powerful opportunity for using an AI-powered agent to drive internal Quality KPIs directly. To that end, let’s return to SLAs and HOOP.
In general, only the largest companies adopt an all-hours, all-channel HOOP, as such an operation requires thousands of human agent hours daily. This generates significant costs and management complexity.
But for companies large and small, AI-powered agents allow for the expansion of HOOP (and/or more aggressive SLAs) at a fraction of the cost. This is because AI-powered agents don’t require sleep or time off or time zone management. They do require human management, of course, but that management can occur comfortably within regular company hours.
So if you think even a small percentage of your customers would enjoy 24/7 support, an AI-powered agent is worth it—even before considering the benefits to your external KPIs.
*Note a slightly different version of this piece was published earlier in 2024 via Troy’s LinkedIn. The latest work developed by Troy alongside the US CDI team will be released first via CDI’s blog.
The third installment of the series titled ‘Improving CS Finance in the new era of AI’ will be released on Monday December 2nd.
If you would like to be notified of additional releases in the series along with white paper content being developed from the series in 2025, you can sign up for direct notifications here.
