Customers are your greatest acquisition—and your greatest asset. This is especially true when it comes to understanding how your business is really doing. Customers will always tell businesses about their performance—good, bad, and even the ugliest encounters.
The customer experience (CX) realm encompasses customer service, support, and generalized experience. CX is also the primary org of most businesses adopting generative artificial intelligence (AI) in an effort to streamline and optimize customer interactions. If you’re new to AI in CX, you may wonder if an in-house AI build is best, or if a third-party AI vendor is a better option to integrate into your toolkit, or what to expect when moving customers toward AI and how their issues will be resolved, and how fast.
The best way to understand any of that is to look at the numbers. Metrics are one of the most valuable ways to understand how the customer experience is improved by generative AI, and how it’s trending over time. To get a better understanding of how CX is boosted by AI strategies and tools, one must take a look at some historical data, even year-over-year trends.
Ahead, we’ll go through three key metrics to measure and where they have trended from 2023, which you can find in Forethought’s 2024 AI in CX Benchmark Report.
Why are CX metrics important to measure?
In business, numbers and measurement play an essential role—no matter the org or team or role. Everyone measures the efficacy of something to ensure they’re on the right track or in need of a pivot. This is especially true of CX.
Measuring CX, perhaps seemingly nebulous, can provide concrete evidence of how well customers are experiencing parts of your company, and how to improve or iterate on that. They can provide insights into satisfaction, advocacy (like how willing they are to suggest your brand to someone else), and overall loyalty.
3 top CX metrics and trends since 2023
There are a number of CX metrics businesses can and should measure, but three of the most important, and especially so for those adopting and integrating generative AI into their customer support tech stack, are: cost per resolution, customer satisfaction, and deflection rate.
These metrics from Forethought’s Benchmark CX Report tell you about time and money saved or improved and how they’ve trended since 2023.
Cost per resolution
Here, this is the average cost of a business’s support team successfully resolving a customer support ticket.
AI solution type
The report found that since 2023 AI that’s trained on company data proved to positively impact cost per resolution over any other form of AI model training.
- Companies who trained AI on their own data are nearly 3.5x more likely to lower cost per resolution.
- Only 5% of companies trained on their own historic data reported a flat cost per resolution, and none reported an increase in costs.
The type of AI training data used showed a steady correlation with a lowered average cost per ticket resolved.
- Companies using their own historic data for training saw an average cost per resolution of $9.
- Companies not using this type of data for training reported an average cost per resolution of $10, the same as the cost for companies not using AI at all.
Help desk
In 2024, the report found that product or service sold, complexity of issues, and processes in place affect costs.
- In-house platforms led to the highest average cost of $16 per resolution.
- When other types of training data were used, in-house built platforms drove costs from $15 to $20.
Revenue
Companies at different levels of revenue established a wide variation in their cost for each customer ticket resolved, with an even larger range of results when AI is used, according to the report.
In 2024, the report notes on cost per resolution according to revenue:
- The lowest overall cost per resolution reported by companies in the $0M – $25M and $2B – $5B ranges was $9 per ticket.
- The highest overall cost per resolution at $12 came in the $201M – $500M band.
- For companies using AI trained on their own historic data, the greatest difference was seen among companies in the $500M – $1B range, who had a cost of $6 per resolution compared with $12 for their competitors and peers who weren’t using AI.
Monthly ticket volume
The cost of resolving a ticket increases with its complexity. However, with numbers from 2024, the use of AI changes the numbers.
- The lowest overall average ticket resolution cost is $8, as seen by companies with simple issues to resolve.
- Companies with complex issues spend an average of $12 per ticket.
Customer satisfaction (CSAT)
How satisfied are customers with a product or service or experience with a business? The customer satisfaction (CSAT) score is a great indicator.
AI solution type
Year-over-year, according to the report, companies using AI for CX showed an improvement in CSAT regardless of the solution type used, though different types did have different levels of impact.
- Companies using a dedicated point solution saw a positive trend with 50% saying that CSAT was up and only 8% seeing a decline.
- Companies using a help desk add-on reported softening CSAT with 14% seeing a negative trend.
Help desk
Most help desk platforms found a flat or improved CSAT year-over-year, according to the report.
- One business found users were most likely to report rising CSAT with 57% seeing a positive trend, including 14% who said it was way up.
- Another service saw 15% of its users report lower CSAT
Revenue
The report found a year-to-year trend in CSAT reported by companies at different levels of revenue.
- There’s a positive trend in companies with $25M – $200M in annual revenue as 43% of these organizations said that deflections were up, including 5% who said that they were way up.
- The largest decline in CSAT came for companies with $5B+ in annual revenue, with 28% seeing a negative trend.
Monthly ticket volume
The report shows that company CSAT scores have changed over the past year, with respondents showing widely varying trends across ticket volume levels.
- Organizations handling 1,001 – 5,000 monthly tickets most likely saw a positive trend, with 53% saying that CSAT was up or way up.
- 14% of companies with a monthly ticket volume of 10,001 – 25,000 saw CSAT trending downward, including 2% who said it was way down.
Deflection rate
In many cases, AI acts as a self-service pathway for many customers looking to solve less complex issues. Deflection rate can indicate users electing to find information through chatbots, knowledge bases, or even communities, rather than speaking to a live agent.
AI solution type
In the report, a year-to-year trend in deflection rate indicated differing, yet still positive, results from companies using different types of AI solutions.
- A dedicated AI platform provided the greatest improvement year-over-year with 59% of companies reporting a higher deflection rate
- Companies using an add-on from a help desk vendor saw a negative trend with 33% reporting deflections were down, including 7% who said they were way down.
Help desk
Year-to-year trends in the report showed deflection rates vary widely as well.
- One business’s users saw a positive trend with 47% percent reporting rising deflections.
- In-house built solution users were moving in the opposite direction with 50% seeing a decline.
Revenue
Companies at all revenue levels tended to report relatively unchanged deflection rates year-over-year.
- Companies with revenue between $25M – $200M were most likely to have seen improving deflection rates with 45% reporting a positive trend.
- Companies over $5B in revenue saw declining deflections with 26% saying that the rate was down and another 5% describing the trend as way down.
Monthly ticket volume
The report indicates that the deflection rate from year to year was fairly flat across most levels of monthly ticket volume.
- The highest-volume organizations report a rising deflection rate with 51% of those handling 25,000+ monthly tickets citing a positive trend.
- On the other hand, 41% of the lowest-volume organizations with 0 – 500 tickets per month saw deflections fall over the last year.
Conclusion
Understanding CX through these metrics is crucial in both overall business decision-making and understanding where and how generative AI is improving a customer’s experience. Customer experience is the single biggest application for AI in business, and one of the most common ways customers are going to interact with the technology. Because it’s happening so fast, and the technology is evolving constantly, there is no room for error. Always look to the numbers to understand the health of your CX.
To understand the full scope of AI and CX, read the Forethought 2024 AI in CX Benchmark Report.
