AI is becoming a bigger part of customer support—but what does that actually mean for teams on the ground?
To find out, we surveyed more than 600 CX professionals across industries, company sizes, and support models. We asked how they’re using AI, what results they’re seeing, and how their approaches compare to others. This year’s responses offer a clearer picture than ever of how AI is being used in practice—and what’s helping teams succeed.
The 2025 AI in CX Benchmark Report brings together all of this data, with breakdowns by industry, team size, ticket volume, helpdesk platform, issue complexity, and more. Whether you’re already using AI or just exploring your options, this report gives you real data to help understand how other teams approach support—and what’s working.
What We Learned:
✅ Agentic AI leads the way.
Companies using agentic AI—AI that doesn’t just respond, but fully resolves customer issues—are seeing better outcomes across the board:
- 64% said their CSAT was improving (compared to 49% for companies not using AI)
- 54% reported better customer retention
- 63% said their resolution costs were trending down
✅ Training AI on your own data makes a difference.
Teams that trained their AI using their own historic ticket and CRM data reported:
- A higher CSAT
- Lower average resolution costs
- Better trends in retention and deflection over time
✅ Support is still a multi-channel effort.
Nearly all respondents offer email support, and most offer voice and chat. However, while voice was named the top channel for fast resolutions in our consumer survey, many teams admitted their voice support still falls short due to long wait times and disjointed systems. This is one of the biggest areas teams want to improve with AI.
✅ Benchmarks show real gaps.
When we compared companies by industry, team size, and even ticket volume, the differences were striking. For example:
- B2C companies using agentic AI saw deflection rates 55% higher than those using non-agentic AI
- Public companies using agentic AI reported an 86% higher deflection rate than their peers using non-agentic systems
Why This Report Matters
Every support team is unique—but that doesn’t mean you have to guess what’s working.
We created this report to help CX leaders see where they land compared to similar companies. Whether you’re in healthcare or commerce, running a team of 5 or 500, or handling 100 tickets a day or 10,000, this report provides a way to benchmark your performance against others like you.
It also gives you a sense of what’s possible. If your team is still using basic automation or hasn’t started with AI at all, you’ll see the kinds of gains other teams are making when they move toward agentic systems that can handle issues from start to finish.
A Note From Lindsay
“We built this report to offer more than just numbers—we want it to be something CX teams can actually use. Whether you’re looking to validate your strategy, make a case for change, or just understand how others are approaching support, our hope is that you’ll find real value here. The best insights come from shared experience—and this report is full of it.”
— Lindsay Fifield, Head of Customer Success, Forethought
Hashtags blocks for sticky navbar (visible only for admin)
{{resource-cta}}
{{resource-cta-horizontal}}
{{authors-one-in-row}}
{{authors-two-in-row}}
{{download-the-report}}
{{cs-card}}

