AI customer service statistics for 2026: adoption, costs, and what customers actually think

If you want the state of AI customer service statistics in three numbers, here they are. Gartner predicts that agentic AI will autonomously resolve 80% of common customer service issues by 2029, cutting operational costs by 30%. Intercom's 2026 Customer Service Transformation Report found that 82% of senior support leaders invested in AI over the last 12 months — yet only 10% say they've reached mature deployment. And in a Gartner survey of 5,728 customers, 64% said they would prefer companies didn't use AI for customer service at all.

That tension — near-universal adoption on one side, real customer skepticism on the other — is the story of AI support right now. This page collects the most credible numbers we could verify, with a link to the original source for every single statistic. No vendor-blog folklore, no unsourced "80% of customers say" filler.

If you're building the case for (or against) automation at your company, pair this with our guides to customer service automation and the metrics that actually matter.

Adoption: how many companies use AI in support

Adoption is broad, but shallow. Most teams have bought something; few have rebuilt their support operation around it.

• 88% of organizations use AI in at least one business function, according to McKinsey's State of AI survey of nearly 2,000 organizations across 105 countries.
• 62% of organizations are working with AI agents — 39% experimenting and 23% already scaling agent systems, per the same McKinsey research.
• 82% of senior support leaders say their teams invested in AI for customer service in the last 12 months, and 87% plan to invest in 2026, per Intercom's survey of 2,400+ support professionals.
• Only 10% of support teams have reached mature AI deployment — AI fully integrated and working at scale — in that same Intercom report.
• 76% of support teams invested in AI in 2024, up from the 54% who said they planned to a year earlier, per Intercom data compiled by HubSpot.
• 81% of CRM leaders believed most customer service professionals would be using AI in their daily work by 2025, per HubSpot's State of Customer Service research.
• Salesforce's seventh-edition State of Service report projects that AI will handle 50% of service cases by 2027, up from roughly 30% in 2025, per Salesforce.

The takeaway: "do you use AI?" is no longer an interesting question. The gap that matters is between the 82% who bought AI and the 10% who actually run their support operation on it.

What customers actually think

This is the section most vendor roundups skip, so let's not. The honest picture: customers like fast answers, and a meaningful share of them still distrust the thing producing them.

• 64% of customers would prefer that companies didn't use AI for customer service, and 53% would consider switching to a competitor if they found out a company was going to use AI for support, per Gartner's December 2023 survey of 5,728 customers. Their top fear: it will get harder to reach a person.
• 75% of consumers prefer talking to a real human in person or over the phone for customer support, per a Five9 study of 4,000 US and UK consumers.
• 56% of consumers are often frustrated by AI customer service chatbots, and 48% don't trust the information AI-powered bots provide, in the same Five9 research.
• The preference is generational: 66% of Gen Z prefer humans versus 86% of Baby Boomers, and 63% of Gen Z say they do trust AI bot answers, per Five9.
• It's not all skepticism: 51% of consumers prefer interacting with bots over humans when they want immediate service, per Zendesk CX Trends data in Zendesk's AI statistics roundup.
• But the bar is high — 68% of consumers believe chatbots should match the expertise of highly skilled human agents, per the same Zendesk data.
• 95% of consumers expect clear explanations for decisions made by AI, and 74% are frustrated when they have to repeat information, per Zendesk's CX Trends 2026 research across 11,000+ respondents in 22 countries.
• 86% of consumers say responsiveness and accuracy influence their purchasing decisions, per the same Zendesk CX Trends 2026 report.

Read together, the data says customers don't hate AI — they hate bad AI they can't escape from. Speed wins them over; wrong answers and dead-end bots drive them to competitors. That's why escalation design matters as much as the model.

Response time and resolution benchmarks

Speed is where AI's case is strongest, and where customer expectations are least forgiving.

• Almost two-thirds of buyers expect a response within 10 minutes of any sales or customer service inquiry, per HubSpot Research.
• 67% of consumers expect their support ticket to be resolved within 3 hours, per HubSpot's State of Customer Service data.
• 75% of customer service reps said 2024 brought the highest ticket volume they had ever seen, per HubSpot — expectations are rising while volume rises too.
• Klarna's AI assistant cut average resolution time from 11 minutes to under 2 minutes in its first month, while matching human agents on customer satisfaction, per Klarna's own press release.
• The same assistant produced a 25% drop in repeat inquiries, which Klarna attributes to more accurate first answers (Klarna).
• AI typically resolves 21–40% of customer requests, with some teams reaching up to 80% of conversations, per Intercom data cited by HubSpot.
• 53% of support teams name faster response and resolution times as a top benefit of AI, per Intercom's 2026 report.

A useful sanity check when you see resolution claims: "deflected" and "resolved" are not the same thing. A bot that makes a customer give up counts as deflection. Which brings us to the most under-quoted statistic in the industry:

• Only 9% of customers report fully resolving their issues through self-service, per Gartner's 2019 customer service research — the pre-LLM baseline modern AI agents are trying to beat.

Cost and ROI figures

The economics are the reason this market exists. The spread between a human-handled contact and an automated one is enormous — when the automation actually resolves the issue.

• Live channels (phone, live chat, email) cost an average of $8.01 per contact, versus roughly $0.10 for self-service, per Gartner's Customer Service and Support leader poll — an 80x difference.
• Gartner's 2029 agentic-AI prediction comes with a cost figure attached: a 30% reduction in operational costs for service organizations that get there (Gartner).
• Klarna estimated its AI assistant would drive $40 million in profit improvement in 2024, doing the equivalent work of 700 full-time agents across 2.3 million conversations in its first month (Klarna).
• The honest coda: by May 2025, Klarna was hiring human agents again, with its CEO acknowledging the company had over-rotated on AI at the expense of quality, per CX Dive's reporting.
• 87% of teams with mature AI deployment report improved metrics since implementing AI, versus 62% of teams overall, per Intercom — ROI concentrates among teams that actually operationalize the tools.

Two lessons hide in these numbers. First, the ROI is real but back-loaded: it shows up after you've done the unglamorous work of knowledge base cleanup and escalation design. Second, pricing models matter — per-resolution fees claw back a chunk of the savings that make automation attractive in the first place. We've run the full cost math on AI support separately (disclosure: we build Asks, a flat-priced AI support agent, so we think about this a lot).

Channel statistics: chat, email, and messaging

• Almost two-thirds of buyers expect a reply within 10 minutes, which effectively rules out email as a first-line channel for urgent questions (HubSpot Research).
• The average live chat wait time is around 46 seconds, per Comm100 benchmark data covering tens of millions of chats, compiled by Help Scout.
• 51% of consumers prefer bots to humans specifically for immediate service (Zendesk) — the channel preference flips when speed is the priority.
• 76% of consumers say they would choose companies that offer text, voice, and visuals in a single conversation, per Zendesk CX Trends 2026 — the multimodal, multi-channel expectation is now mainstream.
• Klarna's AI assistant operates in 23 markets and more than 35 languages, 24/7 (Klarna) — coverage no human team of comparable cost can match.

What's coming: the agentic AI forecasts

Analysts are bullish on the destination and bearish on most teams' route to it. Both halves are worth quoting.

• By 2029, agentic AI will autonomously resolve 80% of common customer service issues without human intervention, per Gartner's March 2025 prediction.
• Over 40% of agentic AI projects will be canceled by the end of 2027 due to escalating costs, unclear business value, or inadequate risk controls, per Gartner's June 2025 prediction. Gartner also estimates only about 130 of the thousands of self-described "agentic AI" vendors are the real thing — the rest are "agent washing."
• Half of the companies that cut customer service staff because of AI will be rehiring by 2027, per Gartner's February 2026 prediction — the Klarna pattern, generalized.
• 33% of enterprise software applications will include agentic AI by 2028, up from less than 1% in 2024, per Gartner.
• The conversational AI market is projected to grow from $11.58 billion in 2024 to $41.39 billion by 2030 — a 23.7% CAGR, per Grand View Research.

The 10 numbers that matter most

Methodology and sources

Every statistic on this page was compiled from publicly available studies, press releases, and vendor research reports, with an inline link to the original source next to each number. Primary sources include Gartner press releases, Zendesk CX Trends research, Intercom's Customer Service Transformation Report, HubSpot's State of Customer Service research, Five9's consumer survey, McKinsey's State of AI survey, Salesforce's State of Service report, Klarna's public case study, and Grand View Research market data. Survey years vary (2019–2026) and are noted or linked in context; predictions are attributed to the analyst firm that made them. None of these figures are Asks data. Links last verified July 2026. If you cite this page, please also credit the original source of the specific statistic.

Frequently asked questions

What percentage of companies use AI for customer service?

Adoption is high but maturity is low. Intercom's 2026 report found 82% of senior support leaders invested in AI for customer service in the last 12 months, and McKinsey puts overall organizational AI use at 88%. But only 10% of support teams describe their AI deployment as mature and working at scale — most are still somewhere between pilot and partial rollout.

Do customers actually like AI customer service?

It's genuinely mixed. Gartner found 64% of customers would prefer companies didn't use AI for service, and Five9 found 75% prefer talking to a human. But 51% of consumers prefer bots when they want immediate service, per Zendesk. Customers reward speed and accuracy and punish dead-end bots — the implementation matters more than the technology.

How much does AI reduce customer service costs?

Gartner benchmarks a live-channel contact at $8.01 versus about $0.10 for self-service, and predicts agentic AI will cut service operational costs 30% by 2029. Klarna publicly estimated $40 million in annual profit improvement from its AI assistant. Real savings depend on how much of your volume AI actually resolves — typically 21–40% today.

Will AI replace customer service agents?

The data points to reshaping, not wholesale replacement. Gartner predicts AI will autonomously resolve 80% of common issues by 2029 — but also that half of companies that cut service staff due to AI will be rehiring by 2027. Klarna, the most cited AI-replacement story, resumed hiring human agents in 2025. The emerging pattern: AI handles repetitive volume, humans handle complex and emotionally loaded cases — which is why handoff design is where good implementations differentiate.