Most call center statistics articles are a chain of citations with nothing at the end of it. A number appears on a vendor blog, a second blog cites the first, a third cites the second, and four years later the figure is still circulating with no study, no sample size and no date attached to it.
So this roundup of call center statistics has one rule: every statistic below names the organisation that ran the research, the year it was run, and links to the source. Where we could find the sample size, it is stated. Where a widely quoted number turned out to be untraceable, it is not in the list. There is a section near the end naming those instead, because knowing which call center stats to stop quoting is worth as much as knowing which to use.
The Short Version
AI is everywhere and resolving less than the headlines suggest. 98% of contact centres use AI in some form (Calabrio, 2025), but only 20% of service leaders have actually reduced agent headcount because of it (Gartner, 2026).
Customers still want a person. 84.9% of consumers prefer a human agent, and 80.1% still prefer one even when told AI would resolve the issue equally well (Metrigy, 2025).
Patience is short and getting shorter. 86% of consumers expect to reach a live agent within 10 minutes, and over 60% report waiting 15 to 60+ minutes in the past year (Genesys, 2025).
The cost of getting it wrong is measurable. Bad customer experiences put an estimated $3.7 trillion of global revenue at risk annually (Qualtrics XM Institute, 2024).
How We Sourced These Contact Center Statistics
- Every figure traces back to a named research organisation: an analyst firm, a vendor's own primary research, a government statistics body or a peer-reviewed paper.
- The year given is the year the research was conducted or published, not the year a blog repeated it.
- Sample sizes are included wherever the original source disclosed them.
- Analyst predictions are labelled as predictions. They are modelled forecasts, not things respondents said.
- Statistics we could not trace to a primary source were excluded and are listed separately at the end.
- Last verified: September 2026.
AI and Automation in the Contact Center
Of all the contact center statistics published in the last two years, these are the ones moving fastest, and the ones where the gap between adoption and results is widest.
How Widely AI Is Actually Deployed
Almost every contact centre now uses AI in some form, but far fewer have moved past chatbots and pilots into deeper, organisation-wide deployment.
| Statistic | Sample | Source |
|---|---|---|
| 98% of contact centres are already using AI in some form | 400+ global contact centre leaders | Calabrio, State of the Contact Center 2025 |
| 92% of organisations have implemented or piloted AI in customer service | 600 business decision-makers across the US, UK and Germany | Five9 |
| 98% of organisations have deployed AI somewhere in the customer journey, but only 15% combine agentic AI with cross-departmental orchestration | 252 director-level and above CX decision-makers globally | Talkdesk, State of Agentic Automation in CX 2026 |
| 48% of organisations with mature service capabilities already use agentic AI, against 24% of low-maturity peers | Deloitte Digital's global service leader survey | Deloitte |
| 51% of small businesses have already integrated AI into customer service | 400 US small business owners | Talkdesk |
| 32% of AI-adopting small businesses use voice AI or IVR, against 74% using chatbots | 400 US small business owners | Talkdesk |
| 64% of CX executives are increasing chatbot and AI investment | 1,300+ CX executives | Zendesk CX Trends 2026 |
| 43% of contact centre leaders plan to build AI in-house rather than buy third-party tools | 400+ global contact centre leaders | Calabrio |
More than 90% of organisations already use AI in customer service, but few go deep. Only 15% combine agentic AI with cross-departmental orchestration, and just 32% of AI-adopting small businesses use voice AI, compared with 74% using chatbots.
What AI Is Actually Delivering
Adoption is nearly universal, but results vary widely: some teams report lower costs and faster agents, while most AI pilots never move the bottom line.
| Statistic | Sample | Source |
|---|---|---|
| 39% of service leaders report a lower cost per contact after deploying AI, and 64% report higher agent productivity | Deloitte Digital's global contact centre survey | Deloitte |
| 43% of organisations expect AI to cut contact centre costs by 30% or more within three years | Deloitte Digital's global service leader survey | Deloitte |
| AI assistance raised issues resolved per hour by 14% on average, and by 34% for novice agents | Controlled study of 5,179 customer support agents at a Fortune 500 software firm | Brynjolfsson, Li and Raymond, NBER |
| An AI copilot cut average handle time by 65% for agents searching internal knowledge | Single client case at a European media and telecom company | McKinsey & Company |
| 95% of generative AI pilots deliver no measurable P&L impact | Enterprise-wide study across all business functions | MIT NANDA, The GenAI Divide |
AI can lower costs and make agents faster, with a 14% average productivity gain that rises to 34% for novice agents. The returns still aren't guaranteed, though: 95% of generative AI pilots show no measurable P&L impact, so results depend on deploying AI against a specific use case you can measure.
What Analysts Predict Happens Next
These are forecasts, not survey findings. They are included because they drive budgets, and because two of the biggest ones contradict each other.
- Agentic AI will autonomously resolve 80% of common customer service issues by 2029, cutting operating costs by 30%. (Prediction) Gartner analyst forecast. (Gartner, 2025)
- Over 40% of agentic AI projects will be cancelled by the end of 2027. (Prediction) Gartner cites escalating costs, unclear business value and inadequate risk controls. A January 2025 Gartner poll of 3,412 attendees found only 19% had made significant agentic AI investment. (Gartner, 2025)
- 50% of organisations that planned to cut customer service headcount will abandon those plans by 2027. (Prediction) Based on 163 customer service and support leaders. 95% said they intend to keep human agents in a strategic role. (Gartner, 2025)
- Half of current customer service jobs will be lost to AI by 2030, concentrated in high-volume, low-complexity work. (Prediction) Forrester report AI Agents Reshape the Customer Service Workforce. (Forrester via Customer Experience Dive, 2026)
- AI is projected to handle 50% of customer service cases by 2027, up from 30% today. Based on 6,500 service professionals across 39 countries. (Salesforce, 7th State of Service Report, 2025)
Skip the reading. See CloudTalk live.
What Customers Actually Expect
Human or AI
- 84.9% of consumers prefer interacting with a human agent over an AI agent, and 80.1% still prefer a human even when assured AI would resolve the issue equally well. Based on 503 consumers in the US and Canada. (Metrigy, 2025)
- 66% of consumers still prefer speaking with a human, even though 80% are willing to use AI-powered customer service. Based on 3,000 consumers across the US, UK and Germany. (Five9, 2026)
- 83% of consumers report having to repeat themselves after being transferred from AI to a human agent. Same 3,000-consumer Five9 sample. This is the single most fixable number in this article. (Five9, 2026)
- 72% of consumers are open to AI-powered interactions, provided they can escalate to a human when needed. Based on 1,006 consumers across the US, UK and Canada. (Five9 Customer Experience Report, 2025)
- 75% of consumers prefer talking to a real human over AI, ranging from 66% of Gen Z to 86% of Baby Boomers. Based on 4,000 consumers in the US and UK, census-balanced by age and gender. (Five9, 2024)
- 54% of consumers said human agents delivered better service than AI over the 2025 holiday season, and 55% had to escalate an AI-handled issue to a human. Based on 1,000 US adults who had completed most or all of their holiday shopping. (Liveops, 2025)
Channel Preference and Self-Service
- 74% of consumers still prefer the phone for complex or urgent matters, and 56% prefer it for general issues. Based on 1,006 consumers across the US, UK and Canada. (Five9 Customer Experience Report, 2025)
- 86% of consumers will try self-service before contacting support. Same Five9 sample. (Five9, 2025)
- 59% of consumers would rather use an instant AI chatbot than wait for a live agent. Same Five9 sample. Read alongside the human-preference figures above: people want a person, but not enough to queue for one. (Five9, 2025)
- 74% of consumers now expect 24/7 customer service. Based on 6,182 consumers and 5,115 business respondents across 22 countries. (Zendesk CX Trends 2026, 2026)
- 75% of CX leaders expect 80% of customer interactions to be resolved without human intervention within a few years. Based on roughly 5,100 consumers and 5,400 CX leaders and agents across 22 countries. (Zendesk CX Trends 2025, 2025)
The Cost of Getting It Wrong
Poor service costs more than one call. These stats show how quickly one bad experience turns into lost customers and lost revenue.
| Statistic | Sample | Source |
|---|---|---|
| Bad customer experiences put an estimated $3.7 trillion of global revenue at risk each year, up roughly $600 billion on the prior estimate | 28,400 consumers across 20 industries and 26 countries | Qualtrics XM Institute |
| 63% of consumers say they would switch to a competitor after one bad experience, up 9 points year on year | Roughly 5,100 consumers across 22 countries | Zendesk CX Trends 2025 |
| 85% of CX leaders say customers will drop a brand that cannot resolve their issue on first contact | 6,182 consumers and 5,115 business respondents | Zendesk CX Trends |
| 86% of consumers say responsiveness and accurate resolution strongly influence their purchase decisions | 6,182 consumers and 5,115 business respondents | Zendesk CX Trends |
| Nearly 74% of consumers will switch providers after a poor contact centre experience | 2,000 US adults | CallMiner |
Customers give you less room for error every year. Nearly two thirds would switch after one bad experience, and $3.7 trillion in revenue is at risk globally, so every unresolved call is a churn risk.
Wait Times, Hold and Abandonment
If you only act on one section, make it this one. These call center stats describe the gap between what callers will tolerate and what most contact centres deliver, and it is the gap a 24/7 AI voice agent closes fastest.
| Statistic | Sample | Source |
|---|---|---|
| 86% of consumers expect to reach a live agent within 1 to 10 minutes, yet over 60% report actually waiting 15 to 60+ minutes in the past year | 5,157 consumers and 1,181 CX decision-makers across 16+ countries | Genesys, State of Customer Experience |
| 75% of callers hang up after 8 minutes on hold, 54% leave within 8 minutes, and only 6% will hold for as little as 2 minutes | 400 US respondents in a customer patience benchmark survey | Nextiva |
| 60% of consumers say short wait times are critical to a good service experience | 1,006 consumers across the US, UK and Canada | Five9 |
| 62% of consumers say their issue was resolved on first contact | 23,730 consumers across 23 countries | Qualtrics XM Institute, Global Contact Center Trends 2025 |
| 64% of consumers say it was somewhat or very easy to connect with a human agent | 23,730 consumers across 23 countries | Qualtrics XM Institute |
| Only 37% of customers offered a callback accept it; around 90% answer when the callback comes | Academic analysis of a US bank's call centre data, published in Management Science | Emadi, Deshpande and Hathaway, UNC Kenan-Flagler |
| Nearly 80% of callers are transferred at least once, and only around 51% are satisfied with the call experience | 500 US consumers who had made high-stakes purchases | Invoca, Buyer Experience Benchmark Report |
For the operational targets behind these numbers, such as what a healthy service level, abandonment rate or average handle time looks like, see our guide to call center metrics for successful benchmarking. This page counts what is happening across the industry; that one tells you what to aim for.
Try CeTe on your own number
Agents: Attrition, Pay and Workload
- Contact centre agent attrition averaged 76% annually. Based on 600 contact centre leaders across the US, Australia, Canada, Japan and the UK. (Deloitte Digital, Global Contact Center Survey, 2024)
- Only 54% of agents remain in role past the two-year mark. ICMI's annual State of the Contact Center research. (ICMI, 2024)
- 63% of service and support leaders are reducing frontline headcount through attrition rather than layoffs. Based on 321 customer service and support leaders worldwide. (Gartner, 2026)
- Only 20% of service leaders say they have actually reduced agent staffing because of AI. Same 321-leader Gartner sample. Set this against the 92% AI adoption figure at the top of this page. (Gartner, 2026)
- 68% of agents feel proud of their role and 75% would recommend the job to a friend. Based on 500+ contact centre agents across Europe and North America. (Calabrio, Voice of the Agent Report, 2026)
- Only 52% of agents received a pay increase in the past year, down from 79% in the previous edition. Same Calabrio agent sample. (Calabrio, 2026)
- Only 38% of contact centres formally measure agent satisfaction and wellbeing. ICMI State of the Contact Center research. (ICMI, 2024)
- 71% of service reps who use AI report having growth opportunities, and 86% say they have developed new skills because of it. Based on 6,500 service professionals across 39 countries. (Salesforce, 2025)
- 71% of UK contact centre leaders say three quarters or more of their advisers work from home at least part of the week. Based on 100 UK contact centre directors and managers. (Pitch Market Surveys, The Forum and Cnect Wales, 2025)
- Only 6% of contact centres that allow home working run a fully remote or fully office model. The other 94% are hybrid. Same UK survey of 100 leaders. (Pitch Market Surveys, 2025)
Pay and Employment (US)
- US customer service representatives earn a median $44,770 a year, or $21.53 an hour, across 2,666,000 jobs. Employment is projected to decline 5% between 2025 and 2035. (US Bureau of Labor Statistics, Occupational Outlook Handbook, 2025 to 2026 edition)
- US receptionists and information clerks earn a median $38,010 a year, or $18.27 an hour, across 947,500 jobs. Employment is projected to decline 2% between 2025 and 2035. (US Bureau of Labor Statistics, 2025 to 2026 edition)
- 440,615 people were employed in the US telemarketing and call centres industry. IBISWorld industry employment data. (IBISWorld, 2025)
- The Philippines IT-BPM sector employs close to 1.9 million workers and generates nearly $40 billion in export revenue. Industry association figures. (IBPAP, 2025)
Call Center Industry Statistics: Cost and Market Size
A warning before the numbers: research firms disagree sharply here, sometimes by tens of billions of dollars for the same segment in the same year. Cite one firm per claim and name it. Never average them.
- The average cost of an inbound call in UK contact centres is £5.58, which is 57% more than email and 83% more than web chat. Based on 200+ UK contact centre operations. The newer edition of this report is paywalled, so this figure is ageing. (ContactBabel, 2024)
- The global call and contact centre outsourcing market is projected to reach $163.86 billion by 2030, a 9.8% CAGR. Market forecast. (Grand View Research, 2025)
- The contact centre outsourcing market is valued at $114.98 billion in 2025, reaching $189.49 billion by 2031 at an 8.55% CAGR. A different firm, a different base value. The gap against the figure above is roughly $50 billion. (Mordor Intelligence, 2025)
- The global CCaaS market is projected to grow from $6.8 billion in 2025 to $32.7 billion by 2033, a 22.3% CAGR. Market forecast. (Grand View Research, 2025)
- The global CCaaS market is projected to grow from $8.33 billion in 2026 to $30.15 billion by 2034, a 17.4% CAGR. A second firm on the same segment, to show how wide the spread is. (Fortune Business Insights, 2026)
- The global business process outsourcing market is projected to reach $695.77 billion by 2033, a 9.9% CAGR. Market forecast, 2025 base year. (Grand View Research, 2025)
Call Center Statistics You'll See Quoted That You Shouldn't Trust
These numbers appear on dozens of pages, often with a confident-looking attribution. We could not trace any of them to a named study with a disclosed sample and date. If you are putting a business case in front of a CFO, leave them out.
| The statistic | Why it doesn't hold up |
|---|---|
| "Chatbots achieve 70% to 90% containment rates" | Repeated across many vendor blogs as an industry benchmark. No named survey, sample size or research body behind any instance we could find. It is a vendor rule of thumb, not research. |
| "62% of small business calls go unanswered" | Attributed everywhere to a 411 Locals study. The underlying sample appears to be 85 businesses with no disclosed year or methodology, and the same 62% figure is separately used for a different claim, which suggests one has been miscopied. |
| "The average small business loses $126,000 a year to missed calls" | The cited source page returns a 404. It is described elsewhere as an estimate rather than a study. Every article quoting it cites either that dead page or another article citing it. |
| "It costs $1 to handle a call through IVR versus $5 to $6 with a live agent" | Ubiquitous across vendor content. No original named source found. |
| "Industry-standard call centre attrition is 30% to 45%" | A folk average with no single traceable study. Use the Deloitte 76% figure above, which has a stated sample, or your own data. |
| "96% of customers say service affects brand loyalty" | Real, but from Microsoft's 2017 State of Global Customer Service Report. Nearly a decade old and routinely presented as current. |
| "One in three consumers walk away after one bad experience" | Real, from PwC, but published in 2018. The Zendesk 2025 figure of 63% is both newer and stronger. |
| "Cost to replace a contact centre agent: $14,113" | Attributed to a Gartner report, but traceable only to a single secondary blog with no Gartner link or publication year. |
There is also no authoritative figure for the total number of contact centre agents employed worldwide. National figures exist for the US and the Philippines, as listed above. Any global headcount number you see is an extrapolation.
What These Numbers Mean If You Run a Contact Center
Four things fall out of the data above, and none of them is the one the AI headlines suggest.
- Adoption has outrun results. 98% of contact centres use AI and 20% of leaders have reduced staffing because of it. The gap is implementation quality, not model quality.
- The handoff is the weak point, not the AI. 83% of consumers repeat themselves after being passed from AI to a human. That is a context-passing problem, and it is solvable this quarter.
- Speed beats channel preference. 74% of people prefer the phone for urgent issues, and 59% would rather use a chatbot than wait. They are not contradicting themselves: they want a person, quickly, and will take a machine over a queue.
- Wait time is where the money leaks. 86% expect an agent inside ten minutes, most wait far longer, and 63% will switch after one bad experience. Concurrency fixes this faster than headcount does.
CloudTalk's AI receptionist answers unlimited simultaneous calls, so the queue disappears, and hands context to the human agent when it escalates, so nobody repeats their account number. Every call is recorded and transcribed, which is how you get your own numbers instead of borrowing everyone else's. Both of those are the problems the research above keeps pointing at.
FAQs: Call Center Statistics
For most operators it is this one: 86% of consumers expect to reach a live agent within ten minutes, yet more than 60% report waiting 15 to 60 minutes or longer (Genesys, 2025). Almost every other number on this page follows from it. People abandon, they switch, and they take a chatbot over a queue, all because the wait is longer than they expected.
98% of contact centres use AI in some form, according to Calabrio's 2025 survey of more than 400 global contact centre leaders, and 92% of organisations have implemented or piloted AI in customer service according to Five9's 2026 research. Adoption is close to universal. Impact is not: only 20% of service leaders told Gartner in 2026 that they have actually reduced agent staffing because of AI.
Human agents, consistently and by a wide margin. Metrigy's 2025 study of 503 North American consumers found 84.9% prefer a human, and 80.1% still prefer a human even when told AI would resolve the issue equally well. The nuance that matters: 59% would rather use an instant chatbot than wait for a live agent (Five9, 2025). Customers want a person, but not enough to sit in a queue for one.
Deloitte Digital's survey of 600 contact centre leaders put annual agent attrition at 76%, and ICMI found only 54% of agents stay past two years. Be careful with the "30% to 45% industry standard" figure that circulates widely: we could not trace it to any named study, and it is roughly half what the best-sourced research reports.
In the US, the median customer service representative earns $44,770 a year, or $21.53 an hour, across 2.67 million jobs, according to the Bureau of Labor Statistics 2025 to 2026 Occupational Outlook Handbook. On a per-contact basis, ContactBabel put the average UK inbound call at £5.58, which is 57% more than email and 83% more than web chat, though that figure dates from 2024 and the newer edition is paywalled.
It depends entirely on which firm you ask, which is why this page names them individually. Grand View Research puts call and contact centre outsourcing at $163.86 billion by 2030, while Mordor Intelligence values the same market at $114.98 billion in 2025 rising to $189.49 billion by 2031. The CCaaS segment sits between $6.8 billion and $8.33 billion depending on the source. Cite one firm, name it, and do not average them.
Yes. Every statistic here links to the organisation that ran the research, and that is who you should credit. If you want to reference this roundup as well, the suggested citation is in the section above. If you spot a figure here that has gone out of date or that you can trace to a better primary source, tell us and we will update it.
