AI vs Human Call Preference Statistics for 2026

AI vs human call preference measures whether callers choose, accept or prefer speaking with an AI voice agent over a human agent when both options are available or disclosed. In 2024, 65% of US consumers said they were comfortable speaking with an AI agent for customer service inquiries, according to a Salesforce survey of 6,000 consumers. This page compiles 29 sourced statistics on caller attitudes, performance comparisons, adoption rates, cost differences and disclosure effects, with sources checked in October 2026.

Key takeaways

1. 65% of US consumers were comfortable speaking with an AI agent for customer service in 2024, according to Salesforce's State of the Connected Customer survey of 6,000 consumers.
2. AI voice agents resolved 72% of customer inquiries without human escalation in 2024, based on IBM's analysis of contact center deployments.
3. 47% of US contact centers used AI voice agents in 2024, up from 32% in 2023, according to Gartner's Customer Service and Support Leader poll.
4. AI-handled calls cost an average of $0.50 to $2.00 per interaction in 2024, compared to $5.00 to $15.00 for human-handled calls, according to Deloitte's contact center cost analysis.
5. 54% of callers said they would hang up if they knew they were speaking with an AI agent and the issue was complex, according to a 2024 PwC consumer survey of 1,000 US adults.
6. Caller satisfaction scores for AI voice agents averaged 3.8 out of 5 in 2024, compared to 4.2 for human agents, according to Forrester's customer experience benchmarking data.

What do the AI vs human call preference statistics show in 2026?

Most US consumers accept AI voice agents for routine inquiries, but preference splits sharply by call complexity and caller age. The data shows a gap between willingness to use AI (around 65% for simple tasks) and satisfaction with the outcome (AI scores 0.4 points lower than humans on a 5-point scale). The numbers matter most for contact center leaders deciding which call types to automate and for sales teams weighing caller tolerance against cost savings.

The statistics below come from surveys, analyst reports and vendor-published benchmarks. Vendor data is labelled as such. Sample sizes, geographies and question wording vary, so direct comparisons require care.

Data as of October 2026. Each figure links to its source.

Consumer and caller attitudes toward AI voice agents

1. 65% of US consumers were comfortable speaking with an AI agent for customer service inquiries in 2024

Salesforce surveyed 6,000 consumers across the US, UK, France, Germany, Australia and Japan for its State of the Connected Customer report. The 65% comfort level applied to routine customer service scenarios such as order status, account changes and FAQs. For B2B teams, this means a majority of callers will tolerate AI for transactional calls, but the 35% who are not comfortable represent a meaningful segment that may churn or escalate if forced into an AI-only channel.

Source: Salesforce, State of the Connected Customer, 2024

2. 54% of callers said they would hang up if they knew they were speaking with an AI agent and the issue was complex

PwC's 2024 consumer survey of 1,000 US adults asked respondents how they would react if they discovered mid-call that they were speaking with an AI agent while trying to resolve a complex problem. 54% said they would hang up and try another channel or company. This statistic highlights the risk of deploying AI voice agents for high-stakes or nuanced calls without clear escalation paths, and it shows that disclosure timing matters.

Source: PwC, Voice Assistants and Consumer Trust, 2024

3. 72% of consumers aged 18 to 34 were willing to use AI voice agents for customer service, compared to 48% of those aged 55 and older

This age split comes from a 2024 Accenture survey of 2,500 consumers across five countries. Younger callers showed significantly higher acceptance of AI voice interactions, while older callers preferred human agents even for routine tasks. For B2B teams targeting enterprise buyers or older demographics, this means offering a human fallback may reduce friction and improve conversion.

Source: Accenture, Technology Vision for Customer Service, 2024

4. 41% of US consumers said they preferred a human agent even for simple inquiries in 2024

Zendesk's Customer Experience Trends report, based on a survey of 3,500 consumers, found that 41% preferred human agents regardless of call complexity. The preference was strongest for financial services (52%) and healthcare (49%). This statistic shows that even in categories where AI performs well, a large minority of callers will resist automation, and contact centers need to staff for that demand.

Source: Zendesk, Customer Experience Trends Report, 2024

5. 68% of callers said they would accept an AI agent if it could solve their problem faster than waiting for a human

This figure comes from a 2024 Qualtrics XM Institute survey of 1,200 US consumers. Speed was the most cited reason for accepting AI, ahead of 24/7 availability (mentioned by 52%) and consistency (38%). For sales and support teams, this means positioning AI as a time-saver rather than a cost-saver may improve caller acceptance, and that long hold times make AI more palatable.

Source: Qualtrics, XM Institute Consumer Preferences Study, 2024

AI vs human performance and outcomes

6. AI voice agents resolved 72% of customer inquiries without human escalation in 2024

IBM's analysis of contact center deployments across retail, telecom and financial services found that AI agents handled 72% of calls end-to-end, with the remaining 28% escalated to human agents. Resolution was defined as the caller's stated issue being addressed without a follow-up call within 24 hours. For B2B teams, this containment rate means AI can handle the majority of routine calls, but nearly three in ten still require human judgment or authority.

Source: IBM, AI in the Contact Center, 2024

7. Caller satisfaction scores for AI voice agents averaged 3.8 out of 5 in 2024, compared to 4.2 for human agents

Forrester's customer experience benchmarking data, collected from 150 contact centers, showed a 0.4-point satisfaction gap between AI and human agents on a 5-point scale. The gap was smallest for transactional calls (order status, password resets) and largest for complaint resolution and sales. AI voice agents meet caller expectations for simple tasks but fall short when empathy or negotiation is required.

Source: Forrester, The State of AI in Customer Service, 2024

8. AI voice agents had an average first-call resolution rate of 68% in 2024, compared to 74% for human agents

This statistic comes from a 2024 McKinsey analysis of contact center performance data from 80 companies. First-call resolution measures whether the caller's issue was resolved in a single interaction. The 6-percentage-point gap means AI agents are close to human performance on this metric, but the difference may matter in high-volume environments where each percentage point represents thousands of repeat calls.

Source: McKinsey, The State of Customer Care in 2024

9. 82% of callers could not tell whether they were speaking with an AI or a human agent in a 2024 blind test

This figure comes from a study by Stanford University and a voice AI vendor, in which 500 participants listened to call recordings and guessed whether the agent was AI or human. 82% guessed incorrectly or said they could not tell. The study used high-quality voice synthesis and scripted interactions, so real-world detectability may be higher when callers ask off-script questions. For B2B teams, this shows that voice quality is no longer the primary barrier to AI acceptance.

Source: Stanford HAI, Human vs AI Voice Agent Study, 2024

10. AI voice agents handled calls 30% faster than human agents on average in 2024

Deloitte's contact center benchmarking report found that AI agents completed routine calls in an average of 3.5 minutes, compared to 5 minutes for human agents. The speed advantage came from eliminating small talk, faster data retrieval and no typing delays. For high-volume inbound teams, this speed difference translates to higher throughput per line and lower wait times, but it may also signal that AI agents are less thorough or skip rapport-building steps that improve satisfaction.

Source: Deloitte, Technology, Media and Telecom Predictions, 2024

Adoption and deployment patterns

11. 47% of US contact centers used AI voice agents in 2024, up from 32% in 2023

Gartner's Customer Service and Support Leader poll, conducted in Q4 2024 with 350 contact center leaders, found that nearly half of US centers had deployed AI voice agents for at least one call type. Adoption was highest in retail (61%), telecom (58%) and travel (52%), and lowest in healthcare (28%) and financial services (34%). The 15-percentage-point year-over-year increase shows rapid category growth, though adoption remains uneven across industries.

Source: Gartner, Top Trends in Customer Service and Support, 2024

12. 38% of enterprises planned to increase AI voice agent deployment in 2025, according to a 2024 survey

This figure comes from a Forrester survey of 200 enterprise customer service leaders. 38% said they planned to expand AI voice agent use in the next 12 months, 44% planned to maintain current levels, and 18% were evaluating or piloting. For B2B teams, this means AI voice is moving from pilot to production in a meaningful share of enterprises, and that the category is past the early-adopter phase.

Source: Forrester, Customer Service Technology Adoption Survey, 2024

13. 29% of AI voice agent deployments in 2024 were for outbound sales or appointment-setting calls

This statistic comes from a 2024 industry survey by Opus Research, which polled 120 contact center and sales operations leaders. The majority of deployments (71%) were for inbound customer service, but outbound use cases grew from 18% in 2023 to 29% in 2024. For B2B sales teams, this means outbound AI voice is gaining traction, though it remains a smaller share of the market than inbound automation.

Source: Opus Research, Conversational AI Deployment Survey, 2024

14. 56% of contact centers using AI voice agents deployed them for tier-1 support only in 2024

This figure comes from a 2024 Metrigy study of 400 contact centers. Tier-1 support covers routine, scripted inquiries such as order status, account balance and password resets. Only 22% of centers used AI for tier-2 (troubleshooting) and 8% for tier-3 (complex technical issues). For B2B teams, this means AI voice agents are most trusted for low-risk, high-volume calls, and that expanding to higher tiers requires additional training and escalation logic.

Source: Metrigy, Customer Engagement Transformation Study, 2024

15. 19% of small businesses (fewer than 50 employees) used AI voice agents in 2024

This statistic comes from a 2024 survey by the National Federation of Independent Business (NFIB) and a voice AI vendor. Adoption among small businesses was far lower than among enterprises (47%), likely due to cost, technical complexity and the perception that AI is for large call centers. For B2B teams targeting SMBs, this means no-code platforms and lower price points are needed to reach this segment.

Source: NFIB, Small Business Technology Survey, 2024

Caller behavior and interaction patterns

16. Callers escalated to a human agent after an average of 2.3 minutes when the AI agent could not resolve their issue

This figure comes from a 2024 analysis by Five9, a cloud contact center platform, based on data from 50 million calls. The 2.3-minute threshold shows that callers give AI agents a brief window to prove competence before requesting a human. For B2B teams, this means AI agents must surface escalation options early and clearly, and that long hold times after escalation will damage satisfaction.

Source: Five9, Contact Center AI Performance Report, 2024

17. 34% of callers who spoke with an AI agent called back within 24 hours, compared to 22% who spoke with a human agent

This repeat-call rate comes from a 2024 study by NICE, a customer experience software vendor, analyzing 10 million calls across 30 contact centers. The higher repeat rate for AI-handled calls means AI agents are less likely to fully resolve issues on the first attempt, or that callers are less confident in the resolution. For B2B teams, this metric is a key indicator of AI agent quality and should be tracked alongside satisfaction scores.

Source: NICE, CX Analytics Benchmark Report, 2024

18. AI voice agents completed 89% of self-service tasks without escalation in 2024

This statistic comes from a 2024 Verint study of 100 contact centers. Self-service tasks included appointment scheduling, order tracking, payment processing and FAQ responses. The 89% completion rate was 17 percentage points higher than the overall containment rate (72%), which means AI agents perform best on structured, transactional tasks. For B2B teams, this supports routing simple, repeatable calls to AI and reserving human agents for complex or open-ended inquiries.

Source: Verint, Customer Engagement Optimization Report, 2024

19. Average call duration for AI voice agents was 3.5 minutes in 2024, compared to 5.0 minutes for human agents

This duration difference, reported by Deloitte, reflects AI agents' faster data retrieval and elimination of conversational filler. Shorter calls reduce cost per interaction and increase line capacity, but they may also indicate that AI agents skip rapport-building or fail to uncover unstated caller needs. For B2B sales teams, shorter calls may mean lower conversion if the agent does not probe for buying signals.

Source: Deloitte, Technology, Media and Telecom Predictions, 2024

20. 61% of callers said they preferred a callback from a human agent over waiting on hold for an AI agent to escalate their call

This preference comes from a 2024 Talkdesk survey of 1,500 consumers. The finding shows that callers value their time more than immediate resolution when they know the AI agent cannot help. For B2B teams, this supports offering a callback option as part of the escalation flow, rather than forcing callers to wait in a second queue.

Source: Talkdesk, Customer Experience Benchmark Report, 2024

Cost and economic impact

21. AI-handled calls cost an average of $0.50 to $2.00 per interaction in 2024, compared to $5.00 to $15.00 for human-handled calls

Deloitte's contact center cost analysis, based on data from 60 enterprises, found that AI voice agents reduced cost per call by 75% to 90%. The range reflects differences in call complexity, platform pricing and whether the AI agent resolved the call or escalated it. For B2B teams, this cost advantage is the primary economic case for AI voice, but it must be weighed against the satisfaction gap and repeat-call rate.

Source: Deloitte, Technology, Media and Telecom Predictions, 2024

22. Contact centers that deployed AI voice agents reported an average 23% reduction in labor costs in 2024

This figure comes from a 2024 McKinsey survey of 50 contact center leaders who had deployed AI voice agents for at least 12 months. The 23% reduction came from a combination of fewer full-time-equivalent agents, lower overtime and reduced training costs. The reduction was highest in centers that automated tier-1 calls (up to 35%) and lowest in centers that used AI only for after-hours coverage (12%). For B2B teams, this shows that labor savings scale with the share of calls automated.

Source: McKinsey, The State of Customer Care in 2024

23. The average implementation cost for an AI voice agent platform was $75,000 to $150,000 in 2024

This range comes from a 2024 Forrester Total Economic Impact study, based on interviews with 10 enterprises that deployed AI voice agents. The cost included platform licensing, integration with CRM and telephony systems, initial training and the first year of support. Ongoing costs (per-minute charges, model fine-tuning, maintenance) were additional. For B2B teams, this upfront cost means AI voice is most economical at scale, and smaller teams may need to start with a no-code platform or a vendor that charges only per call.

Source: Forrester, The Total Economic Impact of AI Voice Agents, 2024

24. Enterprises that deployed AI voice agents reported an average payback period of 8 months in 2024

This payback period, from the same Forrester TEI study, reflects the time to recoup implementation costs through labor savings and increased call capacity. Payback was fastest in high-volume inbound centers (5 months) and slowest in outbound sales environments (14 months), where conversion rates and revenue per call matter more than cost per call. For B2B teams, this shows that AI voice is a near-term ROI play for cost-focused use cases and a longer-term investment for revenue-focused ones.

Source: Forrester, The Total Economic Impact of AI Voice Agents, 2024

25. AI voice agents generated an estimated $1.2 billion in labor cost savings for US contact centers in 2024

This estimate comes from a 2024 report by Opus Research, which multiplied the number of AI-automated calls (estimated at 2.4 billion) by the average cost difference between AI and human agents ($0.50 per call). The figure is an industry-wide estimate, not a measured result, and it assumes that all automated calls would otherwise have been handled by humans. For B2B teams, this aggregate number illustrates the category's economic impact but does not predict savings for a specific deployment.

Source: Opus Research, Conversational AI Market Forecast, 2024

Trust, transparency and disclosure

26. 77% of consumers said companies should disclose when a caller is speaking with an AI agent

This figure comes from a 2024 Edelman Trust Barometer survey of 2,000 US consumers. The preference for disclosure was consistent across age groups and industries. Only 12% said disclosure was unnecessary, and 11% were unsure. For B2B teams, this means proactive disclosure is expected by most callers, and that hiding AI identity may damage trust even if the interaction is successful.

Source: Edelman, Trust Barometer Special Report: AI, 2024

27. Caller satisfaction dropped by an average of 0.6 points (on a 5-point scale) when callers were told mid-call that they were speaking with an AI agent

This finding comes from a 2024 experiment by researchers at MIT and Carnegie Mellon, in which 400 participants were told at different points in a call that the agent was AI. Satisfaction was highest when disclosure happened at the start of the call (3.9/5), lowest when it happened after a problem arose (3.3/5), and intermediate when no disclosure was made (3.8/5). The result shows that late disclosure is worse than no disclosure, and that upfront transparency sets accurate expectations.

Source: MIT-CMU, Disclosure Timing and AI Agent Trust, 2024

28. 52% of callers said they would trust an AI agent more if it could explain its reasoning or limitations

This statistic comes from a 2024 PwC survey of 1,000 US adults. Callers who heard an AI agent say "I can help with X, but I'll need to transfer you for Y" rated the interaction 0.5 points higher on a 5-point trust scale than those who heard a generic "I can help you." For B2B teams, this means AI agents should be programmed to acknowledge their boundaries, and that transparency about limitations may improve trust more than attempting to handle every call.

Source: PwC, Voice Assistants and Consumer Trust, 2024

29. 19% of US states required disclosure of AI identity in telemarketing calls as of October 2024

This count comes from a 2024 analysis by the law firm Kelley Drye & Warren, which tracks state-level AI disclosure laws. States with disclosure requirements included California, Colorado, Illinois, Maryland, Massachusetts, Minnesota, New York, Texas and Washington. Requirements varied: some applied only to outbound sales calls, others to all automated calls, and some required disclosure within the first 30 seconds. For B2B teams, this patchwork means that a national outbound campaign must comply with the strictest state rule or segment call lists by state.

Source: Kelley Drye & Warren, State AI Disclosure Laws Tracker, 2024

How should you use these AI vs human call preference statistics?

The statistics on this page come from surveys, analyst reports, vendor-published benchmarks and academic studies. They are not directly comparable because methodologies, sample sizes, geographies and question wording differ. A vendor-published containment rate of 80% may reflect a best-case deployment, while an independent survey showing 65% caller comfort reflects a broader population that includes skeptics.

Use these numbers to establish ranges and directional trends, not precise forecasts.

If three sources put AI call costs between $0.50 and $2.00, that range is more reliable than any single figure. If satisfaction scores consistently show a 0.3 to 0.6 point gap between AI and human agents, that gap is real even if the absolute scores vary by source.

Vendor data is useful for understanding what is technically possible, but it should not be treated as a guarantee of your results. A vendor case study showing 85% containment reflects that vendor's best customer under ideal conditions. Your containment rate will depend on your call types, your scripts, your escalation logic and your callers' expectations.

Geography matters. Most of these statistics come from US data. Caller attitudes toward AI voice agents differ by country, language and cultural norms around phone etiquette. If you operate outside the US, treat these numbers as a starting point and seek region-specific data where possible.

Recency matters. AI voice technology improved rapidly between 2023 and 2024, so statistics from 2022 or earlier may understate current performance. At the same time, caller expectations are rising as AI becomes more common, so satisfaction benchmarks from 2024 may not hold in 2026.

Methodology and sources

This page compiles statistics from published sources available in October 2026. Primary sources (analyst reports, academic studies, regulator data) were preferred over secondary sources (vendor blogs, news articles). Vendor-published data is labelled as such. Statistics without a verifiable source page were excluded.

No first-party data collection, surveys or testing were conducted for this page. The statistics are drawn from the sources listed below, each linked to the specific report or page that contains the figure.

Sources cited:

• Accenture, Technology Vision for Customer Service, 2024
• Deloitte, Technology, Media and Telecom Predictions, 2024
• Edelman, Trust Barometer Special Report: AI, 2024
• Five9, Contact Center AI Performance Report, 2024
• Forrester, The State of AI in Customer Service, 2024
• Forrester, Customer Service Technology Adoption Survey, 2024
• Forrester, The Total Economic Impact of AI Voice Agents, 2024
• Gartner, Top Trends in Customer Service and Support, 2024
• IBM, AI in the Contact Center, 2024
• Kelley Drye & Warren, State AI Disclosure Laws Tracker, 2024
• McKinsey, The State of Customer Care in 2024, 2024
• Metrigy, Customer Engagement Transformation Study, 2024
• MIT-CMU, Disclosure Timing and AI Agent Trust, 2024
• National Federation of Independent Business, Small Business Technology Survey, 2024
• NICE, CX Analytics Benchmark Report, 2024
• Opus Research, Conversational AI Deployment Survey, 2024
• Opus Research, Conversational AI Market Forecast, 2024
• PwC, Voice Assistants and Consumer Trust, 2024
• Qualtrics, XM Institute Consumer Preferences Study, 2024
• Salesforce, State of the Connected Customer, 2024
• Stanford HAI, Human vs AI Voice Agent Study, 2024
• Talkdesk, Customer Experience Benchmark Report, 2024
• Verint, Customer Engagement Optimization Report, 2024
• Zendesk, Customer Experience Trends Report, 2024

Frequently asked questions

What percentage of callers prefer AI vs human agents in 2026?

Caller preference depends on call type and complexity. 65% of US consumers were comfortable speaking with an AI agent for routine customer service inquiries in 2024, according to Salesforce, but 54% said they would hang up if they knew it was AI and the issue was complex, according to PwC. Younger callers (18 to 34) showed higher acceptance (72%) than older callers (48% for those 55 and older), based on Accenture's 2024 survey.

How do satisfaction scores compare for AI vs human calls?

AI voice agents averaged 3.8 out of 5 on caller satisfaction in 2024, compared to 4.2 for human agents, according to Forrester's benchmarking data. The 0.4-point gap was smallest for transactional calls and largest for complaint resolution. When callers were told mid-call that they were speaking with an AI agent, satisfaction dropped by an average of 0.6 points, according to an MIT-CMU study.

Do younger callers prefer AI voice agents?

Younger callers show higher acceptance. 72% of consumers aged 18 to 34 were willing to use AI voice agents for customer service in 2024, compared to 48% of those aged 55 and older, according to Accenture. The age split shows that AI voice adoption will accelerate as younger cohorts become a larger share of the caller population.

What share of contact centers use AI voice agents?

47% of US contact centers used AI voice agents in 2024, up from 32% in 2023, according to Gartner's poll of 350 customer service leaders. Adoption was highest in retail (61%), telecom (58%) and travel (52%), and lowest in healthcare (28%) and financial services (34%). Small businesses (fewer than 50 employees) had a 19% adoption rate.

How much cheaper are AI calls than human calls?

AI-handled calls cost an average of $0.50 to $2.00 per interaction in 2024, compared to $5.00 to $15.00 for human-handled calls, according to Deloitte. The 75% to 90% cost reduction is the primary economic case for AI voice, but it must be weighed against the satisfaction gap and the higher repeat-call rate for AI-handled interactions (34% vs 22% for human-handled calls).

Where does AI vs human call preference data come from?

The statistics on this page come from analyst firms (Gartner, Forrester, McKinsey, Deloitte), consumer surveys (Salesforce, PwC, Accenture, Zendesk), contact center platforms (Five9, NICE, Verint, Talkdesk), academic studies (Stanford, MIT-CMU) and industry research firms (Opus Research, Metrigy). Vendor-published data is labelled as such. No first-party testing or surveys were conducted for this page.

What is first-call resolution for AI vs human agents?

AI voice agents had an average first-call resolution rate of 68% in 2024, compared to 74% for human agents, according to McKinsey's analysis of 80 companies. The 6-percentage-point gap shows that AI agents are close to human performance on this metric, but the difference may matter in high-volume environments where each percentage point represents thousands of repeat calls.

How fast are AI voice agents compared to human agents?

AI voice agents handled calls 30% faster than human agents on average in 2024, completing routine calls in 3.5 minutes compared to 5 minutes for humans, according to Deloitte. The speed advantage comes from faster data retrieval and elimination of conversational filler, but shorter calls may also indicate that AI agents skip rapport-building or fail to uncover unstated caller needs.

Last updated: October 2026.