The answered-call gap: what the evidence says about small businesses, phones and AI

1. The phone is still where urgent business starts

Customers say they prefer the phone for complex or urgent issues, and in one profession tested, fewer firms answered than five years earlier.

Five9, which sells contact center software, surveyed 1,006 consumers in the US, UK and Canada with Zogby Analytics, and published the results in March 2025. It found that 56% prefer the phone for general issues. For complex or urgent issues, 74% prefer it. This is a vendor survey, so it measures what people say they prefer, not what they do.

A second signal comes from a mystery shop. Clio, which sells legal software, hired a third-party research company in 2024 to phone and email 500 US law firms with an inquiry. Only 40% of firms answered the phone call, down from 56% in 2019. Clio adds that 52% of firms either answered or phoned back, so 48% were essentially unreachable by phone.

The two sources point the same way. People with urgent or complex needs say they want a voice, and in one profession that was tested, fewer firms are picking up than five years earlier. The Clio sample covers law firms only. It cannot speak for dental offices, trades or retail.

Sources: 1, 2.

2. Many calls never reach a person

Vendor data from Invoca and Clio shows a large share of business calls did not reach a person, though the two figures measure different things.

Invoca, which sells call analytics software, published its 2026 Lead Conversion Benchmarks Report in July 2026. It is based on anonymized data from over 70 million calls and 600 million minutes. Its headline: "Across all industries, 56% of calls to businesses are answered by a person." That figure covers all inbound calls. For calls longer than 15 seconds, which filters out misdials and quick hang-ups, Invoca reports an answer rate of 65%.

The report does not say what happened to the rest. It makes no statement about voicemail, and it has no breakdown for dental or healthcare practices. Treat the number as a broad reference point, not a measure of any one trade.

Clio's mystery shop gives a second reading, with a different method. Where Invoca counts calls on its own platform, Clio's researchers placed calls to 500 law firms themselves. In that test, 40% of firms answered the phone, and 52% answered or phoned back.

Both sources come from vendors, and they measure different things: calls to businesses that use Invoca's software, and calls placed to law firms by a research company. The figures are not interchangeable. What they share is a result that is hard to ignore: a large share of business calls did not reach a person. Section 8 explains why independent data on this question is still missing.

Sources: 2, 3.

3. Front desks are stretched, not empty

Hiring front desk staff is hard for dentists who try, yet most say their front desk is adequately staffed, and receptionist jobs are not growing.

The American Dental Association's Health Policy Institute polled dentists in the first quarter of 2026 (796 responses). It found that 28.0% of dentists had recruited administrative staff in the last three months. Of those, 58.0% said recruiting was very or extremely challenging. That 58.0% is a share of the 28.0% who recruited, not of all dentists.

The same poll adds balance. A total of 79.3% of dentists said their practice currently has an adequate number of administrative staff. One third said they were not busy enough. Its summary slide also notes one fifth not having enough administrative staff. So hiring is hard for the dentists who try, and most describe their front desk as adequately staffed.

The wider labour outlook is flat to down. The Bureau of Labor Statistics (BLS) reports that receptionists held about 947,500 US jobs in 2025. It projects employment to decline 2% from 2025 to 2035. It also projects about 105,100 openings a year, on average, over the decade, all from replacing workers who leave.

Taken together, the picture is not an empty front desk. It is a role that is hard to fill when a vacancy opens, in a field where job counts are not growing.

Sources: 4, 5.

4. Big call centers are shrinking first

Global call center headcount has fallen for eight straight quarters, and the BLS projects fewer customer service jobs, without naming AI as the cause.

Revelio Labs, a labour-data firm, published an analysis on October 6, 2026. It found that global call center headcount has fallen for eight straight quarters. As stated by Revelio, headcount is 5.3% below its December 2023 peak. Of 70 quarter-end readings since 2009, eight showed a yearly decline, and all eight are in the last eight quarters. The data are Revelio's own headcount, wage and job-transition records, not a government census.

Revelio's authors say they cannot observe whether firms replace workers with chatbots. The data show a shrinking workforce, not a cause.

Government projections lean the same way. The BLS says customer service representatives held about 2.7 million US jobs in 2025. It projects a 5% decline from 2025 to 2035, with about 289,500 openings a year. Its page does not mention AI. A separate BLS article from February 2025 names customer service representatives among occupations projected to decline 5.0 percent through 2033. The same article says AI is expected to primarily affect occupations whose core tasks can be most easily replicated by generative AI in its current form.

The pattern is consistent: large, high-volume phone operations are changing first. The evidence does not yet show how much of that change comes from AI.

Sources: 6, 7, 8.

5. Small businesses are last to adopt AI

In both the US and Canada, larger firms report more AI use than the smallest businesses, though the two surveys are not directly comparable.

The US Census Bureau's Business Trends and Outlook Survey is a biweekly, nationally representative survey of employer businesses. From December 2025 to May 2026, it shows that overall AI use hovered between 17% and 20%. By size, 37% of firms with at least 250 employees reported using AI in their business operations. For firms with 100 to 249 employees, the figure was 32% in the period ending May 3, 2026. Less than 20% of firms with four or fewer employees reported using AI.

Statistics Canada shows a similar pattern with different wording. In the second quarter of 2026, 19.2% of Canadian businesses reported using AI to produce goods or deliver services over the previous 12 months. For businesses with 1 to 4 employees, the figure was 19.9%. For those with 100 or more employees, it was 27.8%.

The two surveys ask different questions over different periods, so compare the pattern, not the levels. In both countries, larger firms report more AI use than the smallest.

Plans are rising in Canada. In the third quarter of 2026, 25.2% of businesses said they plan to use AI over the next 12 months, up from 14.5% in the third quarter of 2025.

Sources: 9, 12, 13, 14.

6. Many businesses do not see AI as relevant yet

Two in five Canadian businesses say AI is not relevant, and 7.6% of US employer businesses used AI for customer service in the last six months.

In the same Statistics Canada survey, 2 in 5 businesses (40.0%) said the use of AI is not relevant to their business. By size, that view was held by 41.4% of businesses with 1 to 4 employees, 41.0% of those with 5 to 19, and 34.3% of those with 20 to 99.

This is the businesses' own view, and it is a counterpoint to any claim that adoption is simply a matter of time. Plans do show movement. In the third quarter of 2026, 52.7% of Canadian businesses said they did not plan to use AI, and 22.1% were unsure.

The US numbers show how early customer service use is. In a Census supplement collected from November 17, 2025 to February 8, 2026, 7.6% of US employer businesses said they had used AI for customer service in the last six months. A total of 85.4% said no, and 7.0% did not know. The 7.6% is a share of all employer businesses, not of businesses that use AI.

The denominators matter. In an earlier Statistics Canada release from the third quarter of 2025, 78.1% of businesses that did not plan to adopt AI said it was not relevant. That figure applies only to businesses with no plans, and it is a different group from the 40.0% of all businesses above.

Sources: 10, 11, 12, 13, 14.

7. Customers are open to AI when a person stays reachable

Customers appear open to AI when a person stays reachable, but a substantial minority say learning a chatbot was involved would make things worse.

Five9's 2025 survey found that 72% of consumers are open to AI if they can reach a human. That is a vendor result, and it describes a condition, not a blanket welcome.

The Pew Research Center adds an independent reading. In a survey of 5,023 US adults, fielded June 9 to 15, 2025, 41% said they would feel worse about a customer service experience if they later realized they were talking with an AI chatbot. A larger share, 53%, said their views would not change.

Both findings deserve weight. A majority in the Pew sample said learning that a chatbot was involved would not change their view. A substantial minority said it would make things worse. Pew's question is also about learning afterwards, in a hypothetical, not about an actual call.

For a business, the practical reading is cautious. Disclosure and a clear path to a person are the conditions that appear in the evidence. The evidence does not show that all callers accept AI.

Sources: 1, 15.

8. What the data does not show yet

We found no independent source for after-hours call volume, AI versus human booking, front desk turnover or what callers do after voicemail.

Several of the questions a business owner would most like answered have no independent evidence that we could verify.

There is no independent data on how many business calls arrive after hours. There is no independent comparison of AI and human outcomes for booking appointments. There is no turnover data for clinic front desks. The closest public series, the BLS job openings survey, reports quits for all of health care and social assistance, not for dental front desks.

There is also no verified data from 2024 or later on what callers do after they reach voicemail. The answer-rate figures in this review all come from vendors. Revelio, for its part, cannot see whether chatbots are replacing call center workers.

These gaps matter because many widely shared statistics fill them without a traceable source. This review leaves those spaces open rather than guess. Better public data from statistics agencies and professional bodies would help businesses make clearer decisions.

Sources: 2, 3, 4, 8.