White Label AI Receptionist: What the Research Says

How many calls go unanswered?

There is no single authoritative number, and the most quoted one is the weakest. The widely repeated claim that 62% of business calls go unanswered traces to a 30 day test by the marketing firm 411 Locals in January 2016. It covered 85 businesses in 58 industries and found 37.8% of calls answered by a person, 37.8% reaching an answering machine and 24.3% unanswered. The firm counted voicemail as unanswered because no conversation can happen on it ¹. That is a small, vendor run sample that is now ten years old.

Larger and more recent measurements give lower, more varied figures. Invoca, a call analytics vendor, reported in November 2021 that across its customers' aggregate data an average of 26% of calls from potential customers went unanswered, with rates reaching 60% in some industries. The release does not state the number of calls analyzed ³. Moneypenny, a call answering provider, reported in 2016 that 33% of businesses in its survey of 300 micro businesses failed to answer incoming calls, and that 71% of them let calls go to voicemail rather than ring out ².

A 2026 synthesis by PCN, a company that sells answering services, concludes that small and mid sized businesses miss roughly 25% to 60% of inbound calls. It has no sample of its own and combines the sources above with others, so it adds range rather than evidence ¹².

What callers do when nobody answers

The measurements agree on one point: voicemail recovers very little. Moneypenny found that 69% of callers do not leave a message ². Invoca's customer data found that only 2% of callers to unanswered lines left a voicemail ³. Both are vendor studies, and both are several years old.

Surveys that ask people what they say they do give a rosier picture. A CallRail survey of 1,000 US consumers in 2025 is reported to have found that 42% say they leave a voicemail when a business does not answer, that 21% immediately call another business and that 78% say they have abandoned a business after an unanswered call ⁴. We could not find CallRail's own report and read these figures only through a secondary write up, so treat them as unverified. They are also self reported, which is a different thing from measured behavior. The gap between 42% who say they leave a message and 2% measured is the reason this page puts more weight on the call data.

A 2026 Hiya survey of more than 12,000 consumers and 3,600 business professionals in six countries reports, for Canada, that 87% of business professionals have lost a deal or customer because they could not reach them by phone and that unanswered calls cost businesses an average of more than CA$120,000 a year. 88% of Canadian consumers ignore calls from unrecognized numbers ⁵. The article gives no Canada only sample size or method, and Hiya sells caller identification services.

Two industries with testing: law firms and dentists

The best designed measurement we found is Clio's secret shopper work on law firms. The Oklahoma Bar Journal, summarizing Clio's 2024 Legal Trends Report, reports that the share of firms answering a prospective client's call fell from 56% in 2019 to 40% in 2024, that only 20% of firms that missed a call returned it, that 48% of firms were unreachable by phone, meaning they neither answered nor returned messages, and that 64% of prospective clients received no follow up by phone or email. 73% of the secret shoppers said they would not recommend the firms they contacted, against a 39% recommendation rate among those who reached a real person by phone ⁶. The article does not give the sample sizes, and we read it, not the Clio report.

For dental practices we found no independent study. The most specific dataset is a vendor's: Peerlogic tracked 4,280 calls across 26 practices in February 2026, found that 38% went unanswered and that new patient calls converted at 25.24% against 55.77% for existing patients. Calls ending before a real conversation were the top reason for no booking ⁷. The company sells an AI assistant, does not explain how it defined answered or converted, and has no control group.

What a human front desk costs and what it covers

The US Bureau of Labor Statistics lists the median pay for receptionists at $38,010 a year, or $18.27 an hour, in May 2025. There were 947,500 receptionist jobs in 2025 and the bureau projects a 2% decline, about 16,000 jobs, from 2025 to 2035. Answering the telephone and taking messages or forwarding calls is a listed duty, and healthcare and social assistance employs about 42% of receptionists ⁸.

Wages are not the full cost of an employee, and the bureau figure is a median across all employers, so local pay will differ. The coverage problem is simpler: a week has 168 hours and a full time shift covers about 40 of them.

Owner time lost to admin

A 2023 Censuswide survey for Time etc, a virtual assistant company, asked 251 US founders of growth oriented businesses about their week. They reported that small admin tasks took 36% of an average 45.5 hour week ⁹. Multiplying gives about 16 hours, but that is our arithmetic and not a reported result. The sample is small and self selected, and the sponsor sells delegation.

A different survey of 837 US microbusinesses reports that administration combined with legal and compliance took on average 22.4% of owner time, covering staff, finances, scheduling and inventory ¹⁰. We read this through a search summary only.

Neither measures time spent answering phones, and we found no study that does. What they show is that admin is a large part of an owner's week and that estimates range from about a fifth to over a third depending on who is counted.

How fast do businesses respond to enquiries?

Phone calls are one channel; the best old measurement of enquiry speed is for web forms. A 2011 Harvard Business Review study sent a test web enquiry to 2,241 US companies. According to a secondary summary, 37% replied within an hour and 23% never replied, and the average among those that replied within 30 days was 42 hours ^11,13. The same summary reports that contact within an hour made a conversation with a decision maker nearly seven times as likely as contact an hour later ¹³. The article is paywalled, so these figures are reported, not independently checked.

What AI changes: the evidence

The strongest independent evidence is not about voice. A National Bureau of Economic Research study of 5,179 customer support agents found that agents using a generative AI assistant resolved 14% more issues per hour on average, and 34% more if they were novice or low skilled, with minimal effect for experienced workers. The assistant supported human agents ¹⁴.

We found no independent study that measures how many calls an AI phone agent recovers, how often callers hang up on one, or whether bookings rise compared with voicemail. The recovery figures that exist come from vendors, such as the 144 appointments and estimated $47,088 that Peerlogic reports for its own assistant in one month, with no stated method or comparison group ⁷.

What the evidence does not show

• The 62% unanswered figure rests on one 85 business test from 2016 by a marketing vendor, and the original write up could not be opened ^1,12.
• Moneypenny, Invoca and Peerlogic are vendors that sell answering or AI products, and none states a full method or sample for the figures used here ^2,3,7.
• The CallRail 2025 consumer figures were read through secondary write ups, and the original report was not found ⁴.
• The Clio figures reach us through a bar association article, with no sample sizes ⁶. Law firms and dental practices are not typical of all small businesses.
• The Hiya figures are for Canada, self reported, and come without a stated sample or method for the country cut ⁵.
• We found no independent measure of how much time owners or staff spend answering phones, or of the revenue lost to a missed call. Widely repeated annual loss figures, such as $126,000 per business, have no traceable primary source and are not used.
• We found no independent study of AI phone agents' results. The only AI evidence cited is about human agents using an assistant ¹⁴.

Applying this as an agency

The evidence suggests three checks before pitching an AI receptionist to a client. First, measure the client's own unanswered rate: most call tracking tools can report answered, missed and voicemail calls for a month, and that number beats any industry average on this page. Second, find out when calls are missed, such as lunch, evenings and weekends, because a human desk covers about 40 of 168 hours a week ⁸. Third, put a value on a booked call using the client's own average job or visit, since we found no trustworthy industry figure.

Because voicemail recovers so little ^2,3, the realistic goal is to answer the call, take the booking or message and hand urgent calls to a person. In a white label setup the agent answers under the client's business name and the transcript lands in a portal carrying the agency's brand. Convocore's plans for this are on the white label overview.

Method and limits

• Every number on this page comes from a source we opened and read on 9 October 2026, except where a source note below says we could only read a search summary.
• Sources are labelled primary (read at the original publisher), commissioned (paid for by a party with a commercial interest in the result), secondary (a report of the original) or forecast.
• Where sources disagreed or a figure could not be traced to an original, we left it out. Sample sizes and dates are given wherever the source gave them.
• Convocore sells white label AI agents, including a receptionist, and has a commercial interest in phone answering. We have tried to report the evidence as it is, including where it is weak. If you find an error, write to support@convocore.ai and we will correct it.

Sources

1. Missed Call Statistics: origin of the 62% figure (411 Locals study, 85 businesses, 58 industries)Secondary report
   411 Locals, as described in a Moz Q&A forum post and summarized by getaira.io. January 2016.
   We could not open 411 Locals' original write up. The split of 37.8%, 37.8% and 24.3% and the sample come from a search summary of secondary pages, which agree with each other and with the PCN synthesis. Several sites misdate the study to 2023 or 2024.
2. Findings of small business call report revealedCommissioned research
   Moneypenny. 21 November 2016.
   An independent survey of 300 micro businesses with 0 to 9 employees plus call data from 10,000 businesses, commissioned by a call answering provider. The page does not describe the questions or sampling, and its metadata mentions Britain while its text refers to US businesses.
3. Invoca Launches New Solution to Recover Revenue from Missed Sales and Appointment-Setting CallsCommissioned research
   Invoca, press release via SalesTechStar. 9 November 2021.
   Aggregate 2021 Invoca customer data. States 26% of calls unanswered and 2% of those callers leaving voicemail, with no sample size. Invoca sells call analytics.
4. Missed calls cost small businesses (summary of CallRail's 2025 consumer survey of 1,000 US consumers)Secondary report
   Beside, citing CallRail. 2025.
   We read only a search summary. Several sites repeat the 42%, 21% and 78% figures and attribute them to CallRail, but we did not find CallRail's own report. The figures are self reported.
5. Survey: Canadian businesses losing revenue due to unanswered callsCommissioned research
   Hiya, State of the Call 2026. 5 August 2026.
   More than 12,000 consumers and 3,600 business professionals across six countries. The article gives no Canada only sample size or weighting. Hiya sells caller identification services.
6. Answering the Call: Why Responsiveness Is Critical for Law Firm Success (summarizing Clio's 2024 Legal Trends Report)Secondary report
   Oklahoma Bar Journal, Vol. 96, No. 7, by Julie Bays. September 2025.
   We read this article, not Clio's report. It gives no secret shopper sample sizes. Clio's own press page returned an error when we tried to open it.
7. Peerlogic DSO case study: call analysis across 26 practicesCommissioned research
   Peerlogic. February 2026 data.
   Vendor analysis of 4,280 calls across 26 anonymous practices in one month. Definitions of answered and converted are not given, the revenue estimate has no stated method and there is no comparison group. The full case study is gated.
8. Receptionists, Occupational Outlook HandbookPrimary source
   US Bureau of Labor Statistics. May 2025 pay data, 2025 to 2035 projections; read 9 October 2026.
   The handbook is updated yearly. Earlier search results quoted 2024 figures of $37,230 and about 1,007,200 jobs, which this page replaces.
9. Small admin tasks take 36% of a 45.5 hour week (Time etc survey by Censuswide, 251 US entrepreneurs)Secondary report
   Time etc, as reported by smallbusiness.co.uk. Fieldwork September 2023.
   We read search summaries of write ups and could not find the original release. Commissioned by a virtual assistant company. Some sources give October rather than September for fieldwork.
10. How microbusiness owners spend their timeSecondary report
    Ueni. 2024.
    Survey of 837 US microbusinesses reporting admin plus legal and compliance at 22.4% of owner time. We read only a search summary and did not open the page, and Ueni sells website services.
11. The Short Life of Online Sales Leads (Oldroyd, McElheran, Elkington)Primary source
    Harvard Business Review, via BYU ScholarsArchive. March 2011.
    We read the repository record and abstract, which states that most companies are not responding nearly fast enough and contains no figures. The article is paywalled.
12. 2026 Small Business Missed Call Revenue StudyCommissioned research
    PCN Answers (Professional Communications Network). 2026, no publication date given.
    A synthesis of secondary data and scenario modeling with no sample of its own, published by a company that sells answering services. Its revenue figures are modeled, not measured, and are not used on this page.
13. The 42-hour problemSecondary report
    CIENCE. Undated.
    Source of the 2,241 company figures. It cites the HBR article but does not quote it, and we could not check the numbers against the original.
14. Generative AI at Work (Brynjolfsson, Li, Raymond), NBER Working Paper 31161Primary source
    National Bureau of Economic Research. April 2023, revised November 2023.
    We read the paper's abstract page. 5,179 customer support agents; a working paper rather than a final journal article, and about text based support rather than voice.

Questions

Related