White Label WhatsApp AI Agents: What the Research Says

How large is WhatsApp as a business channel?

WhatsApp passed 3 billion monthly users in 2025. Meta chief executive Mark Zuckerberg said so on the company's first quarter earnings call, as reported by TechCrunch ¹. For context, a global audit of digital habits found that 97.3% of connected adults use at least one social media or messaging service each month, and the average internet user spends 2 hours 21 minutes a day on social media ¹².

Business volume is harder to pin down. Meta first disclosed it on its third quarter 2023 call: people and businesses interact more than 600 million times a day across Meta's platforms ³. Meta did not say how that splits between WhatsApp, Messenger and Instagram, or how much of it is marketing rather than support, so the figure shows the size of business messaging, not the share that sits on WhatsApp ³.

What consumers say they want from businesses

Meta commissioned the research firm Kantar to survey 11,056 online adults aged 18 to 64 in 22 markets. Fieldwork ran in April 2025 for 18 markets and September 2025 for 4, and results were weighted to the global internet population ². Because Meta sells business messaging, this is industry research and should be read that way.

Stated preference is not behavior. The survey records what people say, not that adding WhatsApp raises sales for a particular business. One result has a practical edge: 79.3% of consumers say they need proof that a business is legitimate before they engage with it, so a verified business profile matters before the first message is sent ².

How fast people expect an answer, and how fast businesses respond

Sprout Social's 2025 Index found that about three quarters of consumers expect a brand to respond within 24 hours or sooner, and that 73% of social media users say they would buy from a competitor if a brand did not respond to them on social ⁷. These figures describe social channels in general, not WhatsApp, and Sprout's summary page does not state the sample sizes ⁷.

What businesses deliver is slower. In a 2011 Harvard Business Review study, researchers sent a test web enquiry to 2,241 US companies. According to a secondary summary of the study, 37% responded within an hour, 16% within 1 to 24 hours, 24% took longer than a day and 23% never responded. The average response time among companies that replied within 30 days was 42 hours ^4,5.

The same summary reports a companion analysis of 1.25 million leads across 42 companies. In it, firms that tried to contact a lead within an hour were nearly 7 times as likely to have a meaningful conversation with a decision maker as firms that tried an hour later, and more than 60 times as likely as firms that waited 24 hours or longer ⁵. The original article is behind a paywall, so these numbers are reported, not independently checked ^4,5.

The study is 15 years old and concerns web forms, not chat apps. It is included because it is a widely cited measurement of how quickly the value of an enquiry decays, and because the underlying problem, a customer who is ready to talk and a business that is not available, is the one messaging channels are meant to address.

What answering costs in staff time

ServiceNow and ThoughtLab surveyed 34,000 executives, service representatives and consumers worldwide in September and October 2025, including 2,465 in the UK. In the UK data, service agents spent 49% of their working week on customer issues, with the rest going to administration and switching between systems. UK consumers spent more than 445 million hours a year waiting on hold ⁹.

Neither figure is specific to WhatsApp, and we found no reliable published measure of how many hours businesses spend answering WhatsApp messages. The honest way to size it is with a business's own volumes.

How WhatsApp's own rules shape the cost of replying

Since 1 July 2025 Meta charges per message instead of per 24-hour conversation, and charges only when a template message is delivered ⁸. When a customer messages a business, a 24-hour customer service window opens. Inside it, non-template messages such as plain text or images are free, and so are utility templates. Marketing templates are always charged. Outside the window, only template messages can be sent ⁸.

If a customer arrives through a Click to WhatsApp ad or a Facebook Page button, a business that replies within 24 hours opens a 72-hour free entry point window in which any message type is free ⁸. Rates depend on template category and the recipient's country and change only on 1 January, 1 April, 1 July and 1 October, so this page does not quote prices. Meta's rate card is the authority ⁸.

Our reading of these rules, not Meta's wording: replies sent inside the 24-hour window carry no per-message fee, so the pricing structure rewards answering quickly.

What AI changes: the evidence

The strongest independent evidence is a study of 5,179 customer support agents published by the National Bureau of Economic Research. Agents who used 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 authors also report better customer sentiment and higher employee retention ⁶. The assistant supported human agents; it did not replace them.

Consumers are open to it. In the Kantar survey for Meta, 67.7% agreed that getting a response from an AI chatbot is helpful and 42.9% believed AI would improve their messaging experience with businesses ².

Gartner predicts that by 2029 agentic AI will autonomously resolve 80% of common customer service issues without human intervention ¹⁰. That is a forecast, not a measurement.

A Forrester Total Economic Impact study commissioned by Meta and announced by the messaging vendor Gupshup in April 2025 modeled a 270% return on investment over three years for a composite organization using Gupshup-powered automation on the WhatsApp Business Platform. It rests on seven customer interviews and a survey of 224 decision makers, and the vendor's release states some revenue figures inconsistently, so treat it as indicative ¹¹.

What the evidence does not show

• We found no independent, controlled study that measures how much an AI agent on WhatsApp raises conversion or lowers cost for a typical small business. The strongest AI evidence is about human agents using an assistant ⁶.
• Most business messaging research is commissioned by Meta or by vendors that sell on its platform ^2,11.
• The lead response study concerns web forms in 2011 and reaches us through a secondary summary, and its original article is paywalled ^4,5.
• Sprout Social's summary page does not state sample sizes for the figures cited here ⁷.
• We found no reliable published figure for hours spent answering WhatsApp messages, so the staff time section relies on general customer service data ⁹.
• Gartner's 80% figure is a prediction about 2029, not an observed result ¹⁰.

Applying this as an agency

The evidence points to three checks before selling WhatsApp automation to a client. First, how quickly does the client answer today? The response speed evidence suggests most businesses are slow. Second, how many messages arrive each day? The staff time cost scales directly with volume. Third, is the client's number tied to a verified business profile? 79.3% of consumers want proof of legitimacy before engaging ².

In a white label setup each client connects its own WhatsApp number to an agent that runs under the agency's brand and domain, and the client authorizes the connection through Meta. Replies inside the 24-hour customer service window carry no per-message fee ⁸. 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 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. Widely repeated numbers without a traceable origin, such as message open rates, are not used.
• Convocore sells white label AI agents and has a commercial interest in messaging. 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. WhatsApp now has more than 3 billion users a monthSecondary report
   TechCrunch. 1 May 2025.
   Reports Meta CEO Mark Zuckerberg's statement on the Q1 2025 earnings call. We did not read the call transcript.
2. State of Business Messaging, consumer report (Kantar Business Messaging Usage Research)Commissioned research
   Meta, research by Kantar. Fieldwork April and September 2025.
   Commissioned by Meta. 11,056 online adults aged 18 to 64 in 22 markets, weighted to the global internet population. We read the public summary page; the full report is gated.
3. Meta says users and businesses have 600 million chats on its platforms every daySecondary report
   TechCrunch. 26 October 2023.
   Reports the Q3 2023 earnings call. TechCrunch notes Meta did not name the apps or split marketing from informational messages.
4. 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 itself is paywalled.
5. The 42-hour problemSecondary report
   CIENCE. Undated.
   Source of the 2,241-company figures and the companion 1.25 million lead analysis. It cites the HBR article but does not quote it, and we could not check the numbers against the original.
6. 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; working paper rather than a final journal article.
7. Social media customer service statisticsPrimary source
   Sprout Social. 2025 data (Sprout Social Index 2025).
   Sprout's own summary of its Index. The page does not state sample sizes.
8. WhatsApp Business Platform pricingPrimary source
   Meta for Developers. Current documentation, read 9 October 2026.
   Rules and rate cards change on fixed quarterly dates. Check the live page before relying on pricing.
9. UK consumers lose 445 million hours a year on hold, ServiceNow research finds (report: The CX Shift: Customer Expectations in the AI Era)Secondary report
   Compare the Cloud, reporting ServiceNow and ThoughtLab research. Published 2 April 2026; research conducted September and October 2025.
   34,000 executives, service representatives and consumers worldwide, 2,465 in the UK. We read the news report, not the ServiceNow report. ServiceNow sells customer service software.
10. Gartner Predicts Agentic AI Will Autonomously Resolve 80% of Common Customer Service Issues Without Human Intervention by 2029Forecast
    Gartner. 5 March 2025.
    Gartner blocked our automated fetch, so we cite only the claim stated in the release's headline.
11. Gupshup Conversation Cloud and WhatsApp drove 270% ROI for customers, according to Total Economic Impact findingsCommissioned research
    Gupshup, citing a Forrester Consulting study commissioned by Meta. April 2025.
    Vendor press release. Seven interviews and 224 surveyed decision makers. We did not open the Forrester report, and the release gives two different total revenue figures, so we cite only the ROI and sample.
12. Digital 2025: Global Overview ReportPrimary source
    DataReportal (Kepios), compiling GWI, Kepios and data.ai. February 2025.
    We cite only figures we read on the page. Time spent is GWI survey data for the typical internet user.

Questions

Related