This week Meta’s WhatsApp Business Platform team sent its business customers a briefing that opens with an uncomfortable line: most businesses can stand up an AI agent; fewer can say their customers and their own teams trust it enough to hand over the conversation.
We build AI chatbots for UK businesses, on websites and on WhatsApp, so we read it closely. It says three useful things: what earns trust (from a Meta product director), why most businesses lose it (from IDC research), and what a trusted agent delivers (from a case study with real numbers). Here is each one, translated for a business with a shop, a clinic, a van or a small team.
Meta’s Director of Product Management, James Detmer, says trust is not a switch you flip. An agent earns it by knowing when to hand off to a human, and by showing up with your brand voice, product knowledge and tone.
The hand-over is where cheap chatbots fall down. Either they never let go (the customer types “can I speak to someone” three times and gets three cheerful non-answers) or they hand over everything, in which case why have one. A chatbot people trust has plain rules:
Test it: before it goes live, be an awkward customer. Ask for a human. Ask for a price it has not been given. Complain. Watch what it does.
The second half of Detmer’s point is brand voice and product knowledge. Customers trust an assistant that answers “do you cover Stockport?” with a yes and a price, in the way you would say it. They do not trust one that produces a paragraph of generic text and a link to a contact form.
That only happens when the chatbot is built from your information: your services and real prices, your opening hours and areas, your policies, and the way you talk. Every “the bot told a customer we do X and we don’t” story we have heard comes from skipping this step. It is the single biggest difference between a chatbot that is set up properly and one bolted on in an afternoon.
The research in Meta’s briefing is IDC’s August 2026 report on customer service technology. The headline: only 49% of organisations fully preserve customer context when a conversation crosses channels. In plain English, half the time a customer moves from your website chat to a phone call to a WhatsApp message, the business has forgotten who they are and what they asked.
IDC tells big companies to treat customer context as “shared infrastructure” across service, marketing and sales. For a small business the version is simpler: one assistant, one memory. Someone who asked for a quote on Tuesday and messages again on Friday should be greeted by a chatbot that already knows about Tuesday. That is the difference between “this business is on it” and “this business is a mess”, and it is why we run a client’s website chat and WhatsApp from the same assistant rather than two separate bots.
The case study is Stellantis, the group behind Jeep. Its Meta partner, Blip, built a Meta Business Agent on WhatsApp to answer questions about the Jeep Renegade, qualify buyers and pass the serious ones to a salesperson. It was loaded with the official product knowledge base and monitored in real time. The published results, over 15 days:
Figures as reported by Meta and Blip in their published success story.
Ignore the car-industry scale; the pattern is what matters. The agent did not replace the sales team. It handled the tyre-kicking conversations, kept its facts straight and nearly doubled the share of leads worth a human’s time. And customers stayed for six minutes, which nobody does with a bot they distrust.
For a dentist, a garage or a lettings agent the equivalent is a chatbot that answers the twenty questions you get every day, collects what it needs (what, where, when, a photo) and passes you the ones ready to book.
Put together, Meta’s briefing gives you a ten-minute test for any AI chatbot, including one you already have on your site:
Pass all five and you can let it talk to customers. Fail two or more and it is a chatbot in the bad sense, and your customers already know.
Every Infobot ships with the five points in place: a human hand-over path (a call-back request reaches you as a message with the conversation summary), a “never invent” rule with a message-taking fallback, your own knowledge base so it answers in your voice with your prices, one memory across website chat and WhatsApp, and a stop rule. It runs on Meta’s official WhatsApp Business Platform, where we are a verified Tech Provider.
Run the test on ours. Message the Infobot on WhatsApp, ask for a person, ask it something it cannot know, say stop, and see what happens. Try it on WhatsApp or book a 15-minute demo.