There’s a conversation your customer is already having – and it’s probably not with you. They’re asking about a product via a chatbot, booking an appointment on WhatsApp or getting an answer to a query from an automated assistant at eleven o’clock at night. The AI-powered conversational marketing It consists, quite simply, of being there at that moment: using artificial intelligence within messaging apps to provide personalised customer service, make recommendations and sell products at any time of day. And by 2026, it will no longer be a promise for the future but will have become a channel that actually generates revenue.
At Vandelay, we see this every week with clients from a wide range of sectors: email still works, social media remains essential, but chat has emerged as the place where conversations are actually concluded. We’re going to explain what’s happening, what changed with Meta’s latest decision and, above all, how to make the most of it without falling into the typical pitfalls.
What is AI-powered conversational marketing?
Conversational marketing is any strategy that uses dialogue — whether in real time or asynchronously — as the primary means of engaging with the customer. When we add artificial intelligence to the mix, that dialogue no longer depends on there being a person available at the other end. A well-configured AI agent can understand the writer’s intent, check stock levels, recommend a product, book an appointment or escalate the matter to a human when things get complicated.
The difference compared to the rigid chatbots of a few years ago is huge. Before, you’d navigate through menus like “press 1 for sales, press 2 for support”. Now you type naturally and the system interprets what you need. It is this fluidity that has fuelled its widespread adoption, and which ties in directly with another trend we’ve already discussed on the blog: the AI-powered hyper-personalisation.
Why 2026 is the year it really takes off
The figures for AI-powered conversational marketing help to illustrate the scale of the phenomenon. WhatsApp has around 3,000 million monthly active users worldwide, making it one of the platforms with the widest reach on the planet. And messaging as a whole continues to grow: industry estimates suggest there will be around 4,600 million messaging app users by 2026. It is, quite simply, where people are.
The second point relates to engagement. According to data compiled by industry providers, messages sent via WhatsApp Business achieve open rates of around 90–98 %, well above those of email marketing, which typically stand at around 20 %. It is worth treating these figures as industry estimates — they vary depending on the study and the type of campaign — but the trend is clear: a chat message gets read, and it gets read quickly.
The third is the business aspect. The conversational commerce market was valued at around 26,000–30,000 million dollars in 2025–2026, according to consultancy firms such as Juniper Research, with forecasts of sustained growth in the coming years. The exact figures vary depending on the source, but they all agree on one thing: people are increasingly making purchases through conversation. In fact, several recent analyses indicate that shoppers who interact with an AI assistant complete their purchase significantly more often than those who do not.
Meta’s U-turn: WhatsApp no longer supports “general-purpose” chatbots”
Here is the news you really need to bear in mind. Meta has amended the terms and conditions of the WhatsApp Business Platform (the API) to ban the use of general-purpose AI chatbots —open-source ChatGPT-style assistants— within the platform. As reported by TechCrunch in October 2025, the rule comes into force on 15 January 2026 and applies immediately to new accounts registered on or after 15 October 2025.
Meta gives two reasons for this: the volume of messages from these general-purpose assistants was putting a strain on the infrastructure and, furthermore, it strayed from the business-to-customer messaging model for which the API was designed. To put it another way: WhatsApp wants to be a channel for engagement between brands and people, not a back door for deploying ChatGPT on a large scale.
What you can still do (and that’s the interesting bit)
The good news for any business is that the change does not affect the use that actually generates value. AI assistants designed for specific business tasks are still permitted and, in fact, Meta encourages their use. These include:
- Customer service and answers to frequently asked questions.
- Order tracking and enquiries about deliveries.
- Booking and managing appointments or services.
- Product recommendations from your catalogue.
- Recovery of abandoned shopping trolleys and transactional notifications.
The strategic takeaway is clear: the future does not lie in simply plugging a generic model into WhatsApp, but in designing specialised agents that are connected to your inventory, your CRM and your processes. It’s more work, yes, but it’s also far more cost-effective and resilient in the face of policy changes such as this.
RCS, the other front in messaging
Whilst everyone’s focused on WhatsApp, the RCS (Rich Communication Services) standard – the successor to SMS, which is already built into many mobile phones – is gaining ground for the bulk sending of notifications and transactional messages. Various industry analyses position it as a highly effective channel for broad reach — with high open rates — whilst WhatsApp remains the go-to platform for closing sales and providing more complex support. It is not a case of “one or the other”: the sensible approach is to understand what each channel does best and combine them according to the objective of each campaign.
How we approach this in a real-world strategy
It’s one thing to talk about the theory, but quite another to put it into practice without ending up with a robot that puts customers off. These are the steps we follow when setting up an AI-powered conversational marketing project:
- Start with a real question. What do your customers ask you most often? That’s where the first use case lies – not in whatever sounds the most sophisticated.
- Connect the agent to reliable data. Catalogue, stock availability, opening hours, order status. A sales assistant without reliable data can only wing it, and in sales, that comes at a high price.
- Define when a human comes into the picture. AI handles routine tasks — it is estimated that it can manage a large proportion of common enquiries without human intervention — but the handover to a human must be seamless and swift.
- Be mindful of your tone. The chatbot speaks on behalf of the brand. If the brand is approachable, the chat is approachable. This ties in with something we always advocate: the authentic marketing with AI.
- Measure and correct. Conversations initiated and resolved without human intervention, appointments or sales generated. Without measurement, this is just a gimmick; with it, it’s a channel.
Risks that should not be ignored
Not everything about AI-powered conversational marketing is cause for enthusiasm. The first risk is legal: managing conversations involves processing personal data, so any implementation must comply with the GDPR, seek explicit consent and retain only what is necessary. The second risk relates to user experience: poorly calibrated automation that is pushy, repetitive or fails to understand generates more rejection than a voicemail. And the third is the risk of dependency: building your entire customer relationship on a third-party API leaves you exposed to changes in the rules, such as the one we’ve just seen. That’s why we stress the importance of also having your own channels — your website, your mailing list — that don’t depend on a platform’s current policy.
Where to start this week
If you take just one idea away from this, let it be this: chat is already a sales channel, and artificial intelligence is what makes it scalable. You don’t need a massive transformation to get started. You can begin with a bot that answers the five most frequently asked questions, linked to your real-time data, with a well-thought-out handover to a human agent and a few basic metrics to see if it’s working. From there, you can iterate.
At Vandelay, we’ve been helping brands grow for over a decade by combining strategy, creativity and technology, and AI-powered conversational marketing sits right at the heart of that. If you’d like to explore how to apply it to your business — in a measurable way, without any fluff — we can, of course, start that conversation whenever you like.
Sources
- TechCrunch — WhatsApp is amending its terms and conditions to ban general-purpose chatbots from its platform (18/10/2025).
- respond.io — Not All Chatbots Are Banned: WhatsApp’s 2026 AI Policy Explained.
- Juniper Research / Mordor Intelligence — market reports on conversational commerce y conversational AI (2026).
- Trengo / Botpress — WhatsApp Business and chatbot statistics for 2026.
English (UK) 
Español