The Future of AI Marketing: My 3-Year Prediction

The Airport Hotel in Zurich That Changed How I Think About AI

I was sitting in a hotel bar near Zurich Airport last month, jet-lagged, waiting for a connecting flight. The bartender — mid-twenties, sharp — asked what I do. I said AI marketing. He nodded slowly and said, "So you're one of the guys putting people out of jobs."

I didn't get defensive. I ordered another coffee and thought about it seriously for the first time in a while.

Because here's the thing: he's partially right. Some jobs in marketing are already gone. Some agencies are already dead, they just haven't stopped billing yet. And what's coming in the next three years is going to be far more disruptive than anything most marketers are prepared for.

This is my honest read on where AI marketing is going — based on managing over $100 million in ad spend, building Lojain AI, and watching what actually works inside real businesses at Kira Agency.

What's Going to Die (And Deserves To)

Generic content production is already a zombie business. Agencies charging $2,000 a month to write five blog posts and schedule Instagram content? That model is functionally dead. The clients just haven't all figured it out yet. When a founder can generate a month of first-draft content in forty minutes using the right AI stack, you cannot justify that retainer anymore. I've seen agencies try to survive by adding "AI-powered" to their deck. It doesn't work. Renaming the same bad service doesn't change what it is.

Media buying as a standalone skill is also quietly disappearing. Not the strategic thinking behind it — that still matters — but the execution layer. Meta's Advantage+ campaigns, Google's Performance Max, TikTok's Smart Performance Campaigns: the platforms are eating the implementation work themselves. I run a campaign management team. In 2021, we had people whose entire job was manual bid adjustments. That role doesn't really exist the same way anymore.

The agencies that will die in the next three years are the ones selling execution without intelligence. The ones who got good at running ads in 2019 and never upgraded their thinking. There are a lot of them.

What's Going to 10x (And Is Already Starting)

Conversational AI in sales is the biggest shift I'm watching. Not chatbots — I want to be clear about that distinction. Chatbots are scripted decision trees that frustrate people. What I mean is genuine AI sales agents that can hold context, handle objections, qualify leads, and move someone from curiosity to commitment inside a single WhatsApp thread.

We built Lojain AI specifically because I couldn't find anything that worked properly in Gulf Arabic. The dialect differences, the cultural expectations around how a sale is made — no generic platform handled it. Lojain is now running conversations for businesses in real estate, F&B, healthcare, and retail across the GCC. The results are not marginal improvements. A real estate developer in Dubai came to us last year having tried three other automation solutions. Within sixty days of deploying Lojain AI, their cost per qualified lead dropped by 58% and their WhatsApp-to-site visit conversion rate tripled. That's not a tweak. That's a structural change in how their sales pipeline works.

Hyper-personalization at scale is the second thing that's going to 10x. Right now, most brands are doing segmentation — they're putting people in buckets and sending slightly different messages to each bucket. That's not personalization, that's sorting. What's coming is true 1:1 dynamic content driven by behavioral signals, purchase history, conversation patterns, and intent data — all assembled and delivered in real time without a human touching it. The brands that figure this out first will have a customer experience advantage that's almost impossible to close.

And the third: AI-native performance creative. Not AI-assisted — AI-native. The entire creative testing process is being rebuilt. Instead of a creative team producing twelve ad variants and running a two-week A/B test, we're moving toward systems that generate, deploy, and iterate hundreds of variants continuously based on live performance signals. I'm already running early versions of this for clients at Kira Agency. The speed advantage alone is enormous.

The Mistake I Made That Cost Me a Year

I went all-in on AI content tools in 2022 thinking they would transform client acquisition for a wellness brand I was working with — a gym chain in Manchester. I automated their email flows, their social content, their ad copy, everything. The output looked fine. Technically correct. Brand guidelines followed.

But conversions dropped. Not dramatically. Just... steadily, for months. It took me embarrassingly long to diagnose the actual problem: the content had no tension in it. No point of view. It was smooth and bland and forgettable. The AI was optimizing for coherence, not for the friction that actually makes people pay attention.

The lesson I took from that: AI is an extraordinary execution engine and a terrible strategist. You still need a human being — ideally one who understands your market deeply — setting the strategic intent, defining the voice, deciding what position you're taking. The moment you hand that to the AI as well, you lose the thing that made your brand worth paying attention to. I've been much more disciplined about this separation since then. Strategy is human. Execution can be AI.

My Controversial Opinion: Most Brands Are Automating the Wrong Things First

Everyone wants to automate content. I understand the appeal — it's visible, it's easy to show the output, it feels like progress.

But the highest-leverage automation in 2025 is customer conversation. Not content creation. The moment someone raises their hand — clicks an ad, sends a DM, fills out a form — is the single most valuable moment in the entire customer journey. And most brands are still handling that moment with a human who responds four hours later, or a rigid chatbot that sends a PDF and a phone number.

We're leaving an insane amount of money in that gap. A salon owner in Salmiya started using WhatsApp API automation through our platform last year. Before: average response time was three hours, booking conversion was around 12%. After three months: response time under ninety seconds, booking conversion at 34%. Same ad budget. Same services. Just a faster, smarter first conversation.

The conversation is the conversion. If you're spending your automation budget on content scheduling before you've solved your lead response problem, you have your priorities backwards.

Where I'm Putting My Next Three Years

At Kira Agency, we're building everything around what I'm calling the AI sales layer — the infrastructure that sits between a brand's marketing spend and their actual revenue. That means WhatsApp API, Lojain AI, performance creative systems, and the analytical layer that connects all of it into a picture of what's actually working.

The Zurich bartender's question stayed with me. Am I putting people out of jobs? Some jobs, yes. The ones that were already obsolete. But I'm also watching AI create entirely new roles: AI conversation designers, prompt strategists, automation architects. The businesses that will win in the next three years aren't the ones that resist this shift or the ones that blindly automate everything. They're the ones that figure out where human intelligence and AI execution make each other sharper.

That's what I'm trying to build at kiraco.org. And after ten years of doing this — from managing hundred-million-dollar ad accounts to building the first Gulf-Arabic AI sales agent — I'm more convinced of that direction than anything else I've ever bet on.

If you want to talk through what this looks like for your business specifically, I'm reachable.

Talk to Us on WhatsApp