The Future of Marketing in the Age of AI.

Every time someone asks me how I see the future of marketing in the age of AI, I remember a quote by Picasso:

“There is no such thing as art. There are only artists: people with the gift of combining shapes and colors.”

I believe the same is true for marketing. The discipline will survive the market’s attempts to devalue it, even though many of its functions probably won’t.

We’ve spent years trying to “kill” marketing. And, as a disclaimer, I don’t believe ChatGPT or AI will kill Google, nor that AI agents (like Manus or Gemini) will replace agencies or consultancies. Just look at Google’s financial results after integrating AI in search — it’s clearly not that simple.

My reflection focuses on how marketing work changes depending on the type of company and the current context. In short: marketing leaders will need to understand more about technology, psychology, and —above all— interpersonal communication.

Over the last 20 years, online marketing has evolved along three major vectors: platforms, automation, and personalization, with every advance sparking predictions about the death of the others.

As a disclaimer: if you’re only looking for the “top ten” without the context, you can jump here.

Platform evolution

For those of us working in media or over 40, the big turning point was the rise of blogs. They democratized access to communication: anyone could share opinions, build an audience, and reach a global public previously reserved for media outlets and large brands.

Ken Doctor, in Newsonomics, reflects on how this change disrupted the media system’s balance. Traditional media lost their monopoly on creating, managing, and steering public conversation.

Suddenly, anyone could talk about our products or competitors (for better or worse), and monitoring, response times, and coordination became key assets for marketing teams.

The rise of new platforms brought early models of social search and a pre-influencer framework, shifting authority toward individual voices.

Every new channel arrived with an obituary for the previous one: Google didn’t die at the hands of Del.icio.us, nor did Digg or Menéame kill SEO. Facebook, Twitter, Instagram, Snapchat, Pinterest, and Flickr didn’t wipe each other out either. They all found their place and value.

The deepest change for marketing hasn’t been budgetary, but conceptual: learning to manage channel complexity, coexistence, measurement, and attribution.

The new segmentation

Segmentation — now a given — has changed dramatically. In 2005, you could buy clicks on Google for €0.05 and flood anyone searching for “Madrid” with ads, in a pre-GDPR era of Third Party Data.

Today, segmenting means actually knowing your customer (First Party Data) and reaching them at the right moment. It also requires dynamic optimization for your archetypal customer or persona.

Marketing can no longer profitably target “45-year-old dads with a first child” — now it demands collaboration between Data, MarTech, UX, and Acquisition teams to design events, audiences, and messages that algorithms can amplify.

Value-based segmentation: a real example

An app that helps users file their taxes wanted to reach those likely to have deductions.

Since the market didn’t allow segmentation by personal variables (“just had a child,” “bought a house”), they used their CDP to identify common behaviors among users with the best results.

They created a custom event and optimized the campaign around that goal — showing creatives to those most likely to complete the action.

The challenge is technological, yes, but also deeply human: learning to trust that an algorithm might outperform us — and accepting that this doesn’t make us obsolete, but rather more necessary on different levels.

Automation: the silent vector

Automation is perhaps the change that has met the least resistance. Though often framed as a job killer, adoption has been universal.

The growth of digital disciplines (social, SEO, app campaigns) has made us rely on automated processes to manage an increasingly fragmented environment with the same human resources.

Here, company size and type make a difference: not all processes should be automated, but they should be understood. Because automation doesn’t mean disengagement — it means making room for better thinking.

Automation in consulting / SMEs

In the coniadigital conference, Iñaki Huerta explains the risks of automating certain business processes. Value generation and business knowledge can’t be delegated, and it’s essential to structure the transfer of knowledge properly.

General recommendations include:

• Identify repetitive processes that hinder scalability, waste time, and offer low execution value — and automate them.
• Accelerate data extraction and processing, pre-audits, clustering, error analysis, and alert systems to maintain business focus with a personal touch.
• Tools like N8N, Make, or MCP integrations are essential for marketing professionals.

In mid-sized businesses, the need is different

Value generation depends on more departments and their synchronization. Marketing must coordinate across tools like Braze, Klaviyo, Hubspot Marketing Hub — where understanding cost beyond price, opportunity cost, and future limitations is key.

These platforms provide only a partial view of the customer. Each has strengths and limitations, including their own black-box models and scoring systems. Still, what’s lost in data visibility is gained in control and agility.

At the corporate level, the challenge is scale

Multiple departments, agencies, technologies, and dependencies mean AI use must rely on automated processes, not individual instances. The output of one process must become the input of another.

Managing complexity is the main challenge, and some paths worth exploring include:

• Maintaining a 360° customer view in a packaged platform or proprietary environment.
• In proprietary setups, building a CDP and leveraging Reverse ETL tools (Hightouch, Census).
• When building a CDP, ensure it supports model execution or propensity list loading.
• Deciding whether or not to automate creative generation, and structuring campaign naming to understand emotional/rational value axes.

Ultimately, it’s essential to establish a feedback loop for the funnel and empower different experts to add specific value at each micro-moment.

Here, the main tool is a Miro (or Mural) and the key intrapersonal skill is communication and the strategic value of the marketing director — not in process control, but scenario design.

My top recommendations for marketers in 2026

1. More and more processes will become semi-automated, so building a shared vision, guidelines, tone of voice, and data quality will become critical. Invest in crafting your vision now.
2. The "segment of one" or persona definition may have a name and face, but if you can’t capture it in an analytics event, customer acquisition will cost you more. You can automate campaigns, emails, and creatives — but if platforms don’t understand your customer types, you’re limiting yourself.
3. Every marketing platform will introduce generative AI for marketer interactions. The challenge is making them all understand your goals.
4. The pace of change will accelerate, demanding communication, delegation, and goal-oriented management.
5. We’re entering a world of exponentiality with direct impact, digitizing processes, democratizing access, and demonetizing functions.
6. Automating insights with MCP or generative AI is useful only when the information is for personal use. To scale, make sure your insights are structured and normalized in a recognized report.
7. Perfect is the enemy of good. Any tool offering competitive advantage deserves exploration. Focus on implementation time and exit barriers.