AI in Marketing Is Moving Fast. Here Is What Matters

AI in marketing is no longer a future-state conversation. The tools are here, the adoption is accelerating, and the gap between teams using AI well and teams using it badly is already visible in commercial results. The question worth asking is not whether AI will reshape marketing, but which parts of the reshape will matter to your business and which are noise dressed up as progress.

Most of the forward-looking coverage on AI in marketing focuses on the technology itself: what models can do, what platforms are launching, what capabilities are coming. That framing is useful for vendors. It is less useful for marketers who need to make decisions about where to invest attention, budget, and organisational energy over the next two to three years.

I have spent more than 20 years in marketing and agency leadership, managing hundreds of millions in ad spend across 30 industries and running agencies through periods of significant technological change. I watched the paid search revolution in real time. I was at lastminute.com when we launched a campaign for a music festival and generated six figures of revenue in roughly a day from what was, by today’s standards, a relatively simple setup. The technology was new, the results were immediate, and the instinct was to assume the tool itself was the story. It was not. The story was understanding what the customer wanted and getting in front of them at the right moment. AI is not different in that respect.

What Does the Near-Term Future of AI in Marketing Actually Look Like?

Strip away the vendor roadmaps and the conference keynotes and you are left with a handful of shifts that are genuinely going to change how marketing functions operate. Not all of them are glamorous. Most of them are operational.

The first is content production at scale. This is already happening and the trajectory is clear. Moz’s research on AI content points to a market where AI-assisted content is becoming the norm rather than the exception. The practical consequence is that the cost of producing average content is collapsing toward zero. That sounds like good news until you realise it means the bar for what constitutes useful, differentiated content is rising just as fast. The teams that will win are the ones who use AI to handle volume and use human editorial judgment to handle quality control and strategic direction.

The second shift is in personalisation. Not the superficial kind where you insert a first name into an email subject line, but genuine behavioural personalisation at a scale that was previously only available to organisations with very large engineering teams. Tools that can adapt messaging, sequencing, and offer logic in real time based on individual behaviour patterns are becoming accessible to mid-market businesses. Semrush’s overview of AI marketing covers the mechanics of this well if you want a grounding in the technical side.

The third shift, and the one I think is most underappreciated, is in predictive analytics and decision support. Not AI making decisions autonomously, but AI surfacing patterns in data that a human analyst would take weeks to find, and doing it continuously. For anyone who has run a marketing function inside a large organisation, the bottleneck is rarely data collection. It is data interpretation and the speed at which insights translate into decisions. AI closes that gap significantly.

If you want a broader view of where AI sits across the marketing function today and where the category is heading, the AI Marketing hub at The Marketing Juice covers the landscape in depth, from tools and automation to strategy and measurement.

Will AI Replace Marketing Teams?

This question gets asked constantly and the honest answer is: some roles, partially, over time. That is not a dramatic claim. It is what happens with every significant productivity technology. The more useful question is which parts of marketing are most exposed and which are most defensible.

The roles most exposed to AI substitution are the ones built around execution of repeatable tasks: templated content production, basic reporting, campaign setup following established playbooks, and routine A/B testing. These are not low-skill activities in the traditional sense, but they are activities where AI can match or exceed human throughput at a fraction of the cost. Moz’s breakdown of AI tools for automation and productivity gives a practical view of where this is already happening.

The roles most defensible are the ones built around judgment, commercial context, and relationship. Understanding why a brand sits where it does in a market and what it would take to shift that position. Reading a client’s business well enough to know when the brief is wrong. Making the call on a campaign that the data does not fully support because you understand the audience in a way the model does not. These are not soft skills. They are hard-won commercial skills that take years to develop, and they are genuinely difficult to replicate with current AI architecture.

When I was growing an agency from 20 to 100 people, the thing that consistently separated strong performers from average ones was not technical ability. It was the capacity to hold a client’s commercial problem in mind and work backwards from it. AI is not close to replicating that. What it can do is free up the people who have that capacity to spend more time using it.

How Will AI Change the Relationship Between Strategy and Execution?

This is where the future gets genuinely interesting, and where I think most of the current conversation misses the point.

For most of marketing history, strategy and execution have been separated by cost and time. You could have a sophisticated strategy, but the execution required to test it, iterate on it, and refine it at scale was expensive and slow. That created a structural bias toward simpler strategies that were easier to execute consistently.

AI collapses that constraint. When execution becomes cheap and fast, the limiting factor becomes the quality of strategic thinking upstream. A mediocre strategy executed with AI efficiency is still a mediocre strategy, just delivered faster and at greater scale. The organisations that will compound the most value from AI are the ones that invest in strategic clarity at the same time as they invest in AI tooling.

I have seen this pattern play out in performance marketing over and over. Early in my career, the teams that got the most from paid search were not the ones with the most sophisticated bidding setups. They were the ones who understood their customer economics well enough to know what a conversion was actually worth. The same logic applies to AI. The tool is not the edge. The thinking behind the tool is the edge.

HubSpot’s coverage of AI marketing automation makes this point in a different way: the marketers getting the best results from automation are the ones who have invested in understanding their audience and their funnel before they automate anything.

What Is the Real Risk of Moving Too Fast on AI Adoption?

The pressure to adopt AI tools is real and it is coming from multiple directions: boards, clients, competitors, and the general noise of the industry. That pressure is not always aligned with commercial sense.

The most common mistake I see is organisations buying AI capability before they have solved the data infrastructure problem underneath it. AI tools in marketing are only as good as the data they run on. If your customer data is fragmented across platforms, your attribution model is unreliable, and your CRM has not been cleaned in three years, adding an AI layer on top of that does not fix the problem. It amplifies it.

Early in my career, I asked the MD of the company I worked for to approve budget for a new website. He said no. Rather than wait, I taught myself to code and built it myself. The lesson I took from that was not about resourcefulness, though that is part of it. It was about understanding the problem well enough to solve it with what you have. Too many organisations are doing the opposite with AI: buying the solution before they have properly defined the problem.

The second risk is in content quality and brand consistency. Crazy Egg’s analysis of AI marketing assets highlights a tension that is already visible in the market: as AI-generated content proliferates, the brands that maintain a distinctive voice and genuine editorial point of view stand out more, not less. The risk of over-automating content is not just quality degradation. It is brand homogenisation.

The third risk is over-reliance on AI outputs without sufficient human oversight. This is particularly acute in areas like audience targeting, where AI systems can encode and amplify biases present in historical data without anyone in the organisation noticing until the damage is done.

Which AI Capabilities Are Worth Prioritising Right Now?

If I were advising a marketing director on where to focus AI investment over the next 18 months, I would point to three areas that have a clear line to commercial outcomes.

First, AI-assisted content operations. Not AI writing everything from scratch, but AI handling research, first drafts, variations, and distribution optimisation while human editors focus on strategic direction, quality control, and distinctive voice. Buffer’s research on AI tools for content marketing agencies is a useful practical reference for how this works in practice across different team sizes.

Second, predictive lead scoring and customer lifetime value modelling. If you have reasonable CRM data and a meaningful transaction history, AI can surface patterns in customer behaviour that manual analysis would miss. The commercial payoff from better prioritisation of sales and marketing effort against high-value segments is significant and relatively straightforward to measure.

Third, AI-powered testing and optimisation at the campaign level. Not replacing human creative judgment, but accelerating the feedback loop between creative hypothesis and performance data. The teams I have seen do this well treat AI as a way to run more experiments faster, not as a way to avoid making creative decisions.

One area I would approach with more caution is AI-generated creative at the top of the funnel, particularly for brand-building work. The tools are improving rapidly and the trajectory of AI optimisation software suggests this will be a different conversation in three years. But right now, the creative outputs that build genuine brand equity tend to require a level of cultural and emotional intelligence that current generative models do not reliably produce.

How Should Marketing Leaders Think About AI Model Selection?

The proliferation of AI models and platforms is creating a new category of decision fatigue for marketing teams. GPT-4, Claude, Gemini, Llama, and a growing list of specialist marketing AI tools all have different strengths, pricing structures, and appropriate use cases.

The mistake is treating model selection as a one-time strategic decision. The landscape is moving fast enough that what is best for a specific use case today may not be best in six months. HubSpot’s comparison of LLMs for marketing use cases is a useful starting point, but the more important discipline is building the internal capability to evaluate and switch models as the market evolves, rather than becoming locked into a single platform.

The evaluation criteria that matter most for marketing applications are: output quality for your specific content types, integration capability with your existing stack, data privacy and compliance posture, and total cost at your usage volume. The last one is consistently underestimated. Many organisations adopt AI tools based on per-seat licensing costs and then discover that actual usage at scale is significantly more expensive than the initial projections.

Having judged the Effie Awards and seen the work behind some of the most effective campaigns in the market, one pattern is consistent: the campaigns that perform best are built on a precise understanding of the audience and a clear commercial objective. AI can help you execute against those things faster and at greater scale. It cannot substitute for having them in the first place.

What Does Good AI Integration Look Like in a Marketing Function?

The organisations getting the most from AI in marketing right now share a few characteristics that are worth noting.

They have defined specific commercial problems they are trying to solve before selecting tools. They are not adopting AI because it is expected. They are adopting it because they have identified a concrete inefficiency or opportunity that AI can address, and they have a way to measure whether it is working.

They have invested in data infrastructure alongside AI tooling. The two are not separable. An AI system running on clean, well-structured, properly attributed data will consistently outperform a more sophisticated AI system running on messy data.

They have maintained human oversight at the points where AI outputs have the highest commercial consequence. Automated bidding, personalisation logic, and content generation all benefit from periodic human review, not because AI is unreliable, but because the commercial context changes in ways that models do not always pick up quickly.

And they have been honest about what AI is not yet good at. The organisations that have burned budget on AI initiatives tend to be the ones that over-estimated current capability in areas like brand strategy, creative direction, and complex client relationship management.

The marketing teams building durable competitive advantage with AI are not the ones moving fastest. They are the ones moving most deliberately, with a clear view of where AI creates genuine commercial value in their specific context and where it is still a solution in search of a problem.

For a broader perspective on how AI is reshaping marketing strategy, measurement, and team structure, the AI Marketing section of The Marketing Juice covers the full range of topics, from practical tool selection to longer-term strategic considerations.

About the Author

Keith Lacy is a marketing strategist and former agency CEO with 20+ years of experience across agency leadership, performance marketing, and commercial strategy. He writes The Marketing Juice to cut through the noise and share what works.

Frequently Asked Questions