How Marketing teams get real value from AI, and it’s not what you think

Marketing teams get real value from AI when they put craft and context before tools. AI can speed up research, drafting, analysis and execution, but it only improves results when the team knows what good looks like, can draw on reliable organisational knowledge, and understands the customer and market it is working in. The strongest teams don’t start by buying more tools. They start by clarifying the work that matters, building the knowledge base AI needs, and only then choosing tools that strengthen execution.

I’ve spent the past few months inside marketing teams. The ones getting the most from AI aren’t those with the biggest tool stack, but those with the clearest priorities, the strongest customer insight, and the most disciplined way of working.

Key takeaways

• Craft first. AI rewards marketers who already know what good looks like. It does not supply the judgment.
• Context second. Your advantage lives in a knowledge base built from organisational knowledge, domain expertise and customer intelligence.
• Tools last. Choose them in layers, and build your knowledge around the tools rather than inside any one of them.
• The order matters. Most teams start with tools. The leverage sits in craft and context.

Who this is for: founders, CEOs, marketing leaders and growth teams trying to turn AI experimentation into demonstratable commercial impact.

If your team is experimenting with AI but not seeing commercial impact, it usually pays to start with a growth diagnostic rather than another tool.

What actually creates value from AI in marketing teams?

AI creates value when skilled marketers use it inside a clear business context, with reliable organisational knowledge, domain expertise and customer intelligence behind it. The model is the engine. The judgment and the knowledge are what make the output yours.

Most AI initiatives underwhelm. Not because the tools are bad, but because the information supporting them is too often neglected. You can bolt new tools onto existing workflows and feel very busy, but real impact comes from going deeper, from rethinking how work gets done at an ecosystem level.

Getting the most out of AI comes down to three things, in this order of importance:

1. The fundamentals of the craft.
2. The foundational context the technology needs.
3. And only then, the tools.

Many teams start at number three. The leverage sits at one and two.

What telecoms adoption taught me about AI

I grew up professionally in telecommunications, working on how to drive adoption of new technology in the consumer market, over and over again, from 3G through to 5G and all the devices and services riding on that backbone. That front-row seat is why I read this wave the way I do.

Back then, we judged adoption through four lenses:

• Accessibility: can people easily get it?
• Utility: is it genuinely useful to most of the targeted segment?
• Literacy: do they have the fluency to use it?
• Affordability: do they have the means, and can they justify the cost?

Look at ChatGPT through that lens and it’s obvious why it became the fastest-growing consumer application in history. It ticked every box at once.

So when this wave broke at the end of 2022, I committed again to being a practitioner, to understand what this will mean for the way we actually work and live. The conclusion I keep returning to is that the answer does not lie in the technology.

Why craft comes before tools

Craft comes first because AI amplifies judgment rather than replacing it. Give two marketing teams the same models, the same budget and the same tools, and they will produce very different work. The difference is the quality of the thinking around the inputs.

AI amplifies judgment, it does not replace it

Every discipline has fundamentals, and marketing is no exception. The function has mushroomed into a dozen specialisms, but whether you’re building a brand or running email campaigns that actually convert, there are first principles underneath. When the market changes, most of those principles hold. The differences show up in the tactics.

It’s no accident that many of today’s best creators study the advertising and copywriting greats of the past. They mastered the craft of capturing attention. The era of blue links and clicks gave us new disciplines for capturing and measuring attention, which made some things easier, but also obscured the deeper work: understanding the customer and what it really takes to make a product or service resonate.

The marketing skills AI doesn’t replace

AI doesn’t replace that craft. It rewards the people who have it. Knowing what a strong insight, claim, message or offer looks like, and knowing which one breaks through the noise, is still a human call. The hard-earned stripes of doing the work the old-school way are an advantage now, not a relic.

The tasks, resources and information test

Before choosing tools, run a simple test. Every piece of work runs on three things: the tasks that need doing, the resources that do them, and the information that fuels them. AI has already collapsed the cost of the first two, which is exactly why the third is where your advantage now lives.

Which tasks can AI take on?

A few years ago, producing an ebook or white paper involved people, agencies, processes and specialised tools. Now an individual with subject-matter expertise can generate something of meaningful quality in an afternoon. Getting tasks done has been democratised substantially.

Which resources constrain execution?

The same applies to resources. Four years ago, building a website was a serious project for a non-technical person. Today you can describe what you want, point a tool at an example, and get a working mock-up in minutes. The people and tools that used to gate execution are far less of a constraint.

Which information is missing, messy or locked in people’s heads?

Which leaves information as the thing that actually changes and differentiates the outcome. Information is what makes your email, website or campaign different from a competitor’s. It’s the only one of the three that provides a foundation for competitive advantage over time. And yet in most organisations, the information needed to do the work is managed passively, never treated as a strategic asset.

My operating principle is Impact = Clarity multiplied by Focus. AI can increase speed, but speed without clarity multiplies noise. The teams that benefit most from AI are the ones that first clarify the growth constraint, focus the team’s energy, and then use AI to improve execution around that priority.

The three knowledge layers every marketing AI system needs

A marketing AI system needs three layers of knowledge to produce work that is genuinely yours: organisational knowledge, domain expertise, and customer and market intelligence. Built once and maintained thereafter, these layers fundamentally change the quality of output you get from both your people and your AI systems.

Organisational knowledge

The foundational facts about the business. Who you are. What you sell. How you sound. Brand voice, ICPs, personas, product detail, positioning, and the processes behind your outputs. In multi-brand environments this usually lives in pockets. The job is to align it and put it somewhere humans, assistants and agents can all use.

Domain expertise

The hard-won thinking that lives in the team’s heads and in the organisational memory. Frameworks, methodologies, campaign learnings, pattern recognition. Every campaign has a cause and effect: you execute, evaluate, recalibrate. If that learning stays in someone’s notebook or memory, every new campaign runs the risk of starting from scratch.

Customer and market intelligence

The continuously refreshed view of what’s happening outside your walls. Social listening, proprietary research, survey feedback, and the signals your salespeople pick up in the field. A new competitor mentioned in a sales conversation should trigger an analysis that flows back into the team and, if needed, into the positioning. Where these layers live, how they’re maintained and how they’re secured become important questions the moment you reconsider how work gets done across your team and the value chain.

From foundations to daily use: custom assistants and prompting

Once the knowledge base exists, two things turn it into something a team uses every day: custom assistants and the way you brief the work. This is where a small group can multiply its output without losing quality or control.

A while ago I worked with a small, funded startup in a highly regulated industry, preparing to scale into the direct-to-customer channel. The reality was a tiny team, an aggressive launch date, and a legally sensitive topic that couldn’t afford sloppy messaging.

Once we’d defined the growth strategy, one of the first execution tasks was to build the knowledge base. Not as a nice-to-have, but as the only way to make the launch feasible. Two commercial operators pulled together the core layers:

• Organisational knowledge: positioning, tone of voice, product detail, legal boundaries.
• Domain expertise: how the category works, what tends to land with this buyer, what’s off-limits.
• Customer and market intelligence: what we already knew about this audience and the channel dynamics.

With that in place, those two people could multiply themselves across the rest of the team through a set of assistants and automated flows. The same small group could brief campaigns across multiple internal and external channels, quality-check legally sensitive content, keep tone and claims consistent, and wire the same knowledge into internal comms and automated workflows. The interesting part wasn’t the tools. We used the same models everyone else has access to. It was how much leverage came from turning what lived in their heads and scattered documents into a usable, living asset.

That multiplication is the point, and two things make it possible day to day.

First, custom assistants: purpose-defined containers with one job, loaded with your knowledge, built once and used repeatedly. “The brief writer.” “The campaign analyser.” “The content repurposer.” You’re effectively building specialist teammates on top of your information.

Second, how you describe the work. Prompting is really just how you brief. Treat AI as a very clever intern and ask: what context would I have given a capable human to do this properly? The role, the audience, the outcome, the constraints. That question matters far more than any template.

How to choose AI tools without starting with tools

Choose AI tools in layers, and build your data infrastructure around the tools rather than inside any one of them. This is where much of the public conversation lives, and where it’s easiest to feel overwhelmed. Every week there’s a new “must-have” tool or skills library that promises to change how you work. A simple pyramid keeps it in focus.

The foundational tool

The one everyone on the team uses for drafting, reasoning, brainstorming and structured analysis. ChatGPT, Claude, Gemini or Copilot. Pick your house tool and standardise on it, because the compounding only happens when the team shares one foundation.

The customisation layer

Your knowledge bases, custom assistants, skills and workflows, sitting on top of that foundation. This is where you inject the three knowledge layers and codify how your team works.

The niche-tool layer

Specialist tools for specific jobs such as deep research, document interrogation, voice transcription and generation, video and image generation, and workflow automation. Add these only when they solve a real bottleneck.

When your knowledge lives outside the model, every tool can draw on the same context, and you’re far less exposed when the next must-have model arrives.

How to get started: a 30-day AI readiness sprint for marketing teams

Start with a problem, not a tool, then build the context around it, and choose tools last. Three steps make that concrete.

First, name the problem: something painful and repeatable that can genuinely be taken off your hands, such as campaign brief drafting, first-pass performance analysis, or turning long-form content into channel-specific assets.

Second, gather the information needed to solve it: what would you have handed that clever intern? Past examples, templates, brand voice, ICPs, success criteria, constraints.

Third, map how the work gets done across the team: look at the full value chain, not just one task. If you get fast at generating campaign briefs but there’s a bottleneck two steps later with your agency, you haven’t actually won.

The steps below turn that into a four-week sprint a marketing team can run without specialist help.

Before buying another AI tool, ask:

• What problem are we actually trying to solve, and is it worth solving?
• What would we have told a capable new hire to get this right?
• Where does the work really slow down across the value chain?
• What do we know about our customer that a competitor’s model does not?

A few guardrails are worth holding along the way. Keep a healthy respect for personal and sensitive information: if you wouldn’t email it to a stranger, don’t put it into a tool. Verify before you send, share or post. Treat your knowledge base as an asset and build it somewhere protected and accessible. And disclose AI use where it matters to customers and to internal and external stakeholders.

This is the kind of clarity a fractional Chief Growth Officer can help create before you invest in another tool. It is also the work behind M3: Mission Means Machine, the framework I use to connect strategy to the system that delivers it.

Frequently asked questions about AI for marketing teams

What is the best way for marketing teams to get value from AI?

Start with craft and context before tools. Clarify the growth priorities, strengthen marketing judgment, build a reliable knowledge base, and then choose tools that support specific workflows. AI is most useful when it helps a skilled team make better decisions and execute with more focus, not when it is bolted onto work nobody has thought through.

Should marketing teams start with AI tools or strategy?

Strategy first. Tools can accelerate work, but they cannot decide which customer matters, which message is strongest, or which growth constraint to fix first. A practical approach maps the team’s tasks, resources and information gaps, then selects tools that remove friction from high-value work.

What context does AI need to work well for marketing?

Three kinds. Organisational knowledge covers strategy, positioning, offers and tone of voice. Domain expertise gives AI your quality standards and hard-won methods. Customer and market intelligence keeps outputs grounded in real buying behaviour rather than generic marketing language. Together they turn a generic assistant into something that sounds and decides like your team.

Why does marketing craft still matter when teams use AI?

Because AI amplifies the quality of the input and the judgment around it. A marketer still needs to know what a strong insight, claim, message or customer narrative looks like. Without craft, AI produces more content faster, but that content may not create clarity, differentiation or commercial impact.

What is a marketing knowledge base for AI?

A structured set of source material that helps AI understand the business. It can include positioning, customer research, sales notes, proof points, competitive intelligence, product information, brand voice and strategic priorities. Built once and maintained, it turns AI from a generic assistant into a more useful growth execution partner.

How should a growth team choose AI tools?

In layers. First, a foundational tool for general research, synthesis and drafting. Second, a customisation layer of prompts, reusable workflows and knowledge bases. Third, niche tools, added only when they solve a specific bottleneck such as customer research, sales enablement, analytics or content production.

How can a Fractional CGO help a team use AI better?

A Fractional CGO connects AI adoption to growth outcomes. Instead of treating AI as a productivity experiment, the lens asks which revenue constraint matters most, what information is missing, which workflows need focus, and how to measure impact. That turns AI into part of the growth operating system rather than another tool stack.

What should a marketing team do in the first 30 days of AI adoption?

Audit current workflows, identify repeated tasks, collect key business knowledge, define quality standards, identify the tool stack, and test AI on a small number of high-value use cases. The goal is not to automate everything. It is to prove where AI improves clarity, focus, speed or decision quality before you scale.

Turn AI experimentation into focused growth execution

For marketing leaders, founders and growth teams, the question is not which AI tool to buy next. The better question is where better clarity, stronger context and sharper focus would change the result.

The tools will keep changing. They’ll keep getting more intelligent, and the noise around them isn’t going anywhere. Anchoring your strategy on a specific tool stack is a losing game. The craft, the foundations, and the information that differentiates your business: that is where the advantage is being built right now. The teams that win with AI won’t be the ones with the longest list of tools. They’ll be the ones who did the unglamorous work underneath them. If your team is experimenting with AI but not yet seeing commercial impact, start with the operating system around the work: the priorities, the knowledge base, the decision rhythm and the customer intelligence that make AI useful. If you’d like help turning AI experimentation into focused growth execution, book a growth clarity conversation.

About the author

Marinda Malan is a Growth Adviser and Consultant, and Fractional Growth Executive based in Dublin, Ireland. She helps founders, leadership teams and marketing teams diagnose growth constraints, clarify priorities and build practical revenue operating systems. Her work includes growth diagnostics, fractional CGO support and the M3: Mission Means Machine framework. Connect with her on LinkedIn or read more about Marinda.