The Blueprint: How to Build an AI-Enhanced Marketing Function That Actually Delivers
Part 5 of 5 — AI in Marketing: From Experimentation to Commercial Advantage
This is the part of the conversation where most AI marketing content reaches for a framework with five steps and an optimistic conclusion.
What follows is more useful than that: an honest account of what it actually takes to build a marketing function that uses AI well - commercially, responsibly, and in a way that compounds over time.
The businesses that get this right do not necessarily have the biggest AI budget, they simply had the clarity before they started.
Where Should a Business Start When Building an AI-Enhanced Marketing Function?
Start with the business problem, not the technology.
The AI strategy for a marketing function should begin with a clear picture of where the function is underperforming commercially — not with a scan of the latest tools and a conversation about which ones to try.
That means being specific.
- Not "we want better personalisation" — but "we are losing customers after the first transaction and we do not understand why."
- Not "we want to produce more content" — but "our sales team does not have the materials they need to move prospects from consideration to decision, and deals are stalling."
- Not "we want to use AI for insight" — but "our customer data sits in three separate systems and nobody has a clear picture of which customers are most valuable or why."
The use cases that follow from a real problem are more commercially valuable than those following from a technology trend. It is also more likely to get funded, adopted and measured properly, because there is a clear commercial reason for it to exist.
What Is the Right Sequence for Implementing AI in a Marketing Function?
The instinct to move fast on AI is almost natural because the competitive pressure is real. But the businesses making the most progress on AI-driven marketing value are those that moved in the right sequence, not necessarily the fastest.
The right sequence looks broadly like this:
- Data foundations before AI applications;
- Governance before scaling;
- Human review processes before automation;
- Clear proposition and positioning before AI-generated content at volume.
Imagine if AI personalisation applied to fragmented, inaccurate customer data, it personalises the wrong message to the wrong person at the wrong time.
AI content tools applied before a clear brand voice and editorial standard exists produce content that is fast but indistinct. Automated customer journeys built before customer segmentation is properly defined create workflows that move people efficiently through a process not designed for them.
The sequencing conversation will save you from undoing expensive mistakes later.
What Should AI Marketing Governance Actually Look Like in Practice?
Governance in the AI context does not mean a lengthy approval process for every piece of content.
It means agreeing, at the right level of seniority : What AI is for in your marketing function
- Which tools are approved for which use cases,
- What data can and cannot be entered into external AI systems,
- Which outputs require human review before use,
- Who holds accountability for final sign-off, and
- What happens when something goes wrong.
It takes a day to make these decisions, but a very long time to correct if they were never made at all.
The DMA UK's 2026 position on AI in marketing is explicit: brands remain responsible for AI-assisted decisions — including content generated by AI, audiences built by AI and personalisation decisions made by AI.
The technology does not transfer the accountability. The business retains it.
For regulated sectors, the FCA's expectations around explainability, fair treatment and customer vulnerability apply regardless of whether customer-facing output was written by a human or generated by a tool. Getting governance right is what allows the business to use AI confidently and at scale.
How Do You Maintain Quality When AI Is Generating Marketing Output at Scale?
Define what quality looks like before you scale it.
One of the most common mistakes in AI marketing adoption is the absence of a quality standard. Teams start producing content faster, campaign variations multiply, and reporting dashboards fill up with more data, but no one has defined what good looks like, so there is no way to know whether the production is simply faster or driving better outcomes.
This connects directly to the commercial outcome question. If AI is being used to produce more content, the measure of success has to be whether that content is moving the right people toward the right decisions, not the volume.
The practical implication:
Every AI use case in marketing should have a clear commercial KPI attached to it, and someone responsible for reviewing whether the AI-assisted work is meeting that standard on an ongoing basis.
AI output degrades over time if it is not reviewed and recalibrated. The quality discipline has to be built into the process, not added retrospectively.
What Should Humans Be Doing That AI Cannot Replace in a Marketing Function?
"What can AI do?" Almost everything if you let it. But we should be clear about "What should humans be doing when AI takes them out of the weeds?"
If AI handles the drafting, the summarising, the variation testing and the reporting compilation, what does that free up?
In well-run marketing functions, it frees up time for the work that actually requires human judgement:
Understanding customers more deeply, sharpening the proposition, improving the sales and marketing relationship, building better measurement frameworks, and thinking more rigorously about which markets and segments to prioritise.
The teams get the most from AI are those who use AI and redirect human involvement toward higher-value work.
The skill that becomes most important is not AI fluency. It is commercial and strategic judgement — knowing what good looks like, what the business needs, and whether the output in front of you is genuinely serving the customer and the commercial strategy.
What Are the Practical Next Steps for a CEO Building an AI-Enhanced Marketing Function?
These are the five decisions that matter most, in order.
- Conduct an honest audit of where marketing is underperforming commercially. Not where it is slow — where it is failing to create the commercial value the business needs. That is the problem AI should be applied to.
- Assess data readiness before AI readiness. Customer data, campaign data, CRM data, product data — are they clean, connected and accessible enough to power AI-assisted decisions? If not, that is the first investment.
- Make the governance decision at the right level of seniority. AI in marketing is not a technology decision to delegate to the digital team. It is a business decision about how customer relationships, brand standards and commercial accountability are managed. It belongs in the leadership conversation.
- Invest in the human layer, not just the technology layer. Training, review processes, clear editorial standards, commercial judgement — these are what turn AI capability into AI value. Without them, the tools underperform regardless of how sophisticated they are.
- Measure the commercial outcome, not the activity. How much time was saved matters less than whether that time was reinvested in better work. How much content was produced matters less than whether it moved the right customers toward the right decisions.
How long does it take to build an AI-enhanced marketing function?
The foundations — data audit, governance framework, quality standards, team training — can typically be established in sixty to ninety days. Meaningful commercial impact from AI-assisted workflows tends to appear within six months for businesses that have done the foundational work properly. Businesses that skip the foundations often see initial productivity gains but plateau quickly.
Do we need a dedicated AI lead in our marketing team?
Not necessarily. What you need is someone at a senior enough level to define where AI belongs in the marketing operating model, set quality standards and hold governance accountability. In many businesses, that is the CMO or a Fractional CMO. A dedicated AI role only makes sense at scale, once the use cases and governance framework are already established.
What is the single most important thing a CEO can do to improve AI outcomes in marketing?
Stay involved in defining what success looks like — not just in terms of AI tool adoption, but in terms of commercial outcomes. The CEOs seeing the strongest AI marketing returns are the ones who asked "what problem are we solving and how will we know we solved it?" before approving any investment.
How do we know when we are ready to scale AI use in marketing?
Three signals: your data foundations are in good shape, your governance framework is agreed and communicated, and you can clearly connect at least one AI use case to a measurable commercial improvement. If you have all three, scaling is worth doing. If any is missing, scaling will amplify the gap rather than close it.
The Conclusion That Has Run Through All Five Articles
AI in marketing is not a strategy. It is a capability — one that amplifies whatever is already in the system.
If the strategy is clear, the data is reliable, the proposition is sharp and the team has the judgement to use AI well, the amplification is genuinely powerful. If those foundations are not in place, AI makes the weakness louder and faster.
The businesses furthest ahead in three years will not be the ones that experimented most aggressively with every new tool. They will be the ones that decided clearly what their marketing function needed to do commercially, built the right foundations to support it, applied AI where it genuinely improves quality and efficiency, and maintained the human judgement the technology cannot replace.
That is not a technology story. It is a leadership story. And it has always been one.
Talk to ZILU about AI in Marketing
If your organisation is exploring how to move from AI experimentation to practical marketing implementation, ZILU Consultancy can help you assess where to start, what to prioritise, and how to build a more practical AI roadmap for your marketing team.
Book a conversation with Catherine.
Continue to learn about AI in marketing in our 5-part series.
Part 1: AI in Marketing: How many businesses are getting value from it?
Part 2: The CMO's AI Agenda: How much are actually spending on AI- and is it working?
Part 3: Why is AI in marketing underdelivering for most businesses?
Part 4: How Should Businesses Implement AI in Marketing?
This series draws on 2025–2026 research from McKinsey, Gartner, BCG, Deloitte, Salesforce, The CMO Survey, HubSpot, DMA UK, CIM, IAB UK and the Bank of England / FCA. Sources listed in Part 1 of 5.