There Is No Universal AI Marketing Playbook. Here Is How to Find Yours.
Part 4 of 5 — AI in Marketing: From Experimentation to Commercial Advantage
One of the most persistent mistakes in the AI in marketing conversation is treating it as a single conversation.
As though a founder-led business of fifteen people and a regulated financial services firm with a thousand-person marketing function should be following the same playbook.
They should not.
The right AI marketing strategy depends on the stage of the business, the maturity of the team, the quality of the data, the regulatory environment and the commercial priority.
Getting this wrong is expensive — not because the tools fail, but because the wrong use cases get prioritised and the real problems go unsolved.
What Is the Right AI Marketing Approach for a Startup or Founder-Led Business?
At the earliest stage of growth, the AI opportunity is always a leverage.
A founder or small team using AI well can operate with the capabilities of a much larger function — covering content, research, customer communications, positioning testing and basic analytics — without the headcount.
The most practical starting points are close to the work that already exists but takes too long: synthesising customer conversations, drafting first versions of messaging, building SEO foundations, supporting sales outreach and maintaining CRM discipline.
The risk at this stage is inconsistency — particularly in brand voice — and over-reliance on AI output that has not been reviewed by someone with enough context to judge whether it is right.
For founder-led businesses especially, the brand is often the founder's voice and perspective. Generic AI output can quietly erode that distinctiveness before anyone notices.
The discipline here is staying close enough to what is being produced to maintain quality, and being deliberate about where the human voice is irreplaceable.
How Should a Growing SME Use AI in Marketing Without Creating More Chaos?
Businesses at the SME stage often have a marketing function that is busy but not structured. Content gets produced, campaigns run, events get attended — but the underlying commercial architecture is not always clear.
AI used here without addressing those underlying gaps tends to accelerate the wrong things. More content without a clearer message. More campaign variations without better targeting. More activity without better commercial accountability.
The most valuable use of AI at this stage is in building process and insight — not scaling production.
Using AI to
- Extract patterns from customer conversations and surface real objections.
- Tighten segmentation and make the proposition more specific.
- Automate reporting so the marketing leader spends less time compiling data and more time interpreting it.
- Improve the handoff between marketing and sales so what marketing generates is actually usable by the people in the customer conversation.
These applications are less visible than a dramatic increase in content output. They are significantly more valuable to the business.
What Should a Scaleup Prioritise When Implementing AI in Marketing?
Scaleups face a specific challenge. The business has moved fast enough to build real momentum, but often without building the systems and clarity that should underpin it.
Marketing at this stage can become fragmented — different teams or regions running different campaigns, inconsistent messaging, multiple technology platforms that do not connect, and a commercial story that has evolved faster than the marketing infrastructure tracking it.
For companies at this stage, AI can certainly create real value if it is introduced alongside the structural work. However, the clarity and structure in marketing in scaleups are often missing because they are moving too quickly to stop and design properly.
So, enabling AI to work wonder,
The priority is workflow integration. AI-assisted segmentation and nurture programmes work when built on the CRM data that is well-maintained. AI-powered reporting creates value when the measurement framework is clear. AI-generated sales enablement content helps when the proposition is already agreed.
The temptation is to use AI to go faster. But the smarter move is using AI to create the infrastructure and scalability that support growth. Scaleups that use AI to paper over operational gaps will only amplify the gaps.
How Should Large Enterprises Approach AI in Marketing Differently?
At enterprise scale, the AI opportunity is as significant as its governance complexity.
The opportunity at scale is in customer intelligence: unified data, AI-assisted decisioning, personalisation across the full customer lifecycle, media optimisation, content production at volume without sacrificing quality, and — increasingly — agentic AI handling multi-step workflows with minimal human intervention at execution level.
BCG's research shows only a small minority of CMOs currently have AI operating autonomously across campaigns. But the investment is moving in that direction,
The enterprises building the data infrastructure and governance frameworks now will have a significant advantage in three years over those that waited.
The governance question for large organisations is not just about brand safety and compliance risk, though those matter. It is also about how AI integrates with existing technology infrastructure.
Many enterprises are still running fragmented data environments — customer data in one system, campaign data in another, product and transaction data elsewhere, with no clean unified layer. AI applied to this environment produces fragmented intelligence.
The organisations that invest in building a governed data foundation first, and AI capability on top of it, are the ones that will genuinely unlock the personalisation and decisioning potential the research describes.
What Makes AI in Marketing Different for Financial Services, Insurance and Regulated Sectors?
Banking, insurance, financial services and healthcare carry a category of risk in AI adoption that most industries do not face to the same degree.
The Bank of England and FCA's 2024 survey found that
- 75% of UK financial services firms are already using AI — with another 10% planning to within three years. That is a sector that has moved fast. The regulatory framework around it is moving at the same pace.
Customer-facing AI in financial services need to consider its impact on trust, suitability and fairness:
- Whether an AI-assisted customer communication is accurate.
- Whether automated personalisation creates differential treatment that cannot be explained.
- Whether customer vulnerability is being properly considered in AI-driven decisioning.
- Whether the business can demonstrate clearly to a regulator what its AI systems are doing and why.
Allstate's US example is instructive. The company now generates the majority of its claims-related customer communications through AI, grounded in company-specific terminology, producing clearer and more empathetic communications. That is a genuine win,
that can only be achieved with human oversight, purpose-built governance and a clear definition of what the AI was and was not responsible for.
That level of intentionality is not optional in a regulated sector. It is the baseline.
What Is the One Question That Cuts Through All of This Regardless of Business Size?
Regardless of size or sector, the practical question is always the same:
What is the real commercial problem we are trying to solve, and is AI the right tool to solve it?
The CEO who gets this right asks what the business most needs from its marketing function in the next twelve months — and then asks where AI can genuinely help close the gap between where that function is now and where it needs to be.
The right playbook is always the one built around the real problem, not the most impressive case study from a company at a completely different stage of maturity.
Should a startup invest in AI marketing tools before they have product-market fit?
Use AI for research, messaging iteration and content efficiency — but not as a substitute for direct customer conversations. AI is most valuable for speed and scale. At early stage, the priority is still getting the proposition right, which requires human insight, not faster production.
What is the biggest AI marketing mistake scaleups make?
Using AI to accelerate a marketing function that is already fragmented. If the messaging is inconsistent, the CRM is poorly maintained and sales and marketing are misaligned, AI will scale all of those problems simultaneously. Fix the foundations before applying the technology.
How do regulated businesses balance AI marketing innovation with compliance?
Start with use cases that are clearly internal-facing or low-risk: content drafting, internal research synthesis, reporting automation. Build governance and audit capability alongside them. Only move AI into customer-facing communications once the review process, explainability framework and sign-off structure are in place.
Can a small business compete with large enterprises using AI in marketing?
Yes — and in some ways more effectively. Large enterprises are slowed by legacy systems, governance complexity and internal politics. A small business with clean data, a clear proposition and disciplined AI use can move faster and more precisely. The competitive advantage is not budget. It is clarity.
Catherine Mak is Founder of ZILUconsultancy. She works as a Fractional CMO with growth-stage businesses, scaleups and organisations navigating AI adoption in their marketing functions. Clients include Citibank, Bupa Global and PEI Group.
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 5: How to Build an AI-Enhanced Marketing Function That Works?