AI in Marketing: Where Should Businesses Start?

5 hours ago

5 min read

AI is already changing how marketing teams research, create, analyse and run campaigns, and the pressure to “do something with AI” has become impossible to avoid. Fair enough. But buying a platform before you know what problem you’re solving is not a strategy. It’s an expensive way to look innovative.

I help businesses work through technology and marketing change by connecting strategy with practical execution, and my advice on AI is the same as my advice on anything else: start with a business need, choose a manageable workflow and measure whether the change actually helped.

You don’t need an enterprise-wide AI programme to begin. You need one clear first step and the discipline to judge it honestly.

Start with the work, not the tool

Before you compare AI platforms, look at how your marketing gets done today.

Where is the team spending time on repetitive work? Which tasks slow campaigns down, create handoff problems or make it hard to get useful insight from your data? The best first opportunity is often a routine process where a little assistance could save time or improve consistency.

For example, AI could help your team:

Summarise research or customer feedback for a campaign brief.
Create a first draft of content for a marketer to check and improve.
Produce variations of campaign copy for testing.
Summarise performance reports and flag questions for further analysis.
Recommend how to route or prioritise leads for a person to review.

These are starting points, not automatic wins. The right choice depends on your current process, data and team. I’d begin with one clearly defined task rather than trying to transform several parts of marketing at once.

A small cross-functional team mapping a marketing workflow in a modern workspace

Choose a problem you can measure

A useful pilot has a specific purpose and a way to judge whether it worked. “We want to use AI” is not a business objective. It’s a sentence people say in meetings. “We want to reduce the time it takes to prepare a first campaign draft, without lowering quality” is much closer.

Before changing a process, record a simple baseline. Depending on the task, that might include:

How long the work currently takes.
How many people are involved.
How much revision or rework is typical.
Whether the work meets your quality and brand standards.
What happens if the process is late or inaccurate.

Then agree what improvement would make the pilot worth continuing. You might look at time saved, fewer errors, faster campaign delivery or more time for strategic work. Don’t measure success by the volume of AI-generated content. Measure whether the work is better for your business.

Check your marketing operations and data

AI is only useful in context. If your customer data is unreliable, your systems don’t connect or your team can’t agree on how leads are handled, adding AI won’t fix any of that. It will just automate the mess, faster.

Take a quick look at the marketing technology and operations behind the workflow you want to improve:

Where does the data come from, and is it accurate enough to use?
Which tools or teams need to be involved?
Are there manual steps, unclear ownership or duplicate records?
What information should not be entered into an AI tool?
How will you check and approve the AI’s output?

You don’t need to rebuild your martech stack to run a small test. But you do need to understand what the test touches and who is accountable for the process.

A marketing operations lead reviewing campaign data and a process checklist

This is where a focused marketing operations strategy can help. In my work with beBettor, a technology business in a regulated sector, I supported its repositioning, go-to-market strategy and sales process. That work shows why marketing technology, commercial priorities and the way a team operates need to fit together, especially when the market is complex.

Set clear boundaries for the pilot

AI can produce convincing answers that still contain errors. It can also mishandle information if people use tools without clear guidance. For a first pilot, keep a person responsible for reviewing outputs before they reach customers or affect important business decisions.

Agree a few practical rules before you begin:

Use approved tools and follow your company’s data and privacy policies.
Don’t enter confidential or personal information unless the tool and its use have been properly assessed.
Check facts, claims, tone and brand fit before publishing.
Keep human approval for customer communications, targeting decisions and changes to campaign activity.
Make it clear who owns the process and what to do if something goes wrong.

For a more structured approach, the NIST AI Risk Management Framework offers guidance for organisations thinking about AI risks. You don’t need to turn a small experiment into a complex compliance project. You do need to take sensible care with your data, customers and reputation.

Abstract marketing heatmap representing campaign measurement and areas of customer engagement

Keep AI connected to your go-to-market strategy

AI should support how you reach the right customers, explain your value and grow revenue. It shouldn’t become a separate initiative that produces activity without supporting a commercial goal.

That matters for AI GTM: the way AI can support go-to-market planning and execution. For example, a team might use AI to organise customer research, prepare an initial campaign outline or review performance data. The strategic decisions still need to come from a clear understanding of the audience, the offer and the market.

Ask yourself:

Will this help us serve a priority audience?
Does it support a real campaign or sales objective?
Can we connect the work to a useful measure, such as qualified leads or campaign efficiency?
Does the team know how to use the output in the wider go-to-market process?

If the answer is unclear, pause and clarify the strategy before adding another tool.

A practical first 30 days

You can make a sensible start without committing to a full AI transformation.

Week 1: Find the opportunity. Talk to the people doing the work. Choose one repetitive marketing task that has a clear owner and a noticeable cost in time or effort.

Week 2: Map the process. Note the steps, tools, data and approvals involved. Record your baseline and decide what information must stay out of the pilot.

Week 3: Run a controlled test. Use an approved tool to support the task. Keep the scope narrow and have a person check every output before it is used.

Week 4: Review the evidence. Compare the results with your baseline. Did the process improve? Was the output accurate and useful? Did the team spend less time overall, including on review and corrections? Decide whether to stop, adjust or expand the test.

This is a practical way to learn what AI can do for your business without making a large investment based on assumptions.

Move forward with evidence, not pressure

The pressure to “do something with AI” is real, but acting quickly doesn’t mean rushing into a platform or redesigning your entire marketing function. It means choosing a worthwhile problem, setting sensible boundaries and finding out whether AI genuinely helps. The businesses getting this right aren’t the ones shouting loudest about AI. They’re the ones quietly proving which parts of it pay.

I can help you assess your martech stack, improve marketing operations and build a go-to-market strategy that connects AI experiments to business priorities. If you’re unsure where to begin, book a free discovery call. We can look at your current setup and identify a practical next step.