AI strategy in marketing: why 85% of companies still don’t have one (2026 data)

Almost everyone in marketing now uses artificial intelligence for something: copywriting, segmentation, image generation, campaign analysis. And yet very few teams could explain what their plan is. That gap between “we use AI” and “we have a AI strategy in marketing” has a name of its own in 2026: the AI readiness gap. And the latest figures suggest it is much wider than companies would like to admit.

At Vandelay, we see this every week. We come across businesses that are already paying for three or four AI tools, which they use on a daily basis, and yet they still can’t say exactly what they’re improving, who’s responsible, or how they measure whether it’s working. It’s not a lack of interest: it’s a lack of direction. This article explains, with verifiable data, why this happens and what you can do to ensure your investment in AI stops being a collection of isolated experiments.

What is the AI readiness gap?

The AI readiness gap is the disparity between the pressure organisations feel to adopt artificial intelligence and their actual capacity to deploy it in a responsible and measurable manner. In short: it is the gap between what is expected of AI and what the company is prepared to deliver with confidence.

The concept has gained momentum in the wake of the report AI Readiness Gap published by Supermetrics on 14 July 2026, based on a survey of 435 marketing managers in the United States, the United Kingdom, Germany, Australia and Singapore. Its most frequently cited conclusion is an uncomfortable one: 85% of organisations do not have a formal AI strategy or lack clear ownership of such initiatives. Only 15% states that it has a defined roadmap with measurable success metrics.

The nuance matters. The report emphasises that preparing for AI is not, above all, a question of technology, but rather one of responsibility and implementation. The tool is available to anyone; what is in short supply is knowing what to use it for, with what data, and by what criteria for success.

Data for 2026: plenty of AI, little strategy

The picture painted by this year’s studies is consistent: adoption has soared, but maturity has been left behind. Here are the clearest indicators, along with their sources, so that you can check them for yourself.

That only 61 per cent of professionals have fully integrated AI into their workflows — according to the 2026 Marketing Data Report by Supermetrics, published in February—doesn’t mean that the rest don’t use it at all. It means that most use it from time to time, in isolated tests, without it forming part of a consistent process. It’s the difference between starting the engine every now and then and having a car that actually takes you somewhere.

Using AI is not the same as having an AI strategy

Here lies the fundamental misunderstanding. Many companies confuse activity with progress. Generating a hundred copies using a model or creating images in seconds gives a comforting sense of “doing AI”. But a AI strategy in marketing It isn’t measured by how many things you automate, but by which business problem you’re solving and how you know you’re solving it.

Most AI experiments stem from enthusiasm, rather than a defined use case or an integrated data model. The result is silos: the content team tries one thing, the paid advertising team another, and nobody shares what they’ve learnt or measures the actual impact. McKinsey, in its report The State of Trust in AI in 2026, describes this very moment as the transition to the “agent-driven era”, in which trust and governance are developing more slowly than the technology itself. When the capacity to act grows faster than the capacity to control, further automation merely leads to mistakes being made sooner.

This line of reasoning ties in with something we have already explained in our article on Agent-based marketing and AI agents in campaigns: Autonomous agents show great promise, but without a strategy and without clean data, they simply amplify errors. AI does not fix a broken process; it simply runs it faster.

How to develop an AI marketing strategy, step by step

You don’t need a data science department or a multinational’s budget to move from organised chaos to a proper plan. What you do need is structure and a sensible sequence of steps. This is the approach we take at Vandelay when we help an SME or a growing brand chart a course for their AI.

1. Define one or two use cases with a measurable impact. Instead of “using AI for everything”, focus on specific problems: reducing content production time, improving the response to leads or refining segmentation. Fewer fronts, greater depth.
2. Assign an owner. The 85% falls short precisely here. Without someone – by name – taking responsibility for each initiative, AI remains nothing more than a footnote. It doesn’t have to be a new hire; it could be a role within the existing team.
3. Get your data in order before you work on your models. Only 11% of the organisations consider their data to be of high quality and accessible. A model fed with poor-quality data produces convincing but erroneous conclusions. Start there.
4. Brings together analytics and activation. The lack of integration between data and execution platforms is one of the biggest bottlenecks (as highlighted by 40% of SME teams and 34% of large enterprise teams in the Supermetrics report).
5. Set success metrics from day one. What you’re going to improve, by how much, and in how long. Without this, any result looks good and none can be compared.
6. Review with a human touch. AI makes suggestions; people decide and respond. Oversight isn’t a hindrance – it’s what makes your marketing justifiable to a client, a committee or a regulator.

Note that none of these steps begins with “buy another tool”. Strategy comes before software, not the other way round. It is this very approach that underpins more advanced practices such as AI-powered hyper-personalisation or a AI-powered marketing measurement that goes beyond the last click: without foundations, there can be no building.

The real bottleneck: data and who’s in charge

If we had to sum up the readiness gap in two words, they would be ‘data’ and ‘accountability’. The figures make it clear: only 7% of teams receive responses to their data requests in real time, whilst half wait between one and three working days for a simple, one-off enquiry. By that time, the opportunity to respond to a campaign has already passed.

That is why the first-party data have become the true raw material of AI-driven marketing. This is no coincidence: the better you understand and control your own data, the more useful and reliable any layer of artificial intelligence you apply to it will be. Brilliant AI applied to poor data is still a poor decision – just one made with more confidence than it deserves.

Common mistakes when adopting AI without a clear direction

Before you take the plunge, it’s worth recognising the most common pitfalls. The first is acting under pressure: doing it because “everyone else is doing it” rather than because it solves a problem of your own. The second is measuring the activity rather than the outcome: counting how many pieces the AI generates rather than how much it improves the business. The third is to leave the initiative in limbo, with no one to lead it. And the fourth, perhaps the most dangerous, is to delegate judgement: relying on an automated recommendation without understanding where it comes from. Transparency is not just a nice-to-have; in a context where legal obligations regarding the use of AI already exist, it is also a matter of compliance.

Frequently asked questions about AI strategy in marketing

What is the AI readiness gap?

It is the gap between the pressure to adopt artificial intelligence and an organisation’s actual capacity to use it responsibly and in a measurable way. In 2026, 85% of companies lack a formal AI strategy or a clear point of contact, according to Supermetrics.

Do I need a large technical team to implement an AI strategy?

No. What matters most is not the size of the team, but the structure: specific use cases, a lead for each initiative, clean data and success metrics. Many SMEs start with a single, well-chosen use case and an in-house lead.

Where should an SME start?

Based on your data and a real-world problem. Before adding new tools, it’s worth ensuring that your own data is organised and accessible, and choosing one or two areas where AI can drive a metric that matters to the business.

Does AI replace human decision-making in marketing?

It shouldn’t. The figures for 2026 show that governance is advancing more slowly than technology. AI proposes and executes, but the responsibility — and the explanation of why something is done — still lies with people.

If, whilst reading this, you’ve recognised your own situation, don’t worry: you’re in the 85%, not an exception. The good news is that bridging the gap doesn’t require more technology, but rather better judgement. At Vandelay, we help you move from “we use AI” to “we know exactly what it does for our business”, with a bespoke, no-nonsense roadmap. If you’d like to assess where you stand today and what you’d need to be sure of before continuing to invest, Let's have a chat and we put figures to the conversation.

Sources of the data cited: report AI Readiness Gap from Supermetrics (July 2026); 2026 Marketing Data Report from Supermetrics (February 2026); and report The State of Trust in AI in 2026 by McKinsey.