AI and Business Strategy: Most Companies Are Asking the Wrong Questions

Artificial intelligence and business strategy intersect most productively when organisations stop asking “what can AI do?” and start asking “what business problem are we actually trying to solve?” The technology is capable. The strategy, in most companies, is not keeping pace with it.

That gap is where most AI investments quietly underperform. Not because the tools fail, but because the strategic framing was wrong before anyone opened a browser tab.

Why Most AI Strategies Fail Before They Start

I have sat in a lot of strategy sessions over the years. The pattern I see with AI right now is almost identical to what I saw with programmatic advertising in the early 2010s, and with marketing automation before that. A genuinely capable technology arrives. Vendors package it attractively. Leadership teams feel pressure to act. And so businesses invest in the technology before they have defined what success looks like.

The result is a collection of AI tools that are used tactically, measured loosely, and celebrated internally without any honest accounting of whether they moved a commercial needle. I have seen this in agencies. I have seen it in corporate marketing teams. The investment gets made. The press release goes out. The quarterly review slides show adoption metrics. And somewhere underneath all of that, nobody has answered the question that should have come first: what specific business outcome were we trying to improve?

This is not a technology problem. It is a discipline problem. And it is the same discipline problem that has plagued marketing budgets for decades.

If you want a grounded view of how AI tools are actually being evaluated and deployed in marketing right now, the AI Marketing hub at The Marketing Juice covers the practical landscape without the vendor hype. It is worth reading alongside this piece.

What Good AI Strategy Actually Looks Like

Good AI strategy is not a technology roadmap. It is a prioritised list of business problems, ranked by commercial impact, with AI considered as one possible input to solving them. That framing might sound obvious. It is apparently not, given how many organisations approach this the other way around.

When I was running an agency and we were scaling from around 20 people to closer to 100, the operational pressure was constant. Briefing processes, quality control, client reporting, resource allocation. These were real problems with real commercial consequences. If AI tooling had existed then in its current form, the question I would have asked is not “which AI tools should we adopt?” It would have been “which of these operational bottlenecks is costing us the most, and can AI reduce that cost without introducing new risks?” That is a fundamentally different starting point, and it leads to fundamentally different decisions.

The businesses that are getting the most from AI right now tend to share a few characteristics. They have clear ownership of the AI initiative at a senior level. They have defined the problem before selecting the tool. They have an honest measurement framework in place before deployment, not after. And they treat AI as an operational input rather than a strategic identity. They are not trying to be “an AI company.” They are trying to run a better business.

The Measurement Problem Nobody Wants to Talk About

Here is the uncomfortable truth about AI and business strategy: if your measurement infrastructure was weak before you introduced AI, it will be weak after. AI does not fix measurement. It amplifies whatever measurement capability you already have.

I spent years judging the Effie Awards, which are specifically designed to recognise marketing effectiveness. What strikes you, doing that work, is how rarely organisations can draw a clean line between a marketing activity and a business outcome. They can show reach. They can show engagement. They can show brand tracking scores. But the commercial causality, the honest answer to “did this make money?”, is elusive in the majority of cases. And that was before AI entered the picture.

Now organisations are layering AI-generated content, AI-assisted targeting, and AI-optimised creative on top of measurement frameworks that were already struggling to tell the truth. The output volumes increase. The attribution gets murkier. And the quarterly review slides look busier than ever, which is sometimes mistaken for evidence of progress.

If you are building an AI strategy, the measurement question has to come first. What are you measuring today? How confident are you in that measurement? What would a meaningful improvement in that metric actually mean for the business commercially? If you cannot answer those questions cleanly, no AI tool will answer them for you. Semrush’s analysis of AI content strategy touches on this, noting that volume and visibility are not the same thing as commercial impact. That distinction matters more than most teams acknowledge.

Where AI Creates Genuine Strategic Value

There are areas where AI creates real, defensible strategic value. They are worth naming clearly, because the hype tends to blur them.

Speed of analysis is the most straightforward. AI can process and synthesise large volumes of data faster than any human team. For businesses that make decisions based on market signals, competitive intelligence, or customer behaviour patterns, this is a genuine advantage. Not because the AI is smarter, but because it removes the time bottleneck between data and decision.

Content production at scale is real, with caveats. The caveats matter. AI-generated content is only as good as the strategic input that precedes it. I have reviewed enough AI-generated marketing output to know that the quality ceiling is set by the brief, not the model. When the brief is sharp, the output can be genuinely useful. When the brief is vague, the output is expensive mediocrity. Buffer’s overview of AI tools for scaling business operations makes a similar point about the importance of process design before tool adoption.

Personalisation at scale is where the strategic upside is most significant and most frequently overstated. The technology can support highly personalised customer experiences. But personalisation only creates value if it is relevant, and relevance requires a depth of customer understanding that most organisations have not done the work to develop. AI can execute personalisation. It cannot substitute for the customer insight that makes personalisation worth doing.

Operational efficiency is the quietest and most consistent source of value. Automating repetitive tasks, reducing manual reporting, accelerating briefing cycles. These are not glamorous applications, but they compound over time. When I look at where AI is actually delivering measurable return in marketing organisations, it is more often in these unglamorous operational improvements than in the headline-grabbing creative applications.

The Competitive Advantage Question

A question I hear regularly from senior marketers and business leaders is some version of: “If everyone has access to the same AI tools, where does competitive advantage come from?”

It is a good question, and the honest answer is that it comes from the same place it has always come from. Clarity of strategy. Quality of customer understanding. Speed and quality of decision-making. Organisational discipline. AI does not change the sources of competitive advantage. It changes the speed at which advantage can be built or eroded.

Across the industries I have worked in, from retail to financial services to FMCG, the businesses that outperformed their competitors over time were rarely the ones with the best technology. They were the ones with the clearest view of what they were trying to achieve and the discipline to pursue it consistently. AI amplifies that clarity when it exists. It also amplifies the confusion when it does not.

The organisations that will extract the most from AI over the next five years are not necessarily the ones with the largest technology budgets. They are the ones that have done the harder, less exciting work of defining their commercial priorities clearly, building honest measurement frameworks, and treating AI as a means to a business end rather than an end in itself. Ahrefs has covered the practical application of AI tools in ways that reinforce this point: the tool selection question is secondary to the strategic clarity question.

How to Build an AI Strategy That Is Commercially Grounded

This is not a framework in the consulting sense. It is a sequence of questions that most organisations skip past in their rush to adopt.

Start with the commercial problem. Not the marketing problem, not the technology problem. The commercial problem. Where is the business underperforming against its potential? Where are the operational bottlenecks that have a direct cost? Where is the decision-making slower than it needs to be? These are the entry points for a grounded AI strategy.

Then ask whether AI is the right solution. Sometimes it is. Sometimes the problem is a process problem, a people problem, or a data quality problem that AI will not fix and may obscure. I have seen businesses invest in AI-powered analytics on top of data that was inconsistently collected and poorly governed. The output looked sophisticated. The insight was unreliable. The decisions made on the back of it were no better than before.

If AI is the right solution, define the measurement framework before deployment. What does success look like in commercial terms? Not in activity terms. Not in adoption metrics. In commercial terms. Revenue, margin, customer retention, cost per acquisition, time to decision. Pick the metric that matters to the business, establish the baseline, and commit to measuring against it.

Then select the tool. Not before. The tool selection conversation, which tends to dominate AI strategy discussions, should be one of the last conversations, not the first. HubSpot’s comparison of LLM options is useful at this stage, when you have a defined use case and need to match capability to requirement.

Build in a review cadence that is honest. Not a review that celebrates adoption. A review that asks whether the commercial metric moved, and by how much, and whether AI was the cause. That last question is harder than it sounds, which is why most organisations skip it. But skipping it means you are flying blind, and flying blind with an AI co-pilot is not materially better than flying blind without one.

The Organisational Readiness Nobody Audits

One of the least discussed dimensions of AI strategy is organisational readiness. Not technical readiness. Organisational readiness. The capability of the people in the business to use AI tools well, to interrogate their outputs critically, and to integrate AI-generated work into commercial processes without losing the human judgement that gives it value.

I have managed marketing teams across a wide range of sizes and capability levels. The consistent pattern is that tools are only as good as the people using them. A junior content writer using an AI content tool without a strong strategic brief will produce more content faster. It will not produce better content. A senior strategist using the same tool with a sharp brief and critical editorial judgement will produce something genuinely useful. The difference is not the tool. It is the person holding it.

Organisational readiness for AI means investing in the human capabilities that make AI outputs valuable: strategic thinking, critical evaluation, commercial judgement, and the ability to write a brief that is worth executing. These are not new skills. They are the skills that have always separated good marketing from expensive noise. AI raises the stakes on them rather than replacing them.

The Moz analysis of AI SEO tools makes a point worth noting here: the teams getting the most from AI-assisted SEO are not the ones with the most tools. They are the ones with the clearest understanding of what they are trying to rank for and why. That holds across every AI application I have observed.

What the Next Three Years Actually Look Like

Predictions in this space age badly, so I will stay close to what the commercial logic suggests rather than speculating about specific capabilities.

The businesses that treat AI as a strategic input rather than a strategic identity will outperform those that do not. The gap between AI-assisted decision-making and unassisted decision-making will widen in areas where data quality is high and the decision criteria are well-defined. In areas where the data is messy and the criteria are ambiguous, which describes most of marketing, the gap will be smaller than the vendors suggest.

The measurement problem will become more acute before it gets better. As AI-generated content volumes increase across the web, the signal-to-noise ratio in digital marketing data will deteriorate. Research on how LLMs process and cite content points to a shifting landscape for organic visibility that has direct implications for how businesses measure the impact of their content investment. Businesses with strong, first-party measurement capabilities will have a significant advantage over those relying on platform-reported metrics.

And the organisations that will look back in three years and feel genuinely good about their AI investments will be the ones that started with a commercial question, not a technology one. That is not a complicated insight. It is just one that requires discipline to act on when every vendor in the market is telling you to move fast.

There is more on the practical application of AI across marketing disciplines, from content to paid media to analytics, in the AI Marketing section of The Marketing Juice. If you are building or refining an AI strategy for a marketing function, it is a useful reference point alongside the commercial framing in this piece.

About the Author

Keith Lacy is a marketing strategist and former agency CEO with 20+ years of experience across agency leadership, performance marketing, and commercial strategy. He writes The Marketing Juice to cut through the noise and share what works.

Frequently Asked Questions