How AI Can Increase ROI in Marketing Research
For a lot of marketers, the conversation around AI still starts in the wrong place.
It starts with speed.
How fast can it write a draft? How quickly can it summarize a transcript? How many subject lines can it generate in a few seconds? How much time can it save?
Those are reasonable questions but they are not the most important ones.
The more important question is whether AI can help marketers make better decisions, not just faster ones, better ones. Decisions grounded in a deeper understanding of customers, stronger market insight, and clearer value propositions that have a real chance of improving ROI.
That is where AI becomes much more interesting.
Over the past year, more serious conversations about AI in marketing have started moving in that direction. One recent Harvard Business Review (HBR) article on generative AI and market research makes the case that AI is not only improving speed and scale but reshaping how companies gather insight, synthesize evidence, and even simulate customer understanding in new ways. Another article centered on Dyani Marvel’s work at Wondr Nation shows how this is already becoming practical inside modern marketing teams, where AI is being used to improve efficiency while still depending on human strategy, judgment, and refinement. Another broader article on the future of AI in marketing pushes the conversation further, showing how predictive analytics, personalization, and automation are steadily becoming part of the marketer’s everyday toolkit.
Taken together, they point toward something bigger than simple productivity gains.
AI is becoming most valuable when it strengthens marketing research.
From my perspective, the best way to increase the likelihood that AI projects create meaningful impact on revenue is to combine AI with a structured frame of thinking like Jobs to Be Done theory. Then use that structure to analyze and cross-reference quantitative and qualitative data from strong, relevant sources, in order to craft better value propositions.
That is where the real ROI begins.
I’ve been practicing this approach for years and first described an early version of my process in the article “How I use ChatGPT to Uncover Jobs to be Done”. In that piece, I outlined a very early stage of my method for using AI to develop value propositions, an approach that has only become more powerful as AI capabilities continue to evolve. At the time, the concept was met with criticism, skepticism and ridicule from basically everyone (colleagues, family, friends, you name it) but fortunately I never stopped developing my methodology.
(Mentioned articles are linked at the bottom of the page)
AI Does not Create Value Just Because it Saves Time
There is no question that AI can make marketing work more efficient.
It can synthesize long interview transcripts, summarize documents, help write reports, generate content ideas, surface patterns in data, speed up ideation, and reduce friction in repetitive tasks. In the HBR article “How Gen AI Is Transforming Market Research,” one of the clearest findings was that many researchers are already applying those use cases. Most notably for transcript synthesis, data analysis, and reporting. In that sense, AI is already proving its value.
But efficiency is not the same thing as impact.
A team thats is inexperienced with AI can get faster without becoming smarter. They can produce more content but without improving the quality of its strategy. As well as automate more workflows while still relying on weak assumptions about what the customer actually values.
That is why some AI projects sound impressive but fail to produce noticeable commercial results.
The issue usually is not the tool itself. The issue is that the thinking behind the work is still underdeveloped.
AI is powerful, but if it is applied to remedial strategy, it tends to scale remedial strategy (go figure, right?). It’s not some sort of all knowing oracle that knows the context of your questions or requests without you telling it. Sorry to dissapoint you, but you do still need to use your brain to think if you want AI output with true depth and context.
The Real Opportunity is Better Market Understanding
What makes AI especially promising in marketing research is not just that it can process more information, but that it can help marketers work across more kinds of information at once.
That is a much bigger shift.
Traditionally, insight work has often been fragmented. Quantitative data sits in one place while qualitative observations sit somewhere else. Customer interviews, CRM trends, survey responses, reviews, search behavior, competitor positioning, sales notes, and campaign metrics all exist, but they are not always interpreted together. One dataset tells you what happened. Another gives clues about why. Another shows what customers say they want. Another reveals how they actually behave.
The gap between those things is where a lot of bad marketing decisions are made.
This is why AI has so much potential in research. It can help marketers synthesize large volumes of information, compare patterns across sources, identify recurring themes, and surface tensions that would otherwise take much longer to recognize. This was one of the more significant use cases discussed in the HBR article, particularly in its explanation of how generative AI can support existing research practices, fill insight gaps where companies once relied on intuition, and even create new research environments through synthetic data and digital twins. For me, this was the most significant point because it validates the kind of approach I have been practicing and developing since early 2023.
I find this concept valuable because marketing ROI improves when decisions are grounded in stronger evidence.
Why Structure Matters More Than Prompts
This is also where a lot of teams go wrong with AI.
They assume the value is in asking the model to generate answers.
I think the bigger value is in giving the model a strong framework for analysis.
Without proper structure, AI tends to produce plausible but generic output. It can summarize endlessly, but summary alone is not strategy. It can generate recommendations, but recommendations without a disciplined way of evaluating what matters usually leads to weak positioning and forgettable messaging. The same kind of generic results you can expect from a human that lacks education or experience.
That is why I find Jobs to Be Done theory so useful.
JTBD theory gives research a center of gravity.
Instead of asking vague questions about customer preferences, it forces a more useful set of questions. What progress is the customer trying to make? What struggle is pushing them to seek change? What anxieties are slowing action? What desired outcome defines success from their point of view? What alternatives are they comparing? What social, emotional, and functional dimensions are shaping their decision?
Once AI is used inside that kind of structure, the quality of the analysis improves.
It becomes less of a novelty tool and more of a disciplined research assistant.
In practical terms, JTBD helps filter signal from noise. It makes it easier to cross-reference what customers say, what they do, where friction appears, and which patterns actually connect to commercial opportunity. That is a much more serious and valuable use of AI than simply asking it to draft a campaign or write an instagram caption.
Better Value Propositions Come From Better Interpretation
The strongest commercial use of AI in marketing research is not content generation.
It is helping marketers craft better value propositions.
The concept is consistent through all three articles, even though they describe it in different ways. In “The Role of AI in Marketing: Insights From a Wondr Nation VP,” the discussion emphasizes that AI becomes truly useful when marketers apply strategic thinking, customer insight, and brand context to its outputs. In “AI Will Shape the Future of Marketing,” the focus shifts toward how predictive analytics, hyper-personalization, and real-time insights are making marketing more relevant and data-driven. In the Harvard Business Review article “How Gen AI Is Transforming Market Research,” the authors discuss how AI is becoming increasingly valuable for understanding customers, modeling scenarios, and supporting decisions that were once shaped by incomplete evidence.
Put differently, AI becomes much more valuable when it helps a company say something more meaningful to the market.
A value proposition improves when it is grounded in a stronger understanding of what customers are trying to get done, what tradeoffs they face, what outcomes they care about, and what kind of progress they are actually hiring a product or service to deliver.
That is why structured analysis matters so much.
AI helps you cross-reference behavioral data, customer language, reviews, interviews, search patterns, and performance metrics through a JTBD lens, you are much more likely to arrive at a value proposition that resonates. Precisely why I have been doing it for years.
Because when a value proposition resonates more clearly, revenue impact becomes more likely.
The Best AI Projects Improve Judgment, not Just Workflow
There is a reason so many discussions about AI still feel unsatisfying.
They are too focused on the completion of tasks.
They talk about drafts, outputs, productivity, and automation, but not enough about judgment.
What the most thoughtful writing on this topic is starting to suggest is that AI’s long-term value in marketing will come from improving the quality of strategic judgment. Not replacing it. Improving it.
That means helping marketers test assumptions earlier, identify patterns faster, compare more evidence, pressure-test messaging, surface hidden tensions, explore competitor positioning, and analyze customer language at scale. In order to build sharper hypotheses before money is spent on creative, media, or execution.
That is a different mindset from simply using AI to produce more.
It is closer to using AI to think more rigorously.
That, in my view, is where the greatest likelihood of ROI lives.
Why Human Thinking Still Matters
None of this removes the need for human direction.
If anything, it raises the importance of it.
All three articles, in one way or another, point to the same concerns: bias, hallucinations, weak representativeness, privacy issues, copyright concerns, and organizational misuse. Those risks are real but there is another one worth adding to the list.
AI can create the illusion of clarity.
It can make weak assumptions sound sophisticated. It can produce polished language that feels strategic without actually being tied to reality. It can make it easier for teams to confuse output with understanding.
That is why human oversight matters so much.
Not just to fact-check the model, but to guide the entire research process, choose the right inputs, decide which sources are credible, judge which patterns matter, challenge easy conclusions, and recognize when a result is convenient rather than true.
AI can help marketers move faster through complexity.
However, it still takes human strategic discipline to make sure that speed is pointed in the right direction.
A Better way to Think About AI ROI in Marketing Research
If the goal is to increase ROI with AI in marketing research, the starting point should not be content production.
It should be insight quality.
Start with strong, relevant sources. Pull from both quantitative and qualitative data. Use AI to synthesize, compare, and surface patterns. But do not stop there. Apply a structured frame of thinking like Jobs to Be Done, Blue Ocean, Sigma Six or Kaizen, so that the analysis stays focused on what actually drives customer decisions. Then use those findings to build sharper value propositions, better messaging, and make more informed strategic choices.
That is a more demanding use of AI than just using it to write copy but it is also far more likely to result in meaningful revenue impact.
Because in the end, AI does not improve marketing ROI simply by helping us make more content.
AI improves ROI when it helps us understand more, interpret more wisely, and communicate value more clearly to the market.
That’s the kind of AI work that actually matters.
Thats it for now and thank you for making this far!
If you have any questions or comments, feel free to reach out to me on any social platform. I am not hard to find.
Oh, and remember to always unapologetically be yourself, never give up and have fun doing it.
Until next time.
Mentioned Articles;
