The value shift
Part 1: The Upstream Skill Set
In the last edition, I shared that many companies now expect or encourage the use of AI, even though clear guidance remains uneven. That creates pressure on individual marketers to decide what to use AI for, what data to give it, how much to trust it, and when an output is good enough to act on.
I also proposed that AI is changing what makes a commerce marketer valuable.
Not because AI can replace the marketer, but because AI can now produce many of the outputs marketers used to be valued for: retail media plans, shopper data summaries, creative tactics, briefs, measurement, and internal communication. And it can do it almost instantly.
That is why I believe the most valuable commerce, shopper, and trade marketers will build strength in two areas:
Upstream · before the prompt
What question to ask.
Downstream · after the prompt
Which answer to trust.
This edition is about the upstream skill set.
The AI-powered marketer value shift
The old value was
"I can create the comm plan, media strategy, or promotion plan."
The new value is
"I can diagnose the problem, frame the right question, guide AI with context, and turn outputs into commercial performance. And then use AI to create the plan, media strategy, etc."
That shift matters because AI makes it easier to produce answers. But faster answers are not always better answers. AI will confidently generate tactics that seem correct, but can steer the marketer in the wrong direction if the marketer doesn't get off to the right start.
The upstream skill set starts before the prompt.
It changes how marketers use AI: not as an answer machine, but as a thinking partner. One that can help generate options, pressure-test assumptions, organize information, and reveal better paths forward when guided by strong human judgment.
Skill 1
Using curiosity to ask better questions
One simple way to practice this is the Five Whys method: repeatedly asking "why?" to move from the surface symptom to the underlying cause. Innovative companies such as Tesla, Toyota, and Atlassian use this approach.
AI is only as useful as the question it is answering. If you ask AI to solve the surface issue, it will often give you a surface-level response. If you ask it to explore the underlying commercial, shopper, or retailer dynamics, it can become a much more valuable thinking partner.
Five Whys in practice
Scenario: Retail media performance is down at a key retailer.
Weak prompt"Is sponsored search underperforming?"
The curious team asks:
| Why? | Possible answer |
|---|
| Why are sales at the retailer underperforming? | Conversion is down on supported SKUs. |
| Why is conversion down? | Shoppers are clicking but not buying. |
| Why are they not buying? | Price/value perception is weak versus competitors. |
| Why is value perception weak? | The PDP does not clearly communicate pack size, benefit, or use case. |
| Why does that matter? | The shopper does not see enough reason to choose us at the moment of transaction. |
The issue may not be a media problem. It may be a value communication problem showing up inside the media results.
Better prompt"Analyze this situation as a shopper purchase barrier problem, not just a media performance problem. Retail media performance is down, conversion is declining, and shoppers are clicking but not buying. Explore whether the issue is awareness, conversion, availability, value perception, search visibility, content clarity, ratings and reviews, or household penetration. For each possibility, explain what evidence would support or disprove it. And always cite the source."
Skill 2
Giving AI the right context
Context is the ability to give AI the right commercial ingredients. AI cannot know the business situation, retailer priorities, shopper behavior, competitive pressure, or constraints unless you provide them. If not, this is what people mean when they say, "garbage in, garbage out."
Without context, AI defaults to generic best practices. It may produce a polished retail media plan, but that plan may not reflect the retailer's priorities, shopper barriers, category dynamics, the brand's constraints, or the real growth objective.
The better the input, the better the strategic output — and the better the performance.
The commercial context framework
Before asking AI for a recommendation, frame the situation across these context areas:
| Context area | Question |
|---|
| Business objective | Are we driving penetration, conversion, repeat, trade-up, basket size, or retailer support? |
| Current challenge | What is the business situation that is keeping the brand from success? |
| Retailer priority | What does the retailer care about right now? |
| Shopper behavior | What is the shopper doing or not doing? |
| Shopper insight / Purchase barrier | Why is the shopper doing or not doing? What is preventing growth? |
| Category dynamics | Is the category growing, declining, premiumizing, trading down, or shifting channels? |
| Competitive pressure | Who are we losing to and why? |
| Channel reality | Is this digital shelf, retail media, in-store, omnichannel, or sell-in? |
| Constraints | Budget, timing, claims, inventory, promo calendar, retailer requirements. |
| Measurement | What would success look like? |
Weak prompt"Create a retail media plan for Target for back-to-school season."
Better prompt"We need to grow household penetration at Target for Brand X. Back-to-school season is a critical period for the brand and the category. The category is growing, but our brand is losing share among value-seeking shoppers. Our suspected barriers are weak search visibility and unclear value messaging. Recommend three retail media approaches to use during back-to-school season, explain the assumptions behind each, and identify what data we need before acting."
Skill 3
Problem Framing
Problem framing is the ability to define the right problem before generating answers. In commerce or shopper marketing, that often means diagnosing the real purchase barrier — what is stopping the shopper from buying your product along the path to purchase.
If you ask for a plan without defining the purchase barrier, AI may produce a list of plausible tactics that sound reasonable but do not address the real reason shoppers are not buying.
Imagine a brand is losing share at a key retailer.
Weak prompt"What kind of offer and marketing plan do we need to boost sales?"
But that does not take into account the underlying current shopper mindset. The better question is: "What is stopping shoppers from choosing us?"
The answer could be any number of barriers:
Lack of awareness
Low relevance
Not available or easy to buy
Current habit works, no reason to switch
Unconvincing proposition
Confusion about which version to buy
Not knowing how to use the product
Poor value proposition
Each barrier would lead to a different strategy:
If awareness
The answer may be reach and visibility before they get to the store.
If confusion
The answer may be education on how to use our product.
If value perception
The answer may be clearer benefit communication, pack-price framing, or stronger proof. It might have nothing to do with the price.
If habit
The answer may be disruption, trial, or a switching trigger.
Better prompt"Help me frame the right shopper marketing problem before recommending tactics. Our brand is losing share at a key retailer, but we do not yet know whether the issue is awareness, relevance, availability, habit, value perception, product confusion, usage understanding, or an unconvincing proposition. For each possible purchase barrier, explain what shopper behavior we would expect to see, what data would help confirm or disprove it, and what type of strategy would be most appropriate if that barrier is true."
