In this article, you'll learn:
Why early AI merchandising tools aren’t delivering the impact retailers expected
The operational and data barriers that keep AI pilots from scaling
How to integrate AI into real merchandising workflows
AI is ready to transform merchandising, but are retailers ready to be transformed? In a recent survey done by McKinsey & Company, 71% of merchants said AI merchandising tools have had limited to no effect on their business.
At the same time, the potential benefits of AI merchandising tools look promising: smarter assortments, sharper pricing, better promotions, and hours of manual work handed off to machines that can work around the clock.
So, what’s making retailers slow to adopt AI or see results? The problem isn't that AI doesn’t offer the right kind of help, but that most retailers haven’t integrated AI in a way that fits how their businesses actually run. Across industries, AI programs and pilots have been tripped up at the implementation stage. A study released by MIT in 2025 found that 95% of generative AI pilots fail.
To bridge that gap, retailers should stop thinking about AI as a standalone tool and start treating it as an embedded capability to work alongside merchants. In the rest of this article, we’ll look at what’s really holding AI merchandising back and practical steps retailers can take to move from failed pilots to scalable impact.
What is AI Merchandising?
AI merchandising is the use of artificial intelligence to improve how retailers plan, price, promote, and place products across channels. AI merchandising tools are capable of analyzing huge volumes of data (sales, inventory, shopper behavior, digital engagement, supplier performance, and more) to recommend decisions such as:
What to carry
How to price
How to promote
Where to place
When to replenish
More advanced, agentic AI systems can go a step further. Instead of simply generating insights, they can also take actions within predefined limits. For example, an AI merchandising agent might:
Flag underperforming promotions and propose a new mix or price
Detect an out-of-stock risk and trigger a replenishment change
Identify a vendor performance issue and surface negotiation talking points
In an ideal state, AI serves as a digital partner for merchants and planners, taking on tedious, time‑consuming tasks so they can spend more of their time on work that requires nuance, intuition, and strategic judgment.
Barriers to Successful Implementation
If the value of AI is so compelling, why are 71% of merchants saying AI merchandising tools have had limited to no effect? While many things can limit impact, there’s often a primary issue: AI has been added as an afterthought, not built into workflows.
Let’s take a look at a few areas that can limit impact when implementing AI into merchandising:
1. Inconsistent Results
AI is only as good as the data and systems you give it. If your business is dealing with:
Conflicting SKU files between systems
Incomplete or incorrect hierarchy and attribute data
Gaps in store-level inventory accuracy
Fragmented pricing and promotion history
Layering AI on top of that can easily get confident‑sounding but wrong answers.
In addition, because of AI’s tendency to hallucinate, retailers might trust AI to provide additional context, but trusting it enough to change prices, shift inventory, or rework assortments in real time is a much higher bar. Until the underlying data and governance are reliable, full adoption is challenging.
2. Implementation Issues
AI pilot and testing issues, not-so-lovingly dubbed “pilot purgatory,” refer to AI initiatives that get stuck as small, isolated pilots.
These pilots may show promising results in a slide deck, but they don’t connect back to the tools merchants and planners use every day. The gap between a nice demo and operational reality is where most AI projects die.
Without tight integration into merchandising systems, supply chain platforms, and collaborative workflows, even the best models don’t translate into better outcomes on the shelf.
3. Challenging Mindset Shift for Employees
For merchants, planners, and supply chain professionals, AI can fundamentally change their jobs. For many teams, that could like shifting from:
Time spent building reports ➡️ Time spent interpreting and acting on recommendations
“I know this category because I’ve been in it for 15 years” ➡️ “I know how to ask the right questions and pressure‑test AI output”
Manually reacting to issues after they show up in the numbers ➡️ Proactively steering the business based on AI‑flagged signals and scenarios
That’s a big shift. If teams feel AI is being “done to them” instead of built with them, the natural reaction is skepticism. Without training, expectation setting, clear role definitions, and visible benefits, adoption may stay surface‑level.
What Successful AI Merchandising Really Looks Like
Even with those hurdles, the retailers who get AI merchandising right are seeing meaningful gains. Walmart’s Wally is a great example. Walmart built a genAI-powered assistant that is embedded directly into merchants’ day-to-day workflows.
Wally sits on top of Walmart’s proprietary merchandising data and lets merchants instantly pull and synthesize complex insights, diagnose root causes, and run advanced calculations. By automating the reporting and analysis work that once ate up hours, Wally lets merchants focus on strategy, creativity, and vendor relationships.
AI is most powerful when embedded into existing workflows and business functions, not simply when layered on in the name of “productivity.” If your business is ready to take the plunge, here's some ways to set it up for real impact:
1. Start with a Clear, Measurable Use Case
Instead of rolling out “AI for everything,” identify a small number of high‑value, tightly scoped problems, such as:
Reducing out‑of‑stocks in key categories
Localizing assortments for a priority region
Freeing up time spent on report building for merchants
Each use case should have an owner, success metrics, and a defined decision flow (who sees what, when, and how they act on it). That focus makes it much easier to measure value and build confidence.
2. Build on a Solid Data Foundation
AI trained on broken data amplifies, rather than fixes, problems. Before implementing AI into your workflows, it’s important to:
Standardize product, store, and hierarchy data
Improve inventory accuracy in critical locations
Connect core systems (merchandising, supply chain, pricing, promotion, e‑commerce) enough that AI can “see” a full picture
You don’t have to achieve perfection before you start, but you do need to have trusted data to build upon. It’s also critical to identify where AI should not be making decisions without human review.
3. Embed AI into Everyday Workflows
In successful implementations, AI is not a separate portal that merchants log into “when they have time.” It’s woven into:
The same planning and replenishment tools teams already use
Dashboards and signal briefs that automatically surface the most urgent issues
Existing meeting rhythms, like line reviews, promo planning, vendor negotiations, daily huddles
When AI is available in the systems where work gets done and not as an extra step, adoption rises naturally.
4. Set Clear Guardrails and Governance
AI can recommend and even execute actions, but only within well‑defined boundaries. To safely adopt AI, consider:
Defining what AI can do automatically
Defining what requires human approval
Monitoring decisions and outcomes with audit trails and alerts
5. Invest in People and Platforms
To set your team up for success, pair AI deployment with:
Training on how the models work (in plain language), including the strengths and weaknesses of the model
Clear expectations for roles (for example, “category managers as orchestrators and decision‑makers, not report builders”)
Guardrails around what AI is allowed to decide vs. what requires human approval, as well as data access, privacy, and brand / pricing policies
Summary
AI merchandising isn’t falling short because the technology isn’t ready. It’s falling short because too many retailers are treating it as a side project instead of a core capability. When AI is bolted onto messy data, disconnected tools, and unchanged roles, it’s no surprise that 71% of merchants are seeing limited impact.
The retailers who are breaking out of that pattern are doing a few things differently: they start with clear use cases, fix their data foundations, embed AI into everyday workflows, and invest in helping people work with the technology instead of around it.
Ready to Embed AI Across Your Supply Chain? Meet MAX.
AI tools are most powerful when they’re embedded throughout the supply chain, not simply layered on as an afterthought. MAX, the SPS Commerce AI solution, brings this concept to life by embedding intelligence directly into day-to-day workflows. Built on a network of more than 300,000 trading connections and billions of transactions, MAX helps teams identify potential issues early and make faster, more informed decisions across their supply chain.
