AI Is Only as Smart as Your Data
Merchandising Solves This | Week 7
AI is changing the game in retail and consumer goods. Faster forecasting. Smarter inventory decisions. Demand signals that used to take weeks to surface are now available in real time. The brands and retailers who figure out how to use AI well are going to have a meaningful competitive advantage over the ones who don't.
But here is the part nobody talks about enough.
AI does not create insights out of thin air. It surfaces patterns from the data you give it. And if that data is incomplete, inconsistent, or siloed in systems that don't talk to each other, the AI does not fail quietly. It produces confident-sounding wrong answers. Which is more dangerous than no answer at all.
The question worth asking before you invest in any AI tool, before you trust any AI-generated forecast or recommendation, is the same question that has always mattered in merchandising.
How good is your data?
What the Research Is Telling Us
Salesforce recently published its State of Data and Analytics report, Retail and Consumer Goods Edition, surveying over 1,200 technical and business leaders across the global retail and consumer goods sector. The findings make the case for AI clearly while being equally clear about what stands between most companies and its full potential.
94 percent of RCG business leaders say the rise of AI makes it more important than ever to be data-driven. 91 percent of technical leaders say a strong data foundation is the most critical factor in successful AI deployment. And 62 percent of technical leaders who have already implemented AI say they have wasted significant resources training their models on unreliable data.
That last number is the important one. More than half of companies already using AI have burned real time and real money because the data feeding their models was not trustworthy. The tool was not the problem. The foundation was.
And the foundation has a real problem. Only 68 percent of company data is considered trustworthy by the organizations that own it. Nearly one in three data points being used to make decisions is considered questionable by their own standards. 37 percent of business leaders admit they have given a range instead of a specific number because the data was unreliable or inaccessible. 30 percent say they have made gut-based decisions because the data they needed simply was not there.
AI cannot fix any of that. It inherits it.
Why This Matters More in Merchandising Than Almost Anywhere Else
The retail and consumer goods industry is full of decisions that depend on data being right. Not approximately right. Not directionally right. Actually right, at the model level, the channel level, the colorway and size level, the full-price versus markdown level.
Merchandising sits at the center of all of it.
The ABC analysis that tells you which products deserve more investment next season is only as reliable as the sell-through data underneath it. A demand forecasting model trained on incomplete channel data does not produce better forecasts. It produces confident-sounding wrong ones that get built into the buy plan before anyone notices the error. An AI tool asked to identify your top-performing products will surface whatever the data says is performing well. If that data is missing channels, misattributing sales, or not accounting for markdown-driven volume, AI will amplify those blind spots rather than correct them.
The report found that 92 percent of technical leaders say they need faster insights to meet business goals, and 73 percent say their most valuable insights are trapped in unstructured data they cannot access. That is a critical gap. Not because AI cannot help close it, but because AI needs clean, connected data to do so.
When the data is good, AI becomes a genuine force multiplier for merchandising. It can surface demand signals faster than any team could manually. It can flag markdown risk before the season is halfway over. It can identify which colorways are losing momentum in real time rather than at the end-of-season debrief. It can make the ABC analysis dynamic rather than seasonal.
That is the version of AI worth building toward. And it starts with the data.
The Foundation Comes First
The brands and retailers who will get the most out of AI are not necessarily the ones who invest in the most sophisticated tools. They are the ones who invest in the most trustworthy data foundation first.
That means sell-through data that is complete and updated frequently enough to actually inform decisions in season rather than after the fact. It means margin data that reflects what a product actually contributed after the season ended, not just what the sticker price implied. It means inventory data that reflects reality across every channel, not just what the system says at headquarters.
None of that requires AI to build. It requires discipline, the right questions, and a willingness to audit what you already have before adding anything new on top of it.
The report is clear on this point. 91 percent of technical leaders agreed that AI's outputs are only as good as its data inputs. The tool is not the limiting factor. The data underneath it is.
Where to Start
Pull last season's sell-through. Is it complete? Is it broken down by model, channel, colorway and size? Does it reflect what actually sold at full price versus what moved on markdown? Does it match across your systems?
If the answer to any of those questions is no, that is where to start. Not with the AI tool. Not with the forecast. With the data itself.
The brands that get this right, that build a clean, connected, trustworthy data foundation, are the ones that will use AI to pull ahead. And the ones that skip the foundation step will keep getting confident-sounding wrong answers no matter how advanced the tool they plug in.
AI is genuinely exciting. It is changing what is possible in merchandising and across the entire retail and consumer goods value chain. But it is only as smart as the data you give it.
Get the data right. Then AI will follow.
Source: Salesforce State of Data and Analytics, Retail and Consumer Goods Edition, 2025. Based on surveys conducted with 1,236 retail and consumer goods leaders across 17 countries.
The Outdoor Merchant is a product merchandising consultancy specializing in outdoor, cycling, and snowsport industries. Each week in this series, we explore a real business problem that smart merchandising was built to solve.
Follow along for Week 8, and reach out to sarah@theoutdoormerchant.com if any of this is hitting close to home.