How Retailers Use AI So It Creates Time, Not Kills It
Every retailer we talk to is using AI for something now. The question that actually separates the ones getting value from the ones burning hours is simple: does the tool answer the question, or does it hand you a new job?
Time-killer: asking AI to be the system
One of our fashion clients was playing around with a general AI chatbot to build her own open-to-buy. She was proud of it, right up until she had to start reconciling it by hand. Some of it linked to her point-of-sale system. Some of it didn't. She was the one finding the gaps, the one double-checking whether a number made sense, the one doing the integration work a real planning system does automatically. The plan also missed on freshness, one of the biggest drivers of a profitable fashion business.
That's the trap.
A general AI tool will happily generate a plan. It won't tell you it doesn't know your sell-through history, or how to flow inventory to the peaks and valleys of your season to increase revenue and cut down on markdowns. You find that out later, usually at the worst time, and usually after you've spent hours building something that needed to be rebuilt. Had she run with this plan instead of just playing around with it, she would have ended the season with profits sitting on the racks instead of cash in the bank.
Another version of the same trap: a retailer running her own report output through a chatbot to clean it up, then losing track of whether a number came from her point-of-sale system, her planning platform, or the AI layer that reformatted it. When something looks off, she now has three places to check instead of one. That's not saving time. That's adding a debugging step to every report.
Time-creator: asking AI a direct question about a system that already knows the answer
Compare that to a consultant who asked Ask Indie directly what to prioritize for markdown activity. No reconciling, no cross-checking three systems, and on a major POS system, no uploading a report at all, the data's already connected. One question, one answer, grounded in a plan that was already built correctly. His read: "it's definitely helpful." That's the whole point. The AI isn't doing new work. It's skipping the part where a person has to read six reports to find the thing the reports were already telling them.
We've seen the same pattern hold up live. A consultant pulled up Ask Indie™ in front of a retailer, unrehearsed, and it flagged that her cash margin was up $25,000 versus last year despite lower sales, and that tighter buying discipline was the reason. She checked the number herself. It was right, and it took her seconds to get an answer that would have otherwise meant digging through a spreadsheet.
Retailers are finding the same shortcut in smaller ways, too. One uses a general AI tool to write product descriptions instead of staring at a blank field. Another wanted a way to photograph a vendor invoice and turn it straight into a purchase order, instead of keying it in by hand. Both are the same instinct: let AI eliminate a repetitive task that has one clear right answer, and get back to the parts of the job that actually need judgment.
The dividing line
AI creates time when three things are true:
It's answering a question, not generating a new document you now have to verify. If you have to fact-check the output against your own knowledge of the business, you haven't saved time, you've added a review step.
It's grounded in your actual data, not a guess informed by the internet. A tool that doesn't know your sell-through history will confidently give you an answer anyway. Confidence isn't the same as correctness.
You can trace where the answer came from. If you can't tell whether a number is coming from your POS, your plan, or the AI layer, every discrepancy becomes a small investigation.
AI kills time when it's asked to do the job of a system it was never built to be. Not because the technology is bad. Because a chatbot with no memory of your business and no knowledge of specialty retail will always answer the question you asked, even when it's the wrong question, and you're the one who pays for that in hours later.
The fastest AI workflow isn't the one where you do the most work to get a good answer. It's the one where the answer was already sitting in a system that knew your business before you asked.
——-
Onwards and Upwards,
Marc Weiss
Co-founder and CEO
Management One
|
Pass It Along
Know another retailer who'd want this in their inbox?
We send a read like this regularly including a variety of topics indie retailers like you care about: what we're seeing across the community, what's working, what's not, and what to do about it. Forward this post, or sign up below to get it straight from us.
|