The news
Macy’s is expanding an AI tool that forecasts demand for store replenishment after testing it in a pilot, executives said on the company’s second-quarter earnings call.
The retailer is adding an “AI forecast overlay capability” to its replenishment process to improve in-stocks and “deploy our inventory more efficiently,” COO and CFO Tom Edwards said.
Replenishment is “a critical area centered on having the right product in the right place at the right time,” Edwards said. The tool “started out as a pilot” and “moved through testings to prove itself out” before the wider rollout, he said.
Macy’s ended the quarter with $4.45 billion in inventory, up 2.5% from a year earlier, while comparable sales rose 2.7%.
What changed
The replenishment forecast came out of a set of AI tests Macy’s has been running for more than a year.
The plan: Bold New Chapter, the strategy Macy’s laid out in early 2024, named AI and data capabilities for inventory allocation, forecasting and pricing. The supply chain overhaul in that plan targets $235 million of annual savings by 2026, per Supply Chain Dive.
The list: by March, CEO Tony Spring had been “walked through over 35 different use cases.” In June, Edwards called it “35 tests and pilots that we’re running” across customer-facing work, associate tools and “areas like supply chain.” He said Macy’s was “evaluating several inventory forecasting and management initiatives.”
The cut: Max Magni, chief customer and digital officer, told Fortune he dropped about 50 smaller innovation projects after he arrived, including live-streamed shopping. “We are not buying shiny objects,” Spring told investors.
Ask Macy’s: the Gemini-based shopping assistant ran as an internal dark launch in December, went to a small share of shoppers, then to half of users within a day and all users a week later, per Google Cloud. Revenue per visit was about 4.75 times higher among beta users than non-users. It is now being extended to store colleagues.
Replenishment: the forecast tool is moving “from pilot to broader execution,” Edwards said. Spring said AI work is also under way in “the customer journey, inventory management, and localization.”
Edwards set the bar for the supply chain projects in June. “We’re good stewards of returns and ROI and very cognizant that AI can be costly to implement as well as providing great benefits on the other end,” he said, adding that Macy’s was “already using it in places like China Grove.”
How it works
A store replenishment system holds an order rule for each item in each store: a target quantity, a pack size, a lead time from the distribution center. It reorders when the forecast for that item says stock will fall below the target. The forecast is the one input the rules act on.
An overlay leaves the rules, the lead times and the DC assignments where they are and changes only the demand number. A full replacement moves the rules too, onto one new platform with the forecast.
With an overlay, planners keep the screens and exception reports they already use. The DC picks from the same order file.
China Grove, the 2.5 million-square-foot automated center that opened in North Carolina last October, now handles store replenishment for the Macy’s nameplate alongside online orders. It runs a new warehouse management system. Macy’s said automation “more than doubled productivity” at its first two automated sites.
The competition
Five other retailers and distributors are using AI forecasting or replenishment across their networks.
Dollar General: one RELEX platform for forecasting, replenishment and allocation across more than 21,000 stores, 34 distribution centers and about 18,000 SKUs, per the announcement. The three functions ran in separate systems before.
Lowe’s: allocation for seasonal categories on RELEX in 2024, the full assortment and the whole store and DC network from March, full implementation in early 2027, per Supply Chain Dive. The rollout runs about three years, one function at a time.
UNFI: signed with RELEX in 2023 and finished rolling it into every distribution center by the end of this fiscal year, at one point taking about a dozen sites live in a single week. Fill rate rose two points and on-time delivery five in the fourth quarter, the fourth straight quarter of gains.
Sysco: inventory forecasting is one of four projects behind a $500 million AI savings target through fiscal 2029. The stated goal is fill rates up 50 to 100 basis points on less inventory, with one day of inventory worth about $250 million in cash.
Albertsons: fresh replenishment with Afresh, one department at a time. Produce went chain-wide in 2022, meat and seafood after a two-month pilot in 2023, and all four fresh departments were live at 2,257 stores by last October.
Dollar General, Lowe’s and UNFI moved forecasting and ordering onto one platform. Macy’s describes its tool as a layer on the replenishment process it already runs.
What’s next
Edwards expects supply chain efficiencies to reach gross margin in the second half, weighted to the fourth quarter. China Grove is due to take on store replenishment for the other nameplates within two years of opening.
