How Macy's uses data and AI to reshape its omnichannel business

Macy's describes its AI work as a set of problem-led use cases tied to a broader omnichannel strategy, not as a single platform. The company reported more than 35 internal use cases in March 2026, launched a customer-facing assistant, and set out governance and risk language in its annual report. Quantified results have not been published, so the evidence is about direction and discipline more than outcomes.

What has Macy's said about its AI strategy?

In March 2026, as it reported fiscal 2025 results, Macy's said it had identified more than 35 internal AI use cases across supply chain, merchandising, marketing, call centres and customer-facing and omnichannel functions. Chief executive Tony Spring said the company was not buying shiny objects but solving problems. Chief financial officer Tom Edwards said Macy's had significant, underappreciated capabilities in data science, AI and technology.

In June 2026, on the first-quarter call, Edwards said 35 AI pilots and tests were underway across customer service, associate productivity and supply chain. Management said the aim was to support the broader strategy rather than to pursue technology for its own sake.

How does AI support Macy's customers and stores?

Spring said Ask Macy's, a new AI-powered shopping assistant, was producing higher conversion among users in its early stages. No conversion figures, user numbers or launch scale were provided. Macy's annual report says it is applying AI to how customers discover, shop and engage with its brands, through shopping assistance, agentic commerce and hyper-personalised experiences.

The store programme is separate from AI but forms part of the same omnichannel logic. The reimagined store initiative grew from a test at 50 locations to 200 stores, covering about 60 percent of the go-forward fleet and about 75 percent of go-forward store sales. Local leaders have more autonomy over staffing, events and assortment. The company said Reimagine 200 locations had comparable sales growth of 2.4 percent in the first quarter of fiscal 2026.

Read also

How Burberry uses data and AI across clienteling, design and the supply chain

What data underpins the omnichannel strategy?

Macy's operated 665 stores, websites and mobile apps across three brands and served nearly 40 million active consumers, according to its annual report. Its loyalty programmes (Star Rewards, Loyallist and BlueRewards) are described as not standalone marketing tools but as drivers of retention, engagement and omnichannel behaviour. Macy's Media Network, its retail media business, grew revenue 12.5 percent in the fourth quarter, according to a PYMNTS report of March 2026.

The annual report names data and analytics, supply chain and omnichannel capabilities among its investment priorities. It also notes that the ability to collect, use and protect customer data is subject to privacy laws, with twenty US states having enacted privacy statutes as of June 2025.

What guardrails and risks does Macy's describe?

Few retailers write about AI risk as specifically. The annual report says AI is covered in mandatory compliance training, that certain tools require internal review and approval, and that systems are monitored. It also states that even with careful governance, AI can produce incorrect output, and lists confidential data exposure, algorithmic bias and intellectual property infringement among the risks. It adds that agentic AI may be hard to govern and could learn to circumvent guardrails.

• Use cases are screened and approved before use.
• Staff receive mandatory compliance training.
• The company states that incorrect output remains a risk.

What does Macy's approach suggest about organising AI work?

Management's language, that the company is solving problems and not buying shiny objects, points to a portfolio approach. Rather than one large programme, there are dozens of small use cases in different functions, each tied to a business problem. That has advantages: low individual risk, quick learning and local ownership. It also has costs, including duplicated effort, uneven data quality and difficulty comparing results.

The annual report's governance language shows how a large listed retailer handles this. Tools are reviewed and approved, staff are trained, and systems are monitored. For a customer-facing assistant such as Ask Macy's, the same incorrect-output risk applies that the company describes in general terms. Wrong product or availability information in a shopping assistant is a customer service and legal exposure, not only an inconvenience.

• Pilots need clear stop and scale criteria, or they accumulate without decisions.
• Customer-facing tools need tested answers on price, stock and returns policy.
• Data privacy rules limit what customer data may be used, and the report notes twenty US states with privacy laws as of June 2025.
• Retail media and loyalty data raise separate consent and use questions.

For European and Swiss retailers, the same pattern would meet different rules, including the GDPR and the EU AI Act, which this case does not address.

One further point concerns the supply chain. Macy's describes the China Grove distribution centre as automated and as the base for AI-driven improvements, with early results including faster delivery times and customer-visible fulfilment estimates. Physical automation and AI are separate things, and the sources do not say which improvements come from which. Readers evaluating similar claims should ask whether the benefit comes from machinery, from software that schedules it, or from forecasting models, since each has different costs and risks.

The 10-K also notes that the new fulfilment and replenishment centre is intended to improve inventory productivity and delivery accuracy. Those are measurable operational goals, and Macy's could report against them in future filings.

Read also

How Levi Strauss & Co. uses AI: the data programme and its lessons

What is not yet proven?

Macy's results improved in fiscal 2025 and early 2026, with positive comparable sales, but the sources attribute those gains mainly to store reimagination, brand additions and Bloomingdale's performance. None of the reviewed sources quantifies how much AI contributed. The case therefore shows an organised, governed approach, with the financial effect still to be demonstrated.