Retailers Are Handing AI Agents the Merchant Job

Retailers are starting to hand parts of a merchant’s job to AI, from deciding which products to order to bargaining with suppliers on price and payment terms. Walmart, Target, and a wave of European grocers are already running AI agents in production, and the technology is moving up the merchandising stack from data entry to autonomous negotiation.

The pitch to retail companies is simple: instead of buyers wading through data and email threads with suppliers, AI agents can take on the research, the price checks, and in some cases the deal itself. The first results from named deployments show where the technology is winning, where it stalls, and which parts of the merchant’s day are still, for now, off limits to the bots.

What Walmart and Target Are Already Running

Walmart introduced its own internal tool in March 2025. The retailer developed Wally, a GenAI-powered assistant built on Walmart’s proprietary data, to automate data entry and analysis, root-cause identification, how-to support for operational questions, and advanced calculations. The goal, per Walmart’s own announcement, is to act as “a productivity multiplier for our merchants” and to “ultimately” run “autonomously on the merchant’s behalf within configurable guardrails.”

Walmart’s roadmap, written in the same release, is the part that matters for the merchant role: the retailer is working to push Wally from answering questions to executing “tactical actions necessary to bring their strategy to life.” Today, the tool is still framed as an assistant that lets merchants skip the manual reporting so they can focus on “strategic, creative and innovative activities.”

Target went a different direction with its December 2025 launch of Trend Brain, a GenAI-powered trend intelligence platform built specifically for its owned brands. According to Target, the platform “can process massive volumes of visual and written trend data” and shrinks work “that used to take weeks” into “hours.” An early win, the company said, came when Trend Brain flagged greens, berries and blues as must-have colors for fall; those shades then made it into the fall collection.

Guitar Center, the musical instrument retailer, is in the early stages of layering AI on top of both customer and employee workflows. Adolfo Rodriguez, the company’s EVP and chief technology and information officer, told IT Brew in March 2025 that an in-store large language model trained on musical data would soon let shoppers scan a QR code and ask for the rig settings needed to approximate a guitar tone. Behind the counter, Rodriguez said, a GenAI feature already helps with buyback transactions by turning a brand, model, and year into a full product description, with photo backgrounds removed automatically. “AI is a tool. It is not perfect, right? It is an aid that will allow you to be more efficient, to be better. But it does not replace humans,” Rodriguez told the publication.

Agents That Talk to Suppliers

The deeper automation is in vendor negotiations, where AI agents are now sitting on the retailer’s side of the table and running entire conversations.

Walmart’s negotiation agent, Pactum, has been working across the retailer’s supplier base. Pactum’s published case study says Walmart achieved a 3% average gain across negotiations while extending payment terms by an average of 35 days. The Pactum case study also reports that 68% of suppliers engaged through the platform closed an agreement and that 83% of suppliers, surveyed after the fact, described the system as easy to use.

Pactum is not just a Walmart tool. The company’s client list includes Otto Group, Rolls-Royce, Honeywell, Coupang, and Novartis, across procurement, packaging, and pharmaceuticals. The use cases on Pactum’s own site, including “Campaigns Agents” that “conduct large-scale autonomous negotiations of supplier level terms across your entire supplier base,” read like a job description for a junior buyer’s entire week.

Gain, a Tel Aviv-based startup, emerged from stealth on September 29, 2025 with $12 million in seed funding and a stack of AI “Employees” branded with human names. Natalie handles category strategy and negotiation, while Bob runs tactical buying, contract execution, and exception handling. Tempo Beer Industries, Israel’s largest brewer, is one of the named pilots; CEO Daniel Beer said his team is “undergoing testing with Natalie and Bob, and it’s fascinating to see how AI Employees become part of the team.” Gain’s CEO Michael Gabay, who previously founded the retail automation platform Trigo, framed the pitch in opposition to layoffs: “Our AI Employees don’t replace jobs. They support teams by tackling work that otherwise goes undone and delivering real value.”

Where European Retailers Are Pushing Further

European retailers are adopting the technology faster, in part because labor pools are tighter and labor laws are stricter. “There are, rightfully so, in North America concerns over whether or not this could replace jobs in the future,” said Jason Busch, Gain’s co-founder and head of strategy. “We don’t position it that way. We say: There’s so much work that delivers huge value, which is not getting done today, and that’s what we pursue.”

Czech company Duvo.ai, founded by Rohlik Group founder Tomáš Čupr, is taking on the email back-and-forth itself. The company’s AI reads supplier emails on price changes and price proposals, analyzes commodity and labor costs, packaging, transport, and the supplier’s acceptable margin, and replies on the retailer’s behalf to accept or reject proposals at the SKU level. Duvo.ai’s named clients include Holland & Barrett in the U.K. and Heureka in the Czech Republic. Čupr told Modern Retail that the human-in-the-loop discipline is non-negotiable: “If a supplier is angry or rejecting too many of the offers, Duvo.ai will notify the retailer and have them respond.” The system is also deliberately not fully autonomous. “You cannot expect this to be fully autonomous,” Čupr said.

Zenline AI, which builds category management agents, currently works with European retailers including Conrad, Otto, and Futterhaus, in categories ranging from beauty and groceries to electronics and fashion. Founder and CEO Arber Sejdiji said the company’s agents “do the grunt work in the background.” They read external and internal data, clean it, and surface “concrete, margin-tied recommendations for the next actions.” Zenline has said it plans to begin offering its agents to U.S. retailers this fall.

Software That Lifts the Merchant, Not Replaces Them

Older retail software vendors are also working the same problem, with a softer framing. Relex Solutions, a roughly 20-year-old retail software provider, has begun leveraging AI agents to automate weekly merchant meetings, planograms, price updates, and promotions. The company’s named retail clients include Whole Foods and Sprouts Farmers Market. Relex’s director of solution principals, Patrick O’Mara, told Modern Retail that “the merchant’s job is even elevated now.” The merchant, he said, is now focusing “much more on the strategic aspects of the role and the vendor negotiations; working with operations to ensure that the plans the merchant has placed are going into effect; and working with the finance team to ensure that the merchandise financial planning angle is fully covered.”

Crisp, a retail data analytics platform based in Arkansas, takes a similar approach. General manager and EVP Dirk Herdes said the platform helps grocers manage “hundreds if not thousands of items” in fresh produce, deciding weekly how much to order to cut waste and lift turnover. Crisp’s stated customers include small-format stores and distributors in the U.S. and regional grocers in Northern Europe, in grocery and center-store categories such as health and beauty. “It is really democratizing the use of information, so operators can see exactly what is happening,” Herdes said. “They do not have to wait on a report or some internal team to design the analysis; they can take advantage of the data there and make decisions on the fly.”

The Human Is Still in the Loop, by Design

Every vendor in this space is repeating the same line: a human has to approve, escalate, or take over the awkward cases. Busch’s framing was the most explicit: “Our guidance is, you should always have a human in the loop until trust is built, and then the human should still be in the loop, but you can figure out at what cadence you want the check-in to be.”

Guitar Center’s Rodriguez made the same point from the buyer’s seat. “When we think about category planning and merchandising, it is personal, local and human,” a Walmart spokesperson told Modern Retail when the retailer was asked about its internal tools. Guitar Center’s own formulation of the trade-off was identical. “The goal is not simply to automate existing processes, but also to help our teams make better decisions and create more value for customers and associates,” Rodriguez said.

That framing also explains why the technology is showing up first in the parts of the job nobody wants, like back-office data entry, supplier emails at scale, and SKU-level price proposal triage, and not in the parts of the job that decide what a store actually stocks. The agents are doing the work humans did not have time to do. They are not, yet, picking the assortment.

For a related read on how the AI labor question is playing out beyond retail, see this Gallup finding on AI use and tech layoff risk.

Frequently Asked Questions

What AI tools are retailers using today?

Walmart runs Wally, an internal GenAI assistant for merchants announced in March 2025. Target runs Trend Brain, a GenAI trend intelligence platform for owned brands launched in December 2025. Outside the retailers’ own walls, vendors such as Pactum, Gain, Duvo.ai, Zenline AI, Relex Solutions, and Crisp sell AI agents to retailers for negotiation, category management, and demand forecasting.

Is AI replacing retail merchants?

Not according to the vendors and retailers named in public reporting. Walmart, Guitar Center, Gain, and Duvo.ai all describe the technology as a productivity tool, with humans retained for strategy, escalation, and final approval on contentious supplier deals. Gain’s CEO has said the company does not position the technology as a job replacement.

Which retailers are using AI to negotiate with suppliers?

Walmart is the named example, using Pactum’s negotiation agent to extend payment terms across thousands of suppliers. Pactum’s published case study says Walmart achieved a 3% average gain across negotiations and a 35-day average extension on payment terms. Gain is piloting its AI Employees, branded Natalie and Bob, with named customers including Tempo Beer Industries.

How much are suppliers using AI in these deals?

Pactum’s published results say 68% of suppliers engaged through the platform closed an agreement with Walmart, and 83% of suppliers described the system as easy to use in post-engagement surveys.

When will these AI retail agents come to U.S. merchants?

Zenline AI has said it plans to begin offering its AI agents to U.S. retailers this fall. Duvo.ai’s current named clients are in the U.K. and Czech Republic. Gain, which is based in Tel Aviv, is expanding deployments in the United States and Europe.