Walmart AI Strategy: Getting Ready for Agent-Led Shopping

Walmart is betting that AI agents will change how people shop, and it is building for that shift both in its own app and inside other companies’ AI assistants. Its shopping agent, Sparky, runs inside ChatGPT, and Walmart has built a shopping experience for Google’s Gemini. It also expects to keep headcount flat while AI changes the work of its 2.1 million associates.

The ambition is to lead agent-led commerce. Announcing a Google partnership in January 2026, CEO John Furner said: “The transition from traditional web or app search to agent-led commerce represents the next great evolution in retail.” He added: “We aren’t just watching the shift, we are driving it.”

Walmart names the industry shift directly. Former CEO Doug McMillon said online shopping has long been “a search bar and a long list of item responses”, and that this is about to change (OpenAI partnership release, October 2025).

The focus is four “super agents.” In July 2025, Walmart consolidated dozens of AI tools into four agents, as reported by CO/AI:

• Sparky, for customers
• Marty, for suppliers, sellers and advertisers
• an associate agent, for employees
• a developer platform, for building new agents

CTO Suresh Kumar said Walmart was “deliberately choosing to go beyond individual tools and build a unified, company-wide framework.”

The primary focus is the customer and growth. Operational efficiency comes second, mainly through supply chain automation.

Walmart has not said what it will not prioritize. It has said it is cutting redundant AI use cases to control costs.

Walmart builds through partners. CFO John David Rainey said in February 2026: “We’re approaching AI development through partnerships” (CIO Dive).

• OpenAI: customers can buy from Walmart inside ChatGPT, and Sparky has run there since March 2026.
• Google: Walmart built a shopping experience for Gemini using the Universal Commerce Protocol, launching first in the US.

It keeps the customer relationship. In ChatGPT, checkout stays with Walmart. In Gemini, customers link their Walmart accounts and get member benefits.

AI reaches the store floor. In June 2025, Walmart added AI tools to its associate app for 1.5 million associates, including task management, real-time translation in 44 languages and a generative AI assistant, as reported by Retail IT Insights. The conversational assistant had more than 900,000 weekly users. Tools like these take over scheduling and task allocation, work that once trained managers, a shift explored in How AI is Breaking Middle Management.

The workforce stance is specific. McMillon said in September 2025: “It’s very clear that AI is going to change literally every job.” A Walmart spokesperson added: “Over the next three years we expect headcount to remain flat as roles evolve” (Fortune).

New roles are appearing. In August 2025, Walmart created an “agent builder” role, in which employees build AI tools for merchants.

Enablement runs through certifications. In June 2026, Walmart launched an OpenAI certification program for associates, following an earlier Google AI certificate (Modern Retail).

Associates are building their own tools. One associate designed an AI agent that trains bakery staff on cake decorating from a photo. A logistics manager built an app that helps drivers find loads that get them home on time.

Walmart has not disclosed whether training is mandatory or how many associates are certified.

Ownership spans the top team. Daniel Danker joined in 2025 as EVP of AI acceleration, product and design, leading the AI transformation strategy. A separate EVP of AI platforms reports to the CTO. John Furner, CEO since February 2026, is the public champion of agent-led commerce. No AI-specific investment figure has been disclosed.

Governance rests on a published pledge. Walmart’s Responsible AI Pledge (October 2023) sets six commitments: transparency, security, privacy, fairness, accountability and customer-centricity. Under accountability, AI stays under human management.

Shareholders rejected more disclosure on workforce effects. In June 2026, a proposal for a report on how AI affects Walmart’s workforce received about 5% of votes (PYMNTS). The board said technology “should serve people by elevating human capability.”

AI fluency, coded against the Lead with AI Fluency Matrix (September 2026 edition):

• Fluency floor: Not disclosed. Free certifications are offered, but no mandate or standard was found.
• Highest level evidenced: Level 4 (Transforming), company-claimed. Sparky is a working shopping agent at scale. It can build a weekly meal plan and add the ingredients to the basket, and Walmart reports its results every quarter. The customer still confirms checkout.
• Portfolio design: Multi-level, by audience. Customers, suppliers, associates and developers each have their own agent, and agent-builder roles add a builder track. No role-specific fluency targets have been disclosed.

Walmart reports Sparky every quarter. All figures below are company-claimed, as reported by AIM Media House from the earnings call:

• Sparky users were up 70% year on year in Q2 of fiscal 2027 (August 2026).
• Sparky users spent 40% more per order than non-users that quarter, up from 35% earlier in the year.
• Goods sold through Sparky rose 150% year on year in Q1 of fiscal 2027.

The order-value figure needs care. It compares Sparky users with non-users, so part of the gap may reflect who chooses to use Sparky rather than what Sparky does.

Operational results:

• customer care resolution times cut by up to 40%
• fashion production timelines shortened by up to 18 weeks
• shift planning for team leads cut from 90 minutes to 30, in early results

Automation is measured too. About 3,100 of roughly 4,600 US stores receive some freight from automated distribution centers, and more than half of e-commerce fulfillment volume is automated. These are automation figures, not AI fluency.

Labor productivity shows up in results. In August 2026, the CFO credited store associates’ use of technology tools, along with automation, for holding down wage costs as a share of sales.

Two things are missing: a dollar value for AI and a dated AI roadmap. The one multi-year commitment is flat headcount over three years.

Change log

• October 2023: The Responsible AI Pledge is published.
• June 2025: AI tools reach 1.5 million associates, including translation in 44 languages.
• July 2025: Four super agents are announced, and Daniel Danker joins to lead AI acceleration.
• August 2025: Walmart creates an “agent builder” role.
• September 2025: McMillon says AI will change “literally every job,” Walmart expects flat headcount over three years, and it partners with OpenAI on certifications.
• October 2025: The OpenAI partnership brings Walmart shopping into ChatGPT.
• January 2026: The Google partnership brings Walmart shopping into Gemini.
• February 2026: John Furner becomes CEO.
• March 2026: Sparky starts operating inside ChatGPT.
• June 2026: The associate OpenAI certification launches, and shareholders reject an AI workforce report.
• August 2026: Sparky users are up 70% and spend 40% more per order.

• Disclosure: 4 of 4. Walmart answers all four questions. Its measurement is product-level, with no dollar value for AI.
• It is the clearest case of AI as a bet on competitive position. The strategy is about how customers will shop, not only about cost.
• It spreads its bets across platforms, from its own app to ChatGPT and Gemini, while keeping checkout and the account relationship.
• Its workforce commitment is concrete: flat headcount over three years, new agent-builder roles and free certifications. Headcount has held at about 2.1 million since fiscal 2023 while revenue grew 17%.
• The say-do gap is in measurement. The Sparky figures compare users with non-users, and workforce effects are not reported publicly.
• Watch agent-led sales. The test is whether Walmart starts reporting the share of sales that comes through AI agents, its own and others’.

What the numbers show: Revenue grew 17% while headcount stayed at about 2.1 million, in line with Walmart’s stated plan to keep headcount flat for three years while roles change.

Revenue, headcount and revenue per employee, FY2023 to FY2026

• FY2023: $611.3bn revenue, about 2.1 million employees, $291k per employee
• FY2024: $648.1bn revenue, about 2.1 million employees, $309k per employee
• FY2025: $681.0bn revenue, about 2.1 million employees, $324k per employee
• FY2026: $713.2bn revenue, about 2.1 million employees, $340k per employee

Change, FY2023 to FY2026: revenue up 17%, headcount flat, revenue per employee up 17%.

Notes: Revenue is total revenues, as reported. Headcount is associates at fiscal year-end. Walmart’s fiscal year ends on 31 January. Walmart rounds headcount to the nearest 0.1 million, so the per-employee trend mirrors revenue growth.

These figures give context on scale and productivity. They do not show that AI caused any change. Sources: Walmart 10-K filings.

What is Walmart’s AI strategy?

Walmart is preparing for a shift from search to agent-led shopping. It has built four AI super agents for customers, suppliers, associates and developers. Its customer agent, Sparky, works in the Walmart app and inside ChatGPT, and Walmart has built a shopping experience for Google’s Gemini.

‍

What is Sparky?

Sparky is Walmart’s AI shopping agent. It can answer product questions, build meal plans and add items to a customer’s basket. Walmart says Sparky users grew 70% year on year in its August 2026 quarter.

‍

Can you buy from Walmart in ChatGPT or Gemini?

Yes. Walmart partnered with OpenAI in October 2025 so customers can shop inside ChatGPT, and with Google in January 2026 for shopping inside Gemini. Checkout and account benefits stay with Walmart.

‍

Will AI cut jobs at Walmart?

Walmart expects headcount to stay roughly flat over the three years from 2025 while roles change. Its former CEO said AI will change “literally every job,” and the company offers associates free AI certifications.

‍

How is this profile put together?

Lead with AI profiles use public sources only, such as annual reports, earnings materials, filings and on-record interviews, dated at the time of writing. AI fluency is coded against the Lead with AI Fluency Matrix. Company figures are marked as company-claimed unless they have been independently verified.