Walmart: Generative AI Enriched 850M Catalog Items 100x Faster While Autonomous Bots Negotiate Supplier Contracts
Enriching 850 million catalog items and launching generative intent search lifted e-commerce search relevance and basket size by turning multi-query shopping trips into single-prompt solutions.
| Company & Sector | Walmart — Global Retail & Supply Chain |
|---|---|
| Models & AI Stack | Wallaby Retail LLM · Generative Intent Search · Catalog Attribute Extraction · Pactum Autonomous Negotiation AI · My Assistant Enterprise Portal · Me@Walmart Associate AI |
| Rollout Timeline | 2022–2023 — scale-up of autonomous supplier negotiation bots (Pactum) and My Assistant rollout to 50,000+ staff; 2024 — 100x faster enrichment of 850M catalog items and GenAI search launch; 2025–2026 — deployment of Wallaby retail LLM and international expansion across 11 countries. |
| Verification | Verified 2026-10 · 3 primary sources |
Business Context & Scale
Walmart is the world's largest retailer by revenue ($640B+ annually), operating over 10,500 stores and Sam's Club locations across 19 countries with roughly 2.1 million associates.
Across its omnichannel e-commerce platform, Walmart manages a massive catalog exceeding 850 million product listings spanning first-party inventory and third-party marketplace sellers.
Between 2023 and 2025, Walmart Global Tech executed three landmark retail AI deployments: generative search and catalog enrichment across 850M items, the 'My Assistant' workplace platform for 75,000+ corporate employees, and autonomous AI supplier negotiations.
Business Problem & Bottlenecks
The first challenge in an 850-million-item marketplace is sparse, inconsistent supplier product data. Without clean attributes (sizing, compatibility, occasion, materials), legacy keyword search fails on intent-driven queries like 'help me plan a rainy kids soccer party.'
Second is tail-spend procurement: human buyers lack the bandwidth to conduct multi-round negotiations on payment terms, volume discounts, and freight clauses with thousands of smaller indirect and tail-assortment suppliers.
Third is store-associate productivity across 10,500+ supercenters handling millions of daily curbside pickup and last-mile delivery orders.
AI Architecture & Implementation
Walmart Global Tech developed its retail-tuned 'Wallaby' LLM platform and processed over 850 million catalog listings—extracting missing attributes from images and raw vendor text, standardizing specifications, and powering intent-based generative search.
As CEO Doug McMillon announced, enriching and standardizing those 850 million catalog items with Generative AI was completed in roughly 1/100th of the time it would have required using manual human workflows.
In procurement, Walmart deployed autonomous negotiation bots (with Pactum AI) that negotiate payment terms and volume discounts with tail-spend suppliers within buyer-defined guardrails. Internally, Walmart rolled out 'My Assistant' to 75,000+ corporate associates across 11 countries alongside voice AI in the Me@Walmart store app.
“"Using generative AI to improve our product catalog of 850 million items allowed us to complete work that would have required nearly 100 times the current headcount to do in the same amount of time." — Doug McMillon, CEO of Walmart
Measurable Impact & Economics
Enriching 850 million catalog items and launching generative intent search lifted e-commerce search relevance and basket size by turning multi-query shopping trips into single-prompt solutions.
In tail-spend procurement, autonomous AI negotiation bots successfully close agreements with 68% of engaged suppliers, delivering average cost savings of 1.5–3% while shrinking contract turnaround from weeks to days.
Notably, 75% of participating suppliers reported preferring the transparent, self-paced AI negotiation interface over high-pressure human buyer calls.
Key Takeaways for Industry
Consumer-facing generative search is only as good as the underlying catalog: batch LLM attribute enrichment across 850M SKUs is the true foundation.
Tail-spend procurement contracts are ideal for autonomous negotiation bots that trade off payment terms against unit discounts.
Combining a domain-trained retail LLM (Wallaby) with frontier models improves taxonomy accuracy and shopper intent matching.
Giving store floor associates voice-driven inventory lookup inside handheld apps shaves minutes off every curbside fulfillment order.
Verified Primary Sources
Key Questions Answered
How can an AI bot negotiate commercial terms with a human supplier?
The bot operates within a buyer-defined trade-off matrix (e.g., faster payment terms in exchange for a unit discount or larger volume commitment), letting the supplier select the package that best fits their cash flow.
What exactly did Generative AI do across the 850 million catalog listings?
Models parsed raw vendor descriptions and product images to backfill missing filter attributes (material, occasion, compatibility), deduplicate listings, and rewrite cryptic technical titles.
How does Walmart's generative search differ from traditional keyword search?
Instead of searching item by item ('tent', 'sleeping bag'), shoppers type a goal ('autumn camping trip with two kids') and receive a curated multi-category checklist tailored to their context.
Build & Ship Production AI Agents Without Boilerplate Code
Start your free 72-hour trial at Hamidun Academy (academy.hamidun.com) — interactive lessons on Claude Code & Cursor, live Tech Radar, and production agent templates.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.