Strategic Intelligence Brief

Agentic Commerce &
The End of Search-and-Buy

When software makes the purchasing decision, the entire architecture of retail—and B2B commerce—inverts. This is what that means.

Reading Time 12 minutes
Last Updated January 2026
Classification Client Intelligence
Scroll
Key Concept

What Is an Agent?

An AI agent is software that can perceive its environment, make decisions, and take actions to achieve specific goals—operating with varying degrees of autonomy. Unlike traditional automation that follows rigid rules, agents adapt to context and handle exceptions.

In commerce, this means software that doesn't just respond to commands but anticipates needs, monitors conditions, and executes transactions within parameters you define. The human sets the objective and constraints. The agent handles execution.

Executive Summary

The shift from human-driven to agent-driven commerce is not theoretical. It is operational.

For two decades, e-commerce operated on a single interaction model: customers search, browse, compare, and checkout. The entire digital marketing industry—SEO, paid search, conversion optimization—was built to exploit this flow.

Agentic commerce inverts the model entirely. Instead of searching, the customer delegates. Instead of browsing, the customer sets parameters. The instruction might be as simple as "keep my household stocked with essentials under $200 a month." The agent handles the rest.

01

The Structural Shift

When humans shop, they respond to brand recognition, attractive packaging, strategic shelf placement, and promotional messaging. When AI agentsAI AgentSoftware that can perceive its environment, make decisions, and take actions to achieve specific goals—operating with varying degrees of autonomy. shop, they respond to data: accurate pricing, real-time availability, verified ratings, delivery reliability, and return policy clarity.

The emotional and visual triggers that drive human purchasing decisions become less relevant when the decision-maker is software optimizing against defined criteria.

The Customer Experience Inverts

In the new model, the customer doesn't search. The customer delegates. The software maintains standing instructions, monitors context—what you're running low on, what you've purchased before, what preferences you've expressed—and transacts on your behalf within whatever guardrails you've established.

This represents a structural shift, not an incremental improvement. The customer remains in control but no longer performs the labor of shopping.

Who Wins When Agents Shop

Retailers who expose clean, real-time data through reliable APIsAPI (Application Programming Interface)A set of protocols that allows different software systems to communicate. In commerce, APIs enable agents to check prices, inventory, and place orders programmatically. position themselves to win agent-driven transactions. Retailers who depend on impulse purchases, checkout-aisle upsells, and brand-driven premiums will find that leverage diminished.

The strategic question shifts from "how do we capture and hold customer attention?" to "how do we become the default vendor for agents making purchasing decisions on behalf of millions of households?"

Two Models of Commerce

The fundamental interaction pattern is inverting

Traditional Model

Human-Driven

  • Customer searches for products
  • Browses and compares options
  • Adds items to cart manually
  • Completes checkout process
  • Responds to brand and packaging
Agentic Model

Agent-Driven

  • Customer sets parameters once
  • Agent monitors and optimizes
  • Transacts within guardrails
  • Alerts only on exceptions
  • Responds to data and reliability
Now: A Case Study in Execution

Walmart Goes All-In on Agents

The theory above becomes concrete when you examine how the world's largest employer is actually implementing agentic commerce. What follows is that story.

With 2.1 million employees and $648 billion in annual revenue, Walmart's AI implementation offers a reference point for what serious organizational commitment looks like—and a preview of how agentic commerce operates at scale.

The 2025 AI rollout at Walmart isn't a sudden pivot—it's the visible acceleration of nearly a decade of investment. What's new is the urgency. The CEO going to class every week. Every employee getting AI tools on the same day. A dedicated EVP with one mandate: make this happen fast.

$300B
Projected Market
Agentic commerce by 2030 (Bain)
2.1M
Employees
All given AI tools in January 2025
95%
Engineer Adoption
Using AI coding assistants daily
200+
Agents Deployed
Consolidated into 4 super agents
We're at a pivotal moment for Walmart and the future of retail. The rise of agentic AI is more than a technological shift.
Doug McMillon, CEO, Walmart

Walmart's Execution Timeline

Walmart has been building its technology foundation for years. But with the arrival of generative AI, the intensity shifted. What follows is a timeline of their organizational commitment—not the start of the journey, but the moment it accelerated.

December 2024

Leadership Immersion

CEO Doug McMillon and all direct reports began attending weekly AI classes—not a one-time seminar, but ongoing hands-on workshops. Leaders model the behavior they expect.

January 2025

Universal Access

All 2.1 million Walmart employees received free ChatGPT licenses with explicit encouragement to start using them immediately. Not a pilot program. Not a select group. Everyone.

Spring 2025

Strategic Leadership Hire

Created EVP of AI Acceleration reporting directly to the CEO. Hired Daniel Danker—former Instacart CPO, Uber Eats, Facebook, Shazam, 10 years at Microsoft. Not someone learning on the job.

October 2025

OpenAI Partnership

Launched Instant Checkout integration with ChatGPT—customers can shop directly through the AI without leaving the conversation.

December 2025

NYSE to NASDAQ

Walmart transferred its stock listing from the New York Stock Exchange to NASDAQ—the largest exchange transfer in history. A signal: Walmart now sees itself as a technology company that conveniently has physical stores.

The Super Agent Framework

Walmart consolidated 200+ specialized AI agents into four unified systems, each serving a distinct user group

Sparky Agent
For Customers

Shopping assistant inside the app. Helps find products, navigate deals, and complete purchases through natural conversation.

Associate Agent
For Employees

Streamlines scheduling, HR requests, and sales data access for store associates. Reduces administrative friction.

Marty Agent
For Partners

Assists suppliers and advertisers with onboarding, campaign setup, and performance analytics.

Developer Agent
For Engineers

Accelerates testing, deployment, and bug resolution. 95% of bugs now auto-fixed without human intervention.

02

Beyond Retail

The same pattern applies wherever routine purchasing or scheduling decisions can be delegated within defined parameters.

B2B Procurement

A procurement manager at a mid-sized manufacturer might authorize an agent to monitor inventory levels, compare supplier pricing against approved vendors, verify lead times, and place orders automatically—intervening only when something falls outside established bounds.

Service Businesses

An agent that handles appointment scheduling, quote follow-ups, and reminder sequences without human involvement until an exception requires attention.

Agent Governance

An agent that drafts routine documents, files standard forms, and manages intake workflows, freeing practitioners for work that genuinely requires their judgment.

We're not at a point where we can let an agent run loose and let it solve all kinds of problems. We will still keep humans in the loop.
Sravana Karnati, EVP Global Technology Platforms, Walmart

The Workforce Paradox

McMillon has been direct: "AI is going to change literally every job. While some jobs and tasks at Walmart will be eliminated, others will be created."

Walmart expects its global workforce to stay roughly flat over the next three years, even as the mix of jobs changes significantly. They're not using AI to reduce headcount. They're using it to change what people do.

The underlying logic remains consistent: define the objective, establish the constraints, and let the agent execute within those boundaries. Humans handle strategy, exceptions, and the work that requires genuine expertise. Agents handle volume, velocity, and tasks that consume time without requiring thought.

Practical Implications

The framework scales down. These principles apply whether you have 2.1 million employees or 21.

01

The Agent Framework Model

Don't attempt to build one massive AI system. Build specialized agents for specific tasks, then unify them under super agents for each user group. Deploy one, prove it works, measure results, then expand.

02

The Ownership Question

Walmart created an EVP for AI acceleration. Private businesses don't require that level of structure—but they do require someone with clear accountability. Without explicit ownership, initiatives drift into perpetual pilots.

The Competitive Reality

Walmart isn't implementing AI because its leadership is enthusiastic about technology. They're implementing it because the economics of their industry demand it—and because they've concluded that the pace of change leaves no room for gradual adoption.

McMillon has been honest about the discomfort: "I love change except when it relates to me." He acknowledges that change is uncomfortable—even for the person leading it. But he moved anyway.

The competitive pressure Walmart is responding to applies with equal force to every business.

"Listen to your gut. Go hard and go fast. You know those times in your life when you look back and wish you'd gone faster."

— Doug McMillon, CEO, Walmart