Agentic commerce is reshaping the purchasing journey. Is your company ready for autonomous shopping?

We are living in the age of agentic commerce. Agentic commerce is more than chatbots and recommendation engines: it means that AI agents can search, decide, and transact across channels on behalf of users. A customer could delegate their grocery shopping list to an agent, then find everything researched, compared, purchased, and scheduled for delivery.

As outlined in a recent paper, agents are increasingly used for product discovery, decision‑making, and transactions. For retailers, that means planning for brand interactions with AI agents along with human customers.

But agentic commerce is a fundamental re-architecture of the commerce stack—it’s a journey from generative AI (genAI) enabled search all the way to transactions where both the buyer and seller are AI agents.

Agentic commerce is a journey

It begins with a strong unified commerce foundation

Unified commerce is the seamless integration of near‑real‑time platforms that connect all sales channels with core backend systems. Before retailers can accommodate AI agents on their backend or from customers, they need a unified commerce foundation. Fragmented monolith systems and unreliable data quality remain the primary barriers as retailers embark on this journey.  

1. Assisted discovery

Once data has been centralized, personalization and recommendation agents can surface products based on known customer interests. For example, a makeup brand’s assistant might recommend a particular eyeliner, but the shopper would still complete the purchase.

Product discovery is rapidly moving into AI‑driven environments, with traffic from generative AI sources to US retail sites surging by 4,700% in just one year.¹ Already, 55% of customers start researching products on large language models.²

2. Assisted shopping

Conversational interfaces and agents help customers discover, compare, and purchase products—this includes assistance with navigation, queries, and recommendations.

3. Agentic shopping

Generative AI platforms and intelligent brand agents enable customers to discover, compare, and purchase products directly through conversation—ushering in “zero‑click” commerce. The transaction completes inside an assistant shell: a retailer’s system will fulfill the order but the shopper has never visited the retail site.

4. Autonomous shopping

When a brand agent shops autonomously, they act on behalf of customers: proactively searching, deciding, and transacting across channels within approved parameters. An agent could fill forms, apply codes, click buttons, and complete the purchase without needing any input from the customer.

A customer agent acts proactively on the customer’s behalf—monitoring prices and inventory, reordering consumables just-in-time, identifying acceptable substitutes, and initiating negotiations for price matches or delivery upgrades—while escalating exceptions for human review.

5. Agent-to-agent commerce

In the final stage of the agentic commerce journey, agnostic customer agents interact directly with any number of brand agents to complete transactions. For example, a single customer agent tasked with interior design might transact with brand agents selling furniture, decor, and other household items.

Customer agents and brand agents transact directly across the full lifecycle—from intent broadcast and offer qualification to negotiation, payment, and fulfillment orchestration.

A generational shift, an emerging channel

Agentic commerce has moved beyond being a feature embedded within existing retail channels into a standalone channel alongside e-commerce, retail media, and stores. The result is a retail landscape with evolving leadership expectations:

• In 2026, 40% of enterprises plan to leverage AI agents³
• By 2030, 25% of global e‑commerce sales are expected to be enabled by AI agents⁴  

Gen Z and digital‑first consumers increasingly rely on AI assistants as their primary interface to discover, compare, and select products. According to a 2025 study, 61% of Gen Z shoppers used AI tools to assist with a purchase in the past year.⁵

But Canadian companies haven’t adopted AI as quickly as their US counterparts, as shown by a study showing Canadian companies lagging US companies in AI adoption by 37%.⁶ Canadian retailers have to be ready when US brands launch agents for the Canadian market. Without deliberate redesign, control of these users’ experiences will shift from retail brands to agentic platforms. Retail expectations already illustrate this trend:

• 63% of global retailers believe companies without AI agents will fall behind within two years.⁷
• 58% of retailers expect AI agents to handle most customer interactions within five years⁸

The long-term implications are even more striking. Disruptive scenarios anticipate genAI-driven and agentic commerce will be the fastest-growing channels, impacting channel revenue split.  

In short: retailers must either risk delay or reap the rewards of action. Now is the time to act.

Five ways retailers can meet the moment

1. Fortify the data core, and build the unified commerce technology foundation

Agentic commerce is not just “chat + checkout.” It’s orchestration across systems, including the ability to recommend products, check inventory, compare fulfilment options, apply loyalty rewards, manage pricing and promotions, execute payments, and trigger fulfilment and post-purchases within your agentic experience. Retailers should ask:

• Are these systems modular and API enabled?
• Is the product, pricing, inventory, and fulfillment data machine‑readable and agent‑ready—SEO to GEO?

An agent is only as reliable as your data allows it to be, and that process begins with a unified commerce infrastructure that can be scaled. The unified commerce platform orchestrates systems and data, while a multi‑agent layer sits on top—using protocols such as MCP (Model Context Protocol)—to enable fully automated, “no‑visit” orders completed entirely by agents.

2. Secure trust at the point of interaction and establish guardrails

Trust is the prerequisite for scale, regulatory confidence, and customer adoption as AI acts autonomously. Building and maintaining that trust requires strengthening agent‑safe user experiences and authentication.

Strengthening those pillars will require retailers to implement:

• Signed agent identities
• Content provenance and auditability
• Safeguards for high‑risk actions such as checkout, consent, and pricing
• Secure against hijacking

3. Establish clear governance, standards and reporting

Agentic commerce is not simply a technology enablement, but a new customer and commercial channel. It requires deliberate governance across people, process, and strategic alignment. Retailers must establish dedicated KPIs for the agentic channel and strengthen their ability to observe, measure, and optimize agent behavior through enhanced traffic instrumentation and analytics.

Emerging standards that consider interoperability, auditability, and compliance take into account:

• AI‑generated content
• Data sharing with platforms and partners
• Boundaries for the agent to operate autonomously

4. Define and integrate the agentic channel

In addition to treating agentic commerce as a distinct channel, a holistic channel view will still require integrating agent‑driven journeys with:

• E-commerce
• In‑store experiences
• Customer service

Retailers will also need to define where control and data should live for owned and third‑party channels.

5. Use AI efficiency to self‑fund transformation

Seventy-five percent of retailers believe AI agents will be essential as a competitive advantage by 2026, and 76% plan to increase investment in AI agents in the next year.⁹ Early AI investments have already delivered measurable productivity gains, including:

• Up to 34% higher customer service resolution rates¹⁰
• Two‑thirds of customer inquiries resolved in under two minutes¹¹
• 2–5% revenue uplift¹²

Further productivity gains could be a self‑financing flywheel that scales agentic commerce without incremental funding. Leading retailers will be able to reinvest a significant portion of these gains into data readiness, security, consent, and the evolution of governance and operating models.

Questions to begin your agentic journey

Retailers can begin their agentic commerce journey by asking essential questions about the ways agentic AI can be integrated, including:

1. Over the next 12–24 months, what is the expected payoff from agentic commerce, and what level of investment does it warrant?
2.  Which assets should be:
   • Agent‑exclusive?
   • Limited to loyalty programs?
   • Direct‑only to preserve first‑party data and brand equity?
3. What proportion of transactions should flow through third‑party agents versus owned channels?
4. Which proprietary data do we:
   • Ring‑fence?
   • License?
   • Share with LLM partners, and under what economic and governance models?
5. How do we redesign pricing, promotions, and merchandising for agent‑to‑agent dynamics?

Deloitte can help you on your journey

No matter where you are on your agentic commerce journey, Deloitte helps retailers move from a unified data foundation to enterprise‑scale value. Our advantage lies in treating AI agents as a business transformation, not as a one-off technology deployment.

Connect with Deloitte’s agentic commerce leaders to explore how you can lead your company’s journey with confidence.