Designing for a new buyer: How AI agents will transform digital commerce

How Agentic AI will change product discovery, pricing, and transactions, and what technology/digital marketing leaders should do about it.

About six months ago, I posted a short message on LinkedIn with a few references from MIT, stating that AI would become our most important customer. Not only was the prediction very accurate, but it is also becoming increasingly important to prioritize our digital presence in preparation for the new customer segment (Agentic AI).

Agentic AI systems from OpenAI¹, Google², and Perplexity³ can already:

• Search across the entire web

• Compare pricing and product attributes

• Add items to shopping carts

• Apply discount codes

• Verify the checkout flow

• And even place orders automatically

All from a single prompt.

Amazon has already issued cease-and-desist notices⁴ to prevent autonomous agents from placing orders on its platform. That alone should signal how quickly this new customer class is emerging.

Why This Matters (Porter’s Five Forces View)

Through Michael Porter’s Five Forces Framework⁵ lens, two forces spike immediately:

1. Buyer Power = Extremely High

Agentic AI has no emotional connection or loyalty to brands, can conduct thousands of searches concurrently, and has no fatigue.

It seems the Agentic AI optimizes shopping for transparency, accuracy, speed, and price.

2. Threat of Substitute = Extremely High

If product content, price, and other important product information aren’t accessible to AI Agents, they will automatically route to a competitor who is.

We are entering a new era in digital history where our new buyer is fully rational and fully automated.

Two Strategic Opportunities To Prioritize

I’d like to propose two ideas to not only plan for the upcoming disruption, but also to leverage them to stay ahead of the competition.

1. Model Context Protocol (MCP)

The best and concise explanation of MCP⁶ is to think of it as the USB-C port for AI. In other words, all modern electronic devices utilize a single standard USB-C port for charging and connecting to external devices. An organization can utilize MCP to establish a digital USB-C connection, enabling its customers’ AI agents to communicate with their marketing and e-commerce sites, learn from them, and transact.

Think of MCP as a standard interface that lets AI agents:

• Navigate our website

• Read our marketing content, product catalog, return and warranty policy, etc.

• Validate pricing

• Place orders directly

OpenAI, Anthropic, Google, and others enable the addition of MCP tools with a single checkbox. As I reflect on the MCP evolution over the past few months, the Claude Desktop app, for example, has undergone significant improvements, adding the MCP connection from updating the JSON file to now simply checking a box to add a new MCP.

Why is this urgent?

If competitors expose MCP endpoints before we do, their products become instantly more discoverable and more agent-friendly.

2. Conversion Rate Optimization (CRO) strategy for AgenticAI

For 20+ years, the CRO initiative⁷ remained focused on human behavior:

• UX/UI improvements

• A/B testing

• Navigation optimizations (new customers vs. existing customers)

• Mobile responsiveness

• Frictionless customer experience, especially during the checkout process

• Trust signals such as reviews, return policy, and warranty claims

While we are still researching to truly identify what AI Agents prioritize, my hypothesis suggests that they don’t care about branding, layouts, colors, or emotions. Instead, AI agents prioritize:

• Clean, structured product data

• Schema markup

• Clear pricing logic

• Error-free APIs (i.e., shipping rates, credit card transactions, etc)

• Fast response time, page load time

• Machine-navigable flows

• Transparent policies

Assuming my hypothesis holds water, this shifts CRO from Human Experience Optimization to Machine Reasoning Optimization.

In that case, here are some natural questions that we need to unpack:

• Can an AI agent parse our product data without hallucinating?

• Are our company policies machine-readable?

• Can the checkout process be executed autonomously?

• Is our product catalog structured enough for an AI agent to compare SKUs at scale?

The Bottom Line

To my fellow technology leaders: Just as Agentic AI is redefining digital transactions, we can utilize the same technology to disrupt ourselves and our organizations before our competitors do so proactively. This is the moment to combine our technical expertise with business ownership and create solutions that materially strengthen our organization’s competitive positioning.

1. Build Prototype

Use AI Agent to rapidly prototype

• MCP connection

• Agent-driven checkout flow simulations to identify gaps

• Structured product data tool to test if a Product Information Management (PIM) makes sense for your business. If so, build the business case for it.

• Automated SKU comparisons and pricing audits

Demo your prototype to your leadership, sales, and marketing teams, showing them what the future looks like before anyone asks for it.

2. Use AI to Become Fluent in the Business Side

As AI democratizes access to the technology, why don’t we use AI to democratize access to the business world and differentiate our ability to:

• Understand commercial strategy

• Translate customer pain points into technical opportunities

• Communicate solutions in business terms (not technology vocabularies)

• Connect technology decisions to revenue, customer retention, and margins

Please share your thoughts, and let’s learn from one another and drive this next wave of innovation together.

1. https://openai.com/index/buy-it-in-chatgpt/

2. https://blog.google/products/shopping/agentic-checkout-holiday-ai-shopping/

3. https://www.perplexity.ai/hub/blog/shop-like-a-pro

4. https://www.perplexity.ai/hub/blog/bullying-is-not-innovation

5. https://hbr.org/2008/01/the-five-competitive-forces-that-shape-strategy

6. https://www.anthropic.com/news/model-context-protocol

7. https://www.optimizely.com/optimization-glossary/conversion-rate-optimization/