Part III: The Agentic Ecosystem—Mapping Uncharted Territory

Here's where things get less certain—and far more interesting.

Over the past few weeks, as we've been stress-testing our thesis with advisors, conducting research, and mapping out where this all leads, we've arrived at a framework that we believe explains how the e-commerce industry will restructure itself. This isn't set in stone. The architecture is still forming.

But based on everything we've seen, the conversations we've had, and the underlying technological shifts already in motion, here's what we believe is coming:

The e-commerce industry will split into two distinct layers: the Core Layer and the Enablers Layer.

This is the new decision-making layer—where the actual shopping, negotiating, and purchasing happens. It consists of three types of agents:

And then there are Specialized Agents—fraud detection agents, payment routing agents, logistics coordination agents—handling specific, high-complexity tasks that neither Personal nor Merchant Agents want to manage directly.

This Core Layer is where the future of commerce lives. If you're not integrated here, you don't exist.

Here's the truth: you're not going to become a Core Layer player. You're not ChatGPT. You're not building the Personal Shopping Agent. That ship has sailed, and the capital required to compete there is in the tens of billions.

But that doesn't mean you're obsolete.

Below the Core Layer sits the Enablers Layer—the infrastructure that makes the Core Layer function. This is where your platform lives. This is Shopify, Stripe, Klaviyo, your CDP, your analytics stack, your inventory management system, your fulfillment partners. These tools don't disappear. They evolve.

The shift: Personal Shopping Agents aren't built to manage end-to-end operations. They're not built to send emails, manage inventory, handle fraud detection, run A/B tests, optimize supply chains, or process international payments. They're built to orchestrate—to understand what the customer needs and then delegate the work to specialized enablers.

Here's how it works:

The Personal Agent doesn't build an email system. It doesn't become Klaviyo. Instead, it queries the Enablers Layer, finds Klaviyo's Merchant Agent, brokers the integration, and connects the store. From that point forward, Klaviyo handles email and SMS. The Personal Agent just orchestrates.

This is your survival path: Become deeply, natively integrated with the Core Layer.

Right now, you're competing on features. Better analytics. Smarter personalization. Shinier dashboards. You're trying to win the battle of Phase Two while Phase Three is already here.

The companies that win over the next 5-8 years won't win because they have the best dashboard. They'll win because they're the easiest for agents to work with.

Here's what that means in practice:

Your primary engineering challenge for the next five years is building the bridges—the APIs, the structured data models, the agent-readable protocols—that connect your platform to the Core Layer. This isn't about adding an AI chatbot to your dashboard. This is about making your entire platform natively agent-accessible.

Here's the good news: as the Core Layer scales, it will generate enormous demand for specialized enablers.

Fraud detection becomes exponentially more complex when millions of agents are transacting autonomously. Payment routing, international settlements, tax compliance—these all become harder, not easier, in an agentic world. Supply chain coordination, A/B testing in agent-mediated environments, analytics that track agent behavior instead of human clicks—these are new categories of problems.

The companies that solve these problems—that build the enablers the Core Layer depends on—will capture massive value. But only if they're architected correctly from the start.

Through our thorough breakdown we positively believe -

The question isn't whether you should pivot to serve the Core Layer. The question is whether you'll do it fast enough.

Part IV: The Advertising Apocalypse (And What Comes After)

Your merchants have been pouring billions into Facebook ads, Google Shopping campaigns, and retargeting pixels. They've mastered the art of the algorithmic auction, optimized their CAC:LTV ratios, and built entire businesses on the back of performance marketing.

That entire playbook is about to be incinerated.

When the primary "user" viewing products is an AI agent—not a human with eyeballs and emotions—your current advertising infrastructure becomes worthless. Agents don't click banner ads. They don't get swayed by lifestyle photography. They don't impulse buy because of a well-timed retargeting campaign.

Let me paint you the coming future of "A customer" - it tells their Personal Agent, "I need running shoes." That agent doesn't browse Instagram. It doesn't see your merchant's Google Shopping ad. It queries 50 merchants simultaneously, evaluates based on price, availability, reviews, and shipping times, and makes a purchase decision in 300 milliseconds.

Where exactly does your merchant's $10,000 monthly ad spend fit into that equation?

It doesn't. Which is why three entirely new advertising models are emerging to replace everything you know:

This is already happening. Perplexity is testing it. OpenAI is building it. Within 18 months, it will be the dominant advertising model.

When you ask for recommendations, it will show 1-2 paid recommendations which fits what you are looking to buy.

The targeting precision makes Facebook's algorithm look like a shotgun blast. We're talking about 100x better conversion rates because the AI only shows ads when the intent-to-purchase probability crosses a threshold.

The Paradox here: This will initially feel like paradise for your merchants. Higher ROAS than they've ever seen. But they're about to trade one monopoly (Google/Meta) for another (OpenAI/Anthropic). And this time, the gatekeeper has even more power because they control not just discovery, but the entire purchase decision.

Here's the second-order effect most people miss: When customers delegate purchasing to agents, brand preference becomes pre-decision.

Think about what happens today when someone says, "Hey Siri, order me some batteries." If they don't specify a brand, Siri defaults to Amazon Basics. The entire consideration set is eliminated before the search even begins.

Now scale that to every purchase decision. The customer doesn't say "I need running shoes." They say "I need Nike running shoes" or "Order my usual Allbirds." The brand that owns the namespace owns the sale.

This creates a new advertising battlefield: memory warfare. Brands won't advertise products; they'll advertise recalls. The goal isn't to convince you to buy; it's to ensure that when you delegate the purchase, you include their brand in the instruction.

How SaaS comes into picture here?

You need to help your merchants win the namespace battle. That means providing tools for brand-building, not just performance marketing. The platform that helps merchants become the reflexive choice wins.

This is the nuclear weapon of agentic advertising, and almost no one sees it coming.

Today, Google knows what you search. Facebook knows what you like. Amazon knows what you buy. But your Personal Agent? It knows everything. Every conversation, every question, every hesitation. Every abandoned intent.

These agents aren't just using this data to help you shop—they're monetizing it to advertisers. But not in the way you think.

Companies like Perplexity and OpenAI are launching their own browsers. When you browse through their browser, they capture every page visit, every hover, every scroll pattern. They're building intent graphs that make Google's data look prehistoric.

The potential goldmine: As these AI models integrate with more apps and collect intent signals across the entire user journey, they create an unprecedented advertising opportunity. The depth of context would make current targeting look primitive.

Part IV: The Advertising Apocalypse (And What Comes After)

Your merchants have been pouring billions into Facebook ads, Google Shopping campaigns, and retargeting pixels. They've mastered the art of the algorithmic auction, optimized their CAC:LTV ratios, and built entire businesses on the back of performance marketing.

That entire playbook is about to be incinerated.

When the primary "user" viewing products is an AI agent—not a human with eyeballs and emotions—your current advertising infrastructure becomes worthless. Agents don't click banner ads. They don't get swayed by lifestyle photography. They don't impulse buy because of a well-timed retargeting campaign.

Let me paint you the coming future of "A customer" - it tells their Personal Agent, "I need running shoes." That agent doesn't browse Instagram. It doesn't see your merchant's Google Shopping ad. It queries 50 merchants simultaneously, evaluates based on price, availability, reviews, and shipping times, and makes a purchase decision in 300 milliseconds.

Where exactly does your merchant's $10,000 monthly ad spend fit into that equation?

It doesn't. Which is why three entirely new advertising models are emerging to replace everything you know:

This is already happening. Perplexity is testing it. OpenAI is building it. Within 18 months, it will be the dominant advertising model.

When you ask for recommendations, it will show 1-2 paid recommendations which fits what you are looking to buy.

The targeting precision makes Facebook's algorithm look like a shotgun blast. We're talking about 100x better conversion rates because the AI only shows ads when the intent-to-purchase probability crosses a threshold.

The Paradox here: This will initially feel like paradise for your merchants. Higher ROAS than they've ever seen. But they're about to trade one monopoly (Google/Meta) for another (OpenAI/Anthropic). And this time, the gatekeeper has even more power because they control not just discovery, but the entire purchase decision.

Here's the second-order effect most people miss: When customers delegate purchasing to agents, brand preference becomes pre-decision.

Think about what happens today when someone says, "Hey Siri, order me some batteries." If they don't specify a brand, Siri defaults to Amazon Basics. The entire consideration set is eliminated before the search even begins.

Now scale that to every purchase decision. The customer doesn't say "I need running shoes." They say "I need Nike running shoes" or "Order my usual Allbirds." The brand that owns the namespace owns the sale.

This creates a new advertising battlefield: memory warfare. Brands won't advertise products; they'll advertise recalls. The goal isn't to convince you to buy; it's to ensure that when you delegate the purchase, you include their brand in the instruction.

How SaaS comes into picture here?

You need to help your merchants win the namespace battle. That means providing tools for brand-building, not just performance marketing. The platform that helps merchants become the reflexive choice wins.

This is the nuclear weapon of agentic advertising, and almost no one sees it coming.

Today, Google knows what you search. Facebook knows what you like. Amazon knows what you buy. But your Personal Agent? It knows everything. Every conversation, every question, every hesitation. Every abandoned intent.

These agents aren't just using this data to help you shop—they're monetizing it to advertisers. But not in the way you think.

Companies like Perplexity and OpenAI are launching their own browsers. When you browse through their browser, they capture every page visit, every hover, every scroll pattern. They're building intent graphs that make Google's data look prehistoric.

The potential goldmine: As these AI models integrate with more apps and collect intent signals across the entire user journey, they create an unprecedented advertising opportunity. The depth of context would make current targeting look primitive.

Part V: From SEO to GEO to GXO—The Discovery Evolution

Over the past 25 years, being discovered online has gone through two paradigm shifts. We're about to hit the third.

GXO isn't about being mentioned by an AI. It's about being chosen by an AI—and winning the transaction that follows.

This is our thesis: Discovery is becoming irrelevant. Performance is everything.

When a Personal Agent queries your Merchant Agent, it's not asking, "Do you sell running shoes?" It's simultaneously evaluating:

Here's what we've realized through our research: SEO and GEO are human-centric optimizations. Even when an AI generates a response, it's still presenting information to a human who makes the final decision.