Insights from conversations with the former CEO of LVMH North America and the GM of Agentic Solutions at Microsoft AI
The Moat Has Shifted
When software can be built in a week, the competitive advantage of an enterprise technology company can no longer be the software itself. That was the central tension at the heart of a recent fireside chat between Logicbroker CEO Omar Qari and Anish Melwani, who served as CEO and Chairman of LVMH North America.
The conversation was framed by a challenge Jason Maynard, CEO of Qualtrics, laid out moments before: every enterprise software company must disrupt itself or die. With AI driving the cost of development toward zero, the question Qari put to Melwani was direct: if anyone can build a working product in days for next to nothing, where is the moat?
Melwani’s answer: trust, taste, and judgment.
The argument goes like this. When the cost of building collapses, what remains scarce is the knowledge of what to build. A software partner that works across hundreds of similar companies sees patterns that any single customer cannot. That pattern recognition, applied with genuine creative judgment, produces solutions better than what a customer would design for themselves.
Melwani illustrated it with a tailoring analogy: You can go to Savile Row and commission a fully bespoke suit, specifying the cut, the lapel, the fabric, every button. Or you can walk into Bloomingdale’s and buy an Armani suit off the rack. More often than not, the Armani suit draws more compliments, because the taste and judgment built into its design is more refined than most individuals can produce through customization alone.
For retail technology providers, the implication is that the winners will be the ones who know their customers’ pain more deeply than the customers know it themselves, and who exercise precise judgment in how they solve it.
Melwani tied this to the Tyler Perry model of creative success: not a film for everyone, but an intensely specific, lovingly crafted product for a passionate audience that keeps coming back. Niche depth beats broad coverage, especially in a world where anyone can produce something broadly functional at low cost.
The Buy-Build-Partner Decision Is Being Renegotiated
The “buy, build, or partner” calculus that every organization runs when evaluating technology is being rerun right now across the retail and e-commerce industry, and the math looks different than it did two years ago.
As the cost to build falls, more companies are tempted to build in-house. But Melwani’s framing pushes back on that instinct. The Armani suit costs more than making one yourself, but it reflects accumulated expertise you cannot replicate in isolation. Partnering with a provider who has genuine taste and judgment, who has seen the same problems solved across many similar businesses, produces better outcomes than building from scratch, even when building is cheaper than ever.
The right question is no longer “can we build this?” The right question is “do we know this problem well enough to build it better than someone who has spent years solving it?”
Agentic Commerce by the Numbers
In a separate conversation, Qari sat down with Dave Osborne, GM of Agentic Solutions at Microsoft AI. Osborne shared three statistics that reframe the urgency of the moment for anyone in retail and e-commerce.
Automated web traffic is growing eight times faster than human traffic. According to Microsoft’s AI research, automated traffic grew 24% year over year while human traffic grew 3%. To put the eight-times multiplier in physical terms: it is the difference between driving on a highway and cruising at altitude in a 747.
Traffic from AI agents and agentic browsers grew 8,000% year over year. At 8,000% annual growth, even small numbers become very large ones within 18 months. The crossover point, where agent-generated traffic surpasses human-generated traffic in raw volume, is not a distant projection. It is an imminent operational reality. Gartner projects that agentic AI will handle a significant share of enterprise decisions autonomously by 2028, reinforcing the timeline.
Retail and e-commerce is ground zero. Of all AI-driven web traffic, 95% is concentrated in three sectors: retail and e-commerce, streaming and media, and travel and hospitality. Retail and e-commerce alone accounts for 63% of that AI-driven traffic. No other industry is more directly in the path of this shift.
These numbers align with broader research from McKinsey & Company on AI in retail, which has documented accelerating automation across the purchase funnel, and from Forrester’s commerce research, which tracks the growing role of autonomous agents in B2B and B2C transactions.
Three Eras of the Web, Happening Simultaneously
Microsoft AI frames the current moment as three distinct eras of the web that are now running in parallel, each with a different user intent and a different optimization requirement. Understanding all three is foundational to any modern commerce strategy.
The human web is built around “help me find it.” This is the world of traditional search, SEO, and click-based discovery. Google’s Search Central documentation remains the authoritative resource for optimizing in this era.
The LLM web is built around “help me choose.” Users go into ChatGPT, Microsoft Copilot, Perplexity, or Google’s AI Overviews with semantic queries. They are not browsing; they are evaluating. Optimization here falls under Generative Engine Optimization (GEO), a framework documented by researchers at Princeton, Georgia Tech, and the Allen Institute for AI, requiring structured, factually grounded content that generative models can accurately surface and cite.
The agentic web is built around “do it for me.” Agents are not just finding or choosing; they are transacting, fulfilling, and completing tasks autonomously on behalf of users. Anthropic, OpenAI, and Microsoft have all published research on how agentic systems evaluate and interact with commerce infrastructure. For dropship and marketplace operators, this era introduces an entirely new class of buyer that never loads a product page.
The challenge for retailers and brands is that all three eras are active simultaneously. A customer might discover a product through traditional search, validate it through an AI assistant, and complete the purchase through an agent, all in the same session. Optimizing for one era while neglecting the others creates gaps that competitors will fill.
What Retailers and Brands Should Be Doing Now
Several actionable priorities emerge from both conversations.
- Get your data and operations in order.
Trust is not just about discovery. It is about fulfillment. Agents probe deeper into operations than human browsers ever did. If your inventory data is unreliable, your fulfillment is inconsistent, or your product information is incomplete, agents will surface that unreliability. See how Logicbroker handles real-time inventory and order data to understand what a reliable data infrastructure looks like at scale. - Build for all three eras of the web in parallel.
Your search and discovery strategy can no longer be a single-track SEO program. It needs to account for how traditional crawlers index your content, how large language models represent and recommend your products, and how autonomous agents will evaluate and transact with your catalog. Schema.org structured data markup for product pages is a practical starting point that serves all three eras. - Optimize for generative engine visibility.
Generative engines, including ChatGPT, Copilot, Google’s AI Overviews, and Perplexity, favor content that is factual, structured, and authoritative. Product descriptions, catalog data, and brand content need to be written for machine comprehension as much as for human readers. The GEO research paper from Princeton and Georgia Tech offers a useful framework for how content characteristics influence generative engine citations. - Know your customer’s pain better than they do.
In a world where any solution can be assembled quickly, the differentiator is depth of understanding. Retailers who work with technology partners capable of synthesizing patterns across many similar businesses will build better solutions faster than those going it alone.
The Velocity Advantage
One phrase from the conversation worth carrying forward: tailored velocity. It is the combination of deep customer understanding with the ability to move and adapt faster than any competitor. Bespoke judgment, executed at speed.
That combination, not low cost or feature breadth, is what separates the technology partners who will matter in this next chapter from those who will get commoditized by it.
For retailers and brands standing at the intersection of the human web, the LLM web, and the agentic web, that distinction has never been more important to understand. To see how Logicbroker is building for all three, explore our platform or get in touch with our team.
Omar Qari is CEO of Logicbroker. Bryce Roberts is VP of Marketing at Logicbroker. This post is adapted from their fireside chat series on commerce and agentic technology.
