As a trusted Shopify Tax Platform partner, Avalara automates U.S. and global tax, from real-time calculation to streamlined tax returns, all with one solution.

Specifically, I want to talk to vendors who are selling agentic AI as if the category is already defined, brands who are buying it as if the ROI is already proven, and the handful of practitioners who are doing neither and have something honest to say. My goal is not consensus. My goal is one or two answers that hold up when you push on them.

Here are the 3 questions I'm walking in with.

One. Is anyone actually reading the room on agentic AI, or are brands being sold snake oil?

The infrastructure announcements are real. Google's AP2 protocol, Visa's IntelligentCommerce, Mastercard's AgentPay, Adyen's co-authorship of agentic mandates — these are not vaporware. But there is a significant gap between "the rails exist" and "merchants are seeing returns." The $20.9B AI-mediated ecommerce figure everyone is citing for 2026 is 1.5% of total retail. That number is being used to create urgency around decisions that will sit in a merchant's tech stack for a decade. I want to know which vendors are being honest about what merchants should actually be doing right now versus what they should be planning for 2028. Those are not the same answer. Anyone conflating them is selling something.

Two. Can LLMs actually be reliably influenced, or is "agentic optimization" just warmed-over spam?

AI traffic to retail sites grew 693.4% year over year. AI referrals converted 31% more than other traffic sources, and on Black Friday specifically, AI conversions were 38% higher than non-AI conversions. The question I want answered: when a vendor tells a brand they can optimize for agent-driven discovery, do they mean they can structurally improve how product data is indexed by AI systems — or do they mean they are going to stuff metadata and hope? The first is a real capability. The second is SEO spam wearing a new hat. I am going to ask vendors to show me exactly which mechanism they are claiming, and then ask them what happens when the LLM updates its weighting model next quarter.

Three. How do organizations actually build AI muscles that improve customer outcomes?

This is the question I hear brands struggle with most, and vendors answer most vaguely. Building AI capability is not a software purchase. It requires clean data infrastructure, cross-functional ownership, and organizational tolerance for being wrong at machine speed. I want to talk to brands that have moved past the pilot stage and can name the specific decisions — not principles, decisions — that made the difference. What did they have to stop doing? What internal resistance did they have to override? What did it cost, in real terms?

These 3 agentic commerce questions do not have comfortable answers. That is exactly why I am asking them.

HomeCare+ is not a home maintenance subscription. It is a loyalty defense play wearing a utility vest.

For $99 a year, Lowe's sends a red-vested associate to your home twice annually to swap air filters, flush water heaters, lubricate garage doors, and replace smoke detector batteries — seven services per visit across more than 75% of U.S. households. The price point is deliberately below friction. The real product is the recurring physical touchpoint inside a customer's home that no app, no algorithm, and no Amazon delivery driver can replicate.

Lowe's has $86B in annual revenue and a loyalty program that has historically rewarded transactions. HomeCare+ is the pivot from transactional to relational — a direct response to a home improvement market where big-ticket spending is compressed and discretionary DIY is soft. The company needs reasons for customers to think of Lowe's before a project starts, not just when they need a drill bit.

The risk is execution at scale. Human-delivered services are operationally expensive, and inconsistency compounds quickly across 1,700 stores.

The question is not whether this is a smart idea. It is whether Lowe's can staff it.

Amazon can ship you a filter. An app can remind you to change it. Neither can send someone who works for the brand into your home twice a year to do it for you. Retention through relevance. That is the play.

Tuesday, March 24, 2026

Wednesday, March 25, 2026