The Agentic Commerce Frontier 📅 | August 4 - August 10

Welcome back friends. Thanks for reading!

This week in major Agentic Commerce news: a federal appeals court signaling that AI shopping agents may legally count as the user rather than an unauthorized third party; Cloudflare launching programmable wallets for autonomous agents; Shopify reporting a 3x surge in AI-driven traffic and orders; Gravity raising $30.5M to build ad infrastructure for AI assistants; Sumsub and Sumvin linking verified human identity to delegated agent credentials; and ChatGPT surfacing multi-product shopping ads in conversational results.

This week, my primer looks at the human behaviors that merchants relied on to build their pricing, promotions and loyalty offers. And why that might be a trojan horse for Agents.

Thank you again for subscribing, reading, and reaching out to meet and chat!

---

🔥 TL;DR

• A federal appeals court handed Perplexity an important win over Amazon, reversing a temporary ban on its agentic shopping tools. The Ninth Circuit concluded that Amazon was unlikely to prevail on its Computer Fraud and Abuse Act theory because users accessing Amazon through their agents were themselves doing the accessing

  • My takeaway: Very surprised that this news barely made headlines

• Cloudflare introduced programmable wallets designed for autonomous agents, combining stablecoin payments, delegated identity, spending controls, and x402-compatible payment flows

  • My takeaway: The wallet is the side show. The important bit is the off the shelf programmable bit

• Shopify reported that AI-driven traffic and orders each tripled year over year, while 75% of purchases attributed to AI occurred outside merchants’ top 100 product categories. Shopify also said Sidekick had handled 34 million merchant conversations

  • My takeaway: So the long tail of products is favored by AI shopping. Fascinating!

• Gravity raised a $30.5 million Series A to build advertising infrastructure for AI assistants, including tests in which advertiser catalogs and promotions are supplied directly to shopping agents and, with user approval, may ultimately lead to agent-mediated payment

  • My takeaway: In the right conditions, I can see these platforms offering a guaranteed ROI

• Sumsub and Sumvin connected verified human identity with delegated AI-agent credentials, giving merchants and financial institutions a way to associate agent activity with an accountable user

  • My takeaway: So KYC is about to become KTA, “know their agent”

---

🤖 Agentic Commerce Primer:You Made Your Store Agent-Ready. That Doesn’t Mean You Know How to Sell to Them.

Merchants are rebuilding catalogues, APIs and checkout for AI agents. But the commercial rules behind the store were designed around humans, and agents may read those rules very differently.

TL;DR: Agent readiness is becoming a technical checklist: structured product data, real-time inventory, machine-readable policies, agent-compatible checkout and payments. All necessary. But pricing, promotions, loyalty and retention were built around another assumption: the shopper was human. Humans forget, hesitate, tire of comparing, respond to urgency and rarely optimize every transaction. Agents may not. A sale can teach an agent to wait. An abandoned-cart discount can become the real price. A free-shipping threshold can become an equation. Loyalty can turn out to be inertia. The next stage of agent readiness is not making sure agents can shop your store. It is understanding what your store does when they do.

There is now a reasonably clear playbook for becoming agent-ready. Make your catalogue machine-readable. Expose price and inventory. Publish delivery and return policies in formats agents can understand. Support programmatic transactions. Make authentication and payment work when software is acting for the customer.

All of this is sensible, but it answers only one question:

can an agent shop your store?

It says much less about what happens when the agent actually does, because underneath the catalogue, API and payment infrastructure sits something merchants rarely describe as infrastructure at all: decades of assumptions about how humans shop.

Your commercial stack has a human hidden inside it

Physical retail, e-commerce and m-commerce changed where and how we shop, but the shopper remained human. Commerce learned us remarkably well.

We have limited attention, so merchants decide what appears first. We tire of comparing, so they give us bestseller badges, defaults and curated selections. We dislike paying for shipping, so free-delivery thresholds encourage us to add one more item. Shopify explicitly recommends setting those thresholds above average order value to nudge customers toward products they might not otherwise have bought.

We get distracted or hesitate, so abandoned-cart systems bring us back. We dislike checkout friction, and Baymard still measures average cart abandonment at 70.19%, after years of research into the costs, forms and usability problems that cause people to leave. We also form habits. We return because the app is installed, the card is stored, the points are there and comparing alternatives is work.

None of this is particularly sinister. Much of commerce is simply the accumulated result of merchants observing human behaviour and adapting around it.

The shopper became part of the system design.

Now replace some of that behaviour with software. An agent does not necessarily tire after comparing twenty products. It can remember last month’s price, continue monitoring after the consumer has stopped thinking about the purchase, calculate whether adding another product is cheaper than paying delivery, and reopen the market every time something needs replenishing.

Agents are not perfectly rational. They have different limitations: incomplete information, model priors, search budgets, faulty assumptions and whatever rules the consumer has given them. But those are not the same limitations merchants spent decades learning to monetize, and some commercial mechanics may not merely become less effective. They may start producing a different outcome altogether.

Your sale teaches the agent when not to buy

Suppose a product normally sells for $200, and every five or six weeks the merchant runs a promotion at $160.

A human visits during the promotion and sees 20% OFF. Ends Sunday. For the merchant, the urgency is part of the offer. The customer does not necessarily know when the previous discount occurred or when the next one will arrive.

An agent can see something different. It can observe that the product sold for $160 on May 3, June 12 and July 19, and that today’s price is back at $200. If the consumer does not need the product urgently, the rational response may simply be to wait.

Recent research on strategic buying agents models agents that observe prices over a finite shopping window and decide when purchasing now creates more value than waiting. The researchers tested those policies against tens of thousands of historical Amazon price observations.

For a human, the promotion communicates urgency. For the agent, it also provides data.

Humans see a promotion. An agent can see a pricing pattern.

The merchant may discover that repeated sales have trained the shopper never to pay full price. What was designed to accelerate a purchase can become the rule that tells the agent when purchasing makes sense.

Your free-shipping threshold becomes an equation

Now take one of the most familiar e-commerce mechanics. The consumer wants a $43 product, shipping costs $8, and shipping becomes free at $50.

Merchants use thresholds precisely because customers often prefer adding something to the basket over paying for delivery. Shopify recommends positioning the threshold above normal order value so it encourages incremental purchases.

A human might add something useful, pay the shipping or leave. The agent can simply solve the rule. The desired product plus shipping costs $51. If the agent finds a qualifying filler item at $7.25, the basket becomes $50.25 with free shipping.

The threshold worked. Basket value increased. But the additional item is no longer an impulse purchase; it is an instrument used to minimize total transaction cost.

Many retail mechanics depend on consumers optimizing approximately. Agents can optimize them much more precisely.

The same applies to bundles. A “$99 bundle - $140 value” claim means little if the agent decomposes the contents, discovers that the consumer only needs two items and can acquire them elsewhere for $84. The merchandising frame becomes a calculation.

Your recovery discount becomes the real price

Consider a $120 product. The consumer puts it into a basket and leaves, and four hours later receives a message: Still thinking? Here’s 10% off.

The recovery system is doing what it was built to do. Some humans abandon because they hesitate, and a discount can convert demand that would otherwise disappear. Shopify allows merchants to return customers to abandoned checkouts with discounts already applied.

Now give the consumer an agent. The immediate purchase price is $120. The price after waiting four hours is $108. The first time, the agent may simply observe what happened. If the same thing happens again, it has learned something about this merchant.

The posted price may be $120, but the expected price after abandonment is $108. If receiving the product four hours earlier is not worth $12, why complete the first checkout? The agent has not hacked anything. It has simply discovered the merchant’s pricing policy.

The merchant trained the agent.

The recovery discount was meant to change the behaviour of an uncertain buyer. Once an agent remembers the rule, the discount becomes part of the expected price. A discount intended to recover hesitation can become a reward for manufacturing it, and the merchant may eventually have to ask how many agents have learned that its first price is not its best price.

Your loyalty program discovers what was actually loyalty

Now consider a customer who has bought the same coffee from the same retailer every month for six years. The merchant understandably calls that loyalty, but perhaps some of the repeat behaviour comes from something less romantic.

The app is installed. The address is saved. The payment method is stored. The customer already knows which coffee to order. The alternative requires searching, comparing prices, checking delivery and deciding whether saving $3 is worth the effort. Usually it isn’t.

An agent changes the cost of reconsidering the decision. Every month it can compare the usual coffee at $31.99 with an equivalent alternative at $27.50, account for the $1.20 of loyalty value, check that delivery is identical and then decide whether a $3.29 difference is enough to switch.

The point is not that agents destroy brands or loyalty. The consumer can tell the agent, “Always buy this brand unless the alternative is at least 20% cheaper.” That is quite strong loyalty, but it is explicit.

Agentic commerce may expose how much historical loyalty came from preference and how much came from the cost of reopening the decision. For merchants, that changes the objective. The loyalty program may no longer be designed only to make the consumer come back; it may need to give the consumer a reason to tell the agent not to shop around next time.

The agent does not necessarily eliminate loyalty. It makes merchants earn more of it.

Your customer may stop forgetting

There is another human behaviour hiding inside commerce that receives less attention than persuasion: we forget things.

We forget to cancel trials. We miss return deadlines. We fail to notice that a product dropped in price during a protection window. We forget warranty claims or leave loyalty benefits unused. Some businesses depend heavily on these behaviours, while others simply benefit from them at the margin.

An agent can make attention persistent. It can notice that a streaming subscription has been unused for 47 days and renews tomorrow, that an appliance purchased three weeks ago is now $40 cheaper and still inside a 30-day price-protection window, that a return deadline closes Friday, or that an insurance renewal has increased 17% while comparable coverage is available elsewhere.

None of these actions requires extraordinary intelligence. They require memory, persistence and the ability to act at the right moment. Humans are poor at all three when dozens of commercial relationships compete for attention. Software does not have to be.

Some commerce is optimized around attracting attention. Some benefits from the absence of it. Agents can industrialize attention.

Agents will still be influenced

I’m not saying that behavioral commerce disappears as rational agents emerge.

Agents will be influenced too, just by different things. A person may respond to product photography, shelf position, urgency and a crossed-out reference price. An agent may respond to structured attribute completeness, verified delivery, review volume, return policies, merchant data quality and whatever signals its underlying model has learned to value.

Research is already beginning to expose those differences. A 2025 study placed frontier shopping agents inside a controlled marketplace and varied product position, price, ratings, reviews, sponsored labels and platform endorsements. The agents showed measurable but different sensitivities to those signals. The researchers also found that seller-side changes to product descriptions targeted at agent preferences could produce meaningful market-share gains inside the simulation.

Commerce does not necessarily move from persuasion to rationality. It may move from optimizing around human decision architecture to optimizing around machine decision architecture.

That changes the merchant’s questions. Instead of asking only how to make the customer notice a product, merchants may ask how to make the agent confident enough to consider it. Instead of thinking only about how to win the shelf, they may have to think about how to enter the shortlist.

There will still be optimization. The buyer simply behaves differently.

Agent readiness is therefore two different problems

The first is the one the industry is currently solving: technical readiness.

Can agents discover your products? Can they understand the catalogue? Can they verify price and availability? Can they interpret policies? Can they transact and pay?

These are necessary questions, but merchants also need to think about commercial readiness.

What happens to promotion strategy when customers remember every historical price? What happens to loyalty when comparison costs approach zero? What happens to free-shipping thresholds when baskets are optimized mathematically? What happens to cart recovery when agents learn that abandonment produces a discount? What happens to subscription economics when customers no longer forget renewal dates? What happens to a bundle when the shopper decomposes it instantly?

Even the familiar fraud problem belongs here, although it is probably the least interesting example: systems trained to treat automation as suspicious must learn that some legitimate customers now arrive through authorized software.

Being technically prepared for the transaction does not tell you whether the economics surrounding that transaction still work.

Test the behaviour, not just the connection

Imagine a merchant completing an agent-readiness program. The product feed works, inventory is fresh, policies are structured, payment works and the agent successfully purchases a test product. Pass.

I would run another test.

Give 10,000 agents different consumer mandates and let them interact with the commercial system. Some minimize price. Some prioritize delivery. Some strongly prefer the incumbent brand. Some will substitute freely. Some can wait three months, while others need the product tomorrow.

Then let them encounter your promotion calendar, shipping thresholds, bundles, cart-recovery incentives, loyalty mechanics, subscriptions, returns and dynamic pricing.

Do not ask only whether they successfully complete the transaction. Ask what happens to margin, basket composition, discount use, purchase timing, retention and inventory.

The objective is not to prevent agents from optimizing. They are supposed to optimize for the people who sent them. The purpose is to discover whether the commercial rule still creates the response the merchant thinks it creates.

A merchant can look at a system producing exactly the behaviour it was programmed to produce and conclude that everything works. The agent can look at the same system and discover a strategy the merchant never expected a customer to execute consistently.

You made the store ready. Now test the business.

Commerce has adapted to major shifts before. The physical store moved online. The website moved onto the phone. Merchants rebuilt merchandising, advertising, loyalty and checkout around each new interface.

But through those changes, the person doing most of the shopping remained recognizably human.

Agentic commerce is different. The customer behind the transaction is still a person, but increasingly, the behaviour interacting with the merchant may not be.

That is why agent readiness cannot end with catalogues, protocols and payments. A merchant can have immaculate product data, real-time inventory, an agent-compatible checkout and every emerging protocol correctly implemented, and still discover that its sale teaches the customer to wait, its shipping threshold teaches the customer how to extract the cheapest path to free delivery, its recovery system teaches the customer to abandon, its loyalty program reveals how little of repeat purchasing was loyalty, and its subscription model discovers what happens when customers stop forgetting.

You made your store agent-ready. That doesn’t mean you know how to sell to them.

---

🚀 Major Announcements & Funding News

• Perplexity wins reversal of Amazon agent-shopping ban: The Ninth Circuit overturned a temporary restriction on Perplexity’s Comet browser using its agentic shopping functionality on Amazon. The court said Amazon was unlikely to succeed on its CFAA theory because a user employing an agent to access the user’s own Amazon account remained the relevant accessor, a potentially consequential precedent for platform controls over user-authorized agents (Reuters)

• Cloudflare introduces programmable wallets for AI agents: Cloudflare’s new wallet architecture lets developers provision stablecoin-backed wallets to accounts and virtual wallets to individual agents, with controls including spending allowances, transaction limits, and merchant restrictions. The design pairs x402-compatible machine payments with delegated identity so merchants can distinguish an authorized agent from an unidentified automated caller (Cloudflare Blog)

• Cloudflare adds WebMCP interfaces to ordinary websites: A WebMCP developer preview lets site operators expose structured browser-agent actions without building a separate API or changing origin architecture. For commerce sites, the model could make actions such as search, configuration, cart operations, and authenticated workflows more reliable than agents inferring controls from page structure (Cloudflare Blog)

• Cloudflare launches Agent Readiness and Answer Engine Optimization tooling: Agent Readiness evaluates whether AI agents can discover and interpret a site, while Answer Engine Optimization tracks how often AI assistants recommend it. The combination turns agent accessibility and recommendation visibility into observable merchant-distribution metrics rather than purely SEO concerns (Cloudflare Blog)

• Gravity raises $30.5 million for AI-native advertising and agent-to-agent commerce: Lightspeed and Committed Capital co-led Gravity’s Series A, bringing reported funding to $38.5 million. Its stack spans demand, supply, and exchange functions for chatbot advertising, while emerging agent-to-agent formats let advertisers provide machine-readable catalogs and promotions to shopping agents before a possible user-approved transaction (The Next Web)

• Minty brings intent-triggered cashback into Shopify and AI shopping surfaces: Minty made its AI shopping and cashback application available to Shopify merchants, with integrations designed to surface rewards when consumers express purchase intent in conversational environments. Its merchant profiles are positioned for discovery across ChatGPT, Perplexity, Claude, Google AI Overviews, and Apple Intelligence (Digital Transactions)

• Amex Ventures backs Pie’s AI-discovery and voice-agent stack: American Express Ventures made an additional investment in Pie, which helps local merchants improve visibility in AI-driven discovery and uses a Front Desk voice agent to answer questions and handle booking-oriented customer interactions (Digital Transactions)

• Kakao links its Kanana agent to Coupang Eats for in-chat ordering and payment: Kakao’s orchestrator can delegate tasks to Coupang Eats’ service agent, moving from recommendation through authentication, order creation, and Kakao Pay payment without sending the user through a separate merchant interface. The architecture is an early example of commercial agent-to-agent handoffs rather than a single assistant automating every service itself (TechTimes)

• Glassnode makes paid on-chain data callable by agents through x402: Glassnode introduced pay-per-call access to blockchain metrics using USDC and the x402 payment protocol, removing API keys, subscriptions, and conventional signup from the transaction. An MCP-capable agent can discover an endpoint’s price, pay it, and retrieve the requested data autonomously, making the product a concrete machine-to-machine commerce implementation (Glassnode Research)

---

🛡️ Security & Fraud

• Sumsub and Sumvin bind delegated agents to verified human identities: Sumsub verification is being embedded into Sumvin’s agentic credential framework, combining know-your-agent controls with reusable identity. The encrypted credential is designed to let a merchant or financial institution trace an agent’s delegated authority back to a verified, accountable user without repeatedly performing a full identity flow (The Paypers)

• Cloudflare shifts agent defense from one-time risk scoring toward continuous trust: New Agents Week tooling evaluates ongoing automated behavior rather than relying only on an initial bot classification. Components including BotBase and Precursor are intended to distinguish useful agents from systems whose behavior becomes abusive or inconsistent over time (Cloudflare Blog)

• Cloudflare proposes a task-scoped Agent Access Model: The architecture combines identity brokering, continuous mediation, and stateful trust so an agent receives authority for a specific delegated task rather than inheriting broad application access. That pattern maps directly to commerce use cases in which an agent may need permission to compare products and pay up to a fixed amount without gaining unrestricted account control (Cloudflare Blog)

• WriteGuard brings least-privilege controls to MCP write operations: Cloudflare’s private-beta WriteGuard applies fine-grained authorization before an MCP-connected agent can perform state-changing actions. Separating read capability from writes such as purchase, refund, account modification, or message submission is increasingly important as MCP servers become operational interfaces rather than simple retrieval tools (Cloudflare Blog)

• Identity-aware AI Gateway adds per-user and per-agent behavioral baselines: Cloudflare opened identity-aware AI Gateway capabilities in beta and introduced User Insights designed to model typical behavior for individual users and agents and surface anomalous activity. That provides a potential control layer for detecting compromised or over-permissioned agents before their actions propagate into downstream commerce systems (Cloudflare Blog)

---

📈 Consumer & Market Insights

• Shopify says AI-driven traffic and orders tripled year over year: Shopify reported second-quarter revenue of $3.6 billion, up 34%, and GMV of $116 billion, up 32%. More strategically, AI-referred traffic and orders each tripled, 75% of AI-attributed purchases occurred outside merchants’ top 100 product categories, and Sidekick had processed 34 million merchant conversations, evidence that conversational commerce is affecting both demand discovery and merchant operations (Retail TouchPoints)

• ChatGPT advertising expands toward multi-product shopping carousels: Digiday verified placements showing multiple retailer products in a single conversational advertisement at the bottom of a ChatGPT session, following earlier work to automate placements from merchant product feeds. The format brings sponsored assortment presentation closer to the recommendation interface itself rather than relying on a conventional display-ad unit (Digiday)

• Kroger’s AI shopping assistant puts advertising inside the assistance layer: Modern Retail examined Kroger’s decision to embed product-listing ads into its conversational shopping experience. Because the assistant can recommend products, build meal plans, and assemble baskets around constraints such as budget or dietary needs, paid placement is moving upstream into the reasoning and recommendation process rather than being confined to search-result slots (Modern Retail)

• Perion reports 136% year-over-year growth in spend through its Outmax AI Agent: Perion’s August 10 results showed Perion One spend of $156.7 million, up 15% year over year, with retail-media spend up 60% and Outmax AI Agent spend up 136% on a pro-forma basis. Total company revenue fell 5% to $98.2 million, making the agent metric notable as a growth pocket inside a more mixed operating quarter rather than simply reflecting company-wide expansion (Perion Q2 Results)

• Merchant readiness for AI commerce is running ahead of merchant measurement: PYMNTS reports that 61% of surveyed merchants expect AI-generated results to influence purchases more than traditional search results in 2026, 58% expect agents to choose payment methods according to economics such as rewards and fees, and 56% expect autonomous transactions. Yet just 23% said they could clearly identify both AI traffic and resulting purchases, while another 21% could identify agentic traffic without connecting it reliably to completed sales (PYMNTS)

---

🎯 Strategic Hiring Highlights

Product / Platform

• Amazon — Principal Product Manager Technical, Agentic Commerce Experiences — Seattle, WA — $179,900–$243,400 — Amazon / Katchup

• Mastercard — Manager, Product Management, Agentic Commerce — New York, NY — $156,000–$265,000 — Mastercard

GTM / Partnerships

• Adyen — Strategic Growth Manager, Agentic Commerce — San Francisco, CA — $190,000–$240,000 + RSUs — Adyen

• Adyen — Strategic Growth Manager, Agentic Commerce — New York, NY — $190,000–$240,000 + RSUs — Adyen

• Forter — Senior Strategic Partnerships Manager, AI and Agentic Commerce — Remote, US (NY or London preferred) — $180,000–$220,000 + bonus + equity — Forter

Engineering & Architecture

• Cognizant — Agentic AI Commerce Solutioner — Hybrid US (Dallas / Chicago / Atlanta / NY–NJ / San Francisco / Los Angeles) — $220,000–$240,000 + bonus + stock awards — Cognizant

• Stripe — Staff Engineer, Agentic Commerce — Singapore — S$244,000–S$366,000 + equity — Stripe

• Gap Inc. — Principal – Architecture (AI, Unified Commerce) — Columbia, SC — $209,700–$272,600 — Gap Inc

Advisory & Strategy

• Accenture Song — Agentic Commerce & AI Lead – Consumer Goods and Retail — Barcelona / Madrid, Spain — Salary not listed — Accenture

• Accenture Song — Growth Tech & Agentic Commerce Advisory Senior Manager (all genders) — Germany / Austria / Switzerland (multiple locations) — Salary not listed — Accenture

---

📖 Articles Worth Reading

• Agentic payments will need more than wallets: Finextra contributor Tohid Naeem argues that giving an agent a wallet or card solves value transfer but not the harder governance questions: who the actor is, what mandate it has, which limits apply, and how decisions can be audited or disputed. The distinction between payment capability and authorization infrastructure is increasingly central as agents gain transaction privileges (Finextra)

• Payment Fraud Is Moving Upstream, and Visa Just Bought the Data to Follow It: PYMNTS’ August 4 analysis argues that agent-mediated commerce shifts fraud controls earlier in the purchase journey, where behavioral signals, intent, permissions, and interaction context may matter more than a final payment credential alone (PYMNTS)

• Splitit CEO Says Card-Linked Payments Can Save AI Sales: The argument is that conventional BNPL flows can introduce underwriting, redirects, and human interaction that interrupt autonomous checkout, while card-linked installment models can preserve the existing card authorization path. Whether that becomes an advantage will depend on how tightly agent runtimes can integrate credit choice without creating new consent or suitability problems (PYMNTS)

• Knowing My Payment in the agentic economy: This Finextra-sponsored piece proposes a useful trust vocabulary for agentic payments built around explainability, authentication, consent, delegated authority, verification, and validation (Finextra)

• Who is liable when AI goes rogue? Lawyers see new risks: Reuters examines how autonomous-agent behavior complicates negligence, computer-access, attribution, and multi-party liability analysis when actions cannot be cleanly traced to a single human decision (Reuters)

• Deep Dive: Engineering the Agentic Control Orchestration in Banking: Sam Boboev argues that deploying individual assistants is structurally different from running governed agentic operations, where policy, exception handling, model independence, and immutable audit trails need to sit in a centralized control plane (Finextra)

• AI Won’t Replace Brand Operators, It Will Kill Average Ones: Anthony Connelly’s August 10 essay argues that AI is commoditizing baseline ecommerce execution; copy, listings, analytics summaries, ad variants; while increasing the value of operators who can make cross-functional decisions spanning pricing, inventory, creative, advertising, expansion, and margin (Retail TouchPoints)

---

🧭 Looking Ahead

Unpacking the 2026 State of Agentic Commerce Report

• Date: August 19, 2026

• Location: Virtual

• Focus: Salesforce will discuss findings drawn from 3,450 commerce leaders and behavioral data covering 1.5 billion buyers, including AI search traffic, conversational queries, catalog readiness, and retention in agent-mediated commerce Event page

Agentic Commerce is Here: How AI Is Rewiring Retail Decision-Making

• Date: August 20, 2026

• Location: Virtual

• Focus: NIQ and Advantage Solutions will examine how AI-driven decision systems affect merchandising, pricing, promotion, product placement, and growth strategy across retail Event page

eTail Boston 2026 / eTail East

• Date: August 10-12, 2026

• Location: Boston, USA

• Focus: Ecommerce and omnichannel operations, with AI commerce and digital-commerce execution themes

Stripe Tour Sydney

• Date: August 19, 2026

• Location: ICC Sydney

• Focus: Payments, software platforms, AI-commerce tooling, and internet-economy growth

Stripe Tour Singapore

• Date: August 25, 2026

• Location: Sands Expo and Convention Centre

• Focus: Regional payments, cross-border commerce, AI-enabled business infrastructure

W3C / GS1 Workshop: E-commerce for Humans and AI Agents

• Date: September 8-9, 2026

• Location: Zurich, Switzerland / hybrid

• Focus: Standards, interoperability, product data, identity, and agent participation in ecommerce workflows

MRC San Diego 2026

• Date: September 14–16, 2026

• Location: Hyatt Regency Mission Bay Spa and Marina, San Diego

• Focus: Payments, fraud prevention, chargebacks, merchant risk, and digital-trust operations

Agentic Commerce & Payments Summit 2026

• Date: September 15, 2026

• Location: Stockholm, Sweden

• Focus: Agentic AI in commerce and payments, intelligent transactions, automation, fraud prevention, and AI-driven decisioning

Money20/20 USA 2026

• Date: October 18–21, 2026

• Location: Las Vegas, NV

• Focus: Payments, fintech, financial services, connected commerce, banking partnerships, and money-movement infrastructure

---

General information only. Not legal, tax, investment, or professional advice. No warranty as to accuracy or completeness. Verify independently and consult your own advisers.

If you believe any information is inaccurate, please contact AgenticCommerce@proton.me and we will make a good-faith effort to review and correct it where appropriate.

Sponsored or affiliate content, if any, is disclosed.

---

For full transparency, a quick note on how this newsletter comes together:

• I use an agent to help monitor and organize a large daily feed of news, announcements, events, jobs, and industry developments. I still make the editorial calls myself; the value here is in separating meaningful signal from the surrounding noise

• The Agentic Commerce Primer is my original thinking, written and structured by me. I may use AI selectively to polish sentences or improve clarity, but the ideas, argument, and structure remain my own

• The Agentic Commerce Ecosystem Map is researched, assembled, and maintained entirely by hand, with no AI assistance, through the dozens of conversations I have with builders, operators and investors in the ecosystem (whom I thank again for the conversations)