The Agentic Commerce Frontier 📅 | July 21 - July 27

Welcome back folks, and thanks for reading this week’s Agentic Commerce Frontier. The main news this week: Stripe’s reported interest in acquiring model-routing infrastructure, new work on tamper-evident transaction evidence, stablecoin treasury and payment products, and growing concern about the cost and governance implications of deploying agents at scale.

This week, I opine on how the personalized shopping for each person promised by Agentic Commerce may actually translate into a more concentrated market for everyone.

Finally, lots of chatter last week on AI slop. 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)

---

🔥 TL;DR

• Stripe is reportedly pursuing OpenRouter in a transaction valued at roughly $10 billion, potentially combining payment orchestration with a high-volume marketplace for model access and token-based consumption

  • My takeaway: Stripe looked at buying PayPal’s past and OpenRouter’s future. The play is to become the infrastructure through which agents choose models, spend money and transact

• A new autonomous-commerce architecture proposes cryptographically verifiable ordering and fraud evidence across agent transactions, addressing a layer not fully covered by payment authorization protocols

  • My takeaway: Payments are half the problem, there is an emerging theme on who can prove what an agent actually did

• Cross River framed agentic commerce as a banking-primitives problem, emphasizing liability, stablecoin settlement, and merchant-owned financial products rather than treating agents as a front-end feature

  • My takeaway: Stating the obvious

• Stripe is powering new stablecoin payment and account products for Ramp, including 24/7 bank-funded bill payments and accounts that let businesses hold, transfer, and manage stablecoins

  • My takeaway: Crypto bros will bite my head off, but crypto will go mainstream through dull back-office workflows, not flashy crypto apps

• Two-thirds of surveyed U.S. consumers now use AI before making purchase decisions, but shoppers continue to demand visibility and control over recommendations and payment choices

  • My takeaway: This reminds me of Hemingway’s famous line in The Sun Also Rises: '“Gradually and then suddenly”

• Commerce infrastructure teams are beginning to quantify the compute requirements of shopping agents, including latency constraints, concurrency, model routing, and checkout-time capacity planning

  • My takeaway: The fastest shopping agent will often beat the smartest one. Latency is becoming a conversion metric

---

🤖 Agentic Commerce Primer: When Every Shopper Has a Personal Agent, Everyone May Buy the Same Thing

A personal agent may understand you very well. That does not mean it will recommend something different.

TL;DR: Shopping agents can search more patiently, remember more context and understand our needs better than any retailer. But some of the diversity we see in commerce today may come from human inconsistency, imperfect searches and different sources of trust. If millions of agents rely on a small number of models, catalogues and evidentiary rules, richer personal context may still lead to the same shortlists. We could end up with more personalized shopping for each person and a more concentrated market for everyone.

I have been thinking about the heatwaves that moved across France and much of Western Europe this summer. They lasted long enough to change household behaviour. People wanted to know how severe the heat would become, whether they needed a fan or a portable air conditioner, and how quickly anything useful could be delivered.

Now imagine that many of those households had already given an agent permission to monitor the situation.

Yet the practical mandate could be surprisingly similar: it is becoming too hot, how long will this last, and what can cool the home before the worst days arrive?

They would begin filtering. Products that cannot arrive in time disappear first. So do products with uncertain inventory, incomplete specifications, poor reviews or unclear returns. After that process, three or four models may offer the safest combination of price, efficiency and availability.

Nobody copied a neighbour. The agents did not communicate, and no merchant coordinated the purchases. Each system made a reasonable decision for the household it represented.

Even so, orders could cluster around the same products within a short period. Those products sell out, the agents move to the same alternatives, and delivery capacity tightens across the market.

Personalization is not the same as diversity

The natural expectation is that richer personal context will produce a wider variety of choices. Different people have different needs; agents can search deeper for products that fit them; niche products should therefore become easier to discover.

That was my first instinct as well.

But the variety of information entering a decision does not tell us how varied the decision process will be. Two consumers can give their agents very different context while those agents still consult the same merchant feeds, reviews, fulfilment data and reputation systems.

There is another aspect of this that we underestimate. Some of the diversity in human commerce may come from the fact that people are inconsistent.

An agent carrying purchasing authority will probably be more systematic. It needs to explain why a product met the mandate and why the transaction was reasonable. That encourages clear rules and structured evidence.

Consider a washing machine that is exactly the right size and colour, but whose merchant forgot to populate the colour field in the product feed. A person can look at the photograph and see that it fits. An agent may exclude it because an important requirement cannot be verified.

The product did not lose the comparison. It never entered it.

Agents may share the same commercial worldview

There may eventually be millions of agents, but that does not mean there will be millions of genuinely independent decision systems.

A large number of consumer applications can sit on top of a much smaller number of foundation models. The French competition authority recently reported that OpenAI, Google and Anthropic together accounted for more than 84% of users in the agent sector it examined. That does not prove their shopping recommendations will be identical, but it shows why different interfaces should not automatically be treated as independent commercial viewpoints.

Agents need data that is current and dependable. Inventory must be real. Delivery promises must be credible. Product specifications and return policies need to be clear enough for the system to act on them.

A small manufacturer may make a more repairable appliance or offer unusually good local service. A human buyer can learn this through a conversation, a specialist review or a recommendation from someone they trust. If those advantages are absent from the sources an agent uses, the product may be visible online and still invisible to the buying system.

Research on AI search offers a useful, though imperfect, comparison. One large study found that AI search surfaced fewer long-tail sources and less varied responses than traditional search. Another study using real ChatGPT interactions found that generative AI expanded the range of questions people asked while producing less diverse answers than Google for comparable searchable queries.

The first important agent decision may therefore not be which product ranks first. It may be which products are eligible to be considered at all.

A product ranked seventh at least had a chance to compete. A product removed during eligibility checks is absent from the market the consumer experiences.

The shortlist becomes the practical market

No consumer experiences the whole catalogue. Search, advertising and rankings already determine what receives attention. Agents will make that filtering more explicit and potentially more consequential.

An agent may identify ten thousand possible products, remove most of them during eligibility checks, compare fifty, shortlist five and recommend one. The products outside that path remain online, but they have no realistic opportunity to be chosen.

This matters more than ordinary recommendation bias because the human pause can disappear. A person who dislikes a recommendation can remove a filter, continue browsing or abandon the purchase. An agent with a mandate may monitor the conditions and complete the transaction when they are met.

If many agents monitor the same products using similar signals and thresholds, purchases that humans might have spread over several days could arrive within a much narrower window.

How the market begins to reinforce the agents

Agents do not start with a blank sheet of paper. Their models, platform policies and training already contain assumptions about familiar brands, review volume, reliability and acceptable uncertainty.

Eligibility filters then remove much of the market. A missing delivery date, incomplete specification or unclear return policy can eliminate a product before comparison begins. What remains is a smaller set of products that are easy to evaluate and easier to defend: complete data, reliable fulfilment, clear policies and enough transaction history.

When a shared event activates many mandates, that small group receives more orders. More sales create more reviews. Retailers carry more inventory. Delivery performance becomes easier to measure. Those signals reduce uncertainty for the next agent.

The outcome of today becomes an input into tomorrow’s decision.

Research on recommender systems has documented popularity bias and the reinforcement effects it can create. More recent work modelling these feedback loops found that recommendation adoption can sometimes diversify an individual user’s consumption while still amplifying popularity concentration across the market.

The long tail may be searchable but unreachable

Shopping agents should extend that promise. They can search longer, remember unusual requirements and examine products a person would never reach manually.

There is a strong counterargument to everything written so far: perhaps agents will become the best discovery systems the long tail has ever had.

Research on agentic markets suggests that cheaper consumer search can improve learning and consumer surplus. The same work also finds that more informative screening can weaken competition in some conditions, because consumers become better at narrowing the field before purchase.

An independent clothing maker may offer the right fit, materials and repair service, yet lack the structured returns data, inventory feed or review history the agent expects. The product could fit the consumer perfectly while remaining difficult for the system to assess.

Merchants will learn to sell to the median agent

Once agent referrals become material, merchants will adapt. Search engines changed how websites were written. Marketplaces changed packaging, fulfilment and customer service. Payment rules changed checkout design.

Merchants will improve product feeds, stock accuracy and delivery estimates. They will make returns machine-readable and provide stronger guarantees. I would welcome much of this. It should reduce failed purchases and remove ambiguity that consumers have tolerated for too long.

The adjustment may have another effect. Warranty structures, delivery terms and product attributes could become more alike because they fit the fields dominant agents can interpret. Merchants will invest in qualities that are easy to verify and compare. Craftsmanship, comfort, longevity or local service may receive less attention when they are difficult to encode.

The product gradually gets designed for the median agent.

Rational decisions can still produce unstable demand

Human demand contains friction. People hear about an event at different times. They search, hesitate, speak with friends and return a day later. That behaviour spreads purchases over time.

Agents can compress the same process. In a category with scarce stock and a common external trigger, many systems may be watching the same feeds and waiting for similar conditions. They do not need to hesitate once the mandate is satisfied.

Retail inventory, fulfilment and pricing systems were largely designed around human traffic. They may interpret a sudden wave of agent orders as broad, durable demand. A dynamic-pricing system raises the price. Many agents reassess and move to the same second-choice product.

This is most plausible where supply is scarce, products are relatively comparable and timing matters: travel, tickets, emergency household goods or seasonal equipment. It may never be important for most everyday purchases. I am still uncertain about the scale.

What concerns me is not only the possibility that agents make bad decisions. It is that many agents can make sensible decisions using similar information and still create an outcome the market is poorly equipped to absorb.

Some uncertainty may be useful

Recommender systems already face a choice between repeatedly selecting what is most likely to work and occasionally testing a less-certain alternative. The second behaviour is usually called exploration.

Commerce needs a more careful version because an agent is not recommending a song. It may be spending real money, creating a return or damaging the user’s trust.

A consumer might allow more of it where the downside is limited. Mineral water is a simple example. If the requirement is natural water and an unfamiliar brand is substantially cheaper, trying a different pack of six bottles creates little risk. The same policy would be inappropriate for a child’s car seat or an expensive trip.

Agent providers should measure concentration across users, not only satisfaction within each session. Merchants need explanations for exclusion, not only ranking. Retailers should monitor whether agent-driven demand is becoming unusually correlated.

What this changes

For merchants, making a product agent-readable will not be enough. The practical questions are: how often do agents seriously consider us, where do we disappear, and why?

For agent builders, recommendation quality has to be examined at two levels. Does the answer fit the person? What happens when the same system makes similar decisions for millions of people?

For investors, some of the less obvious opportunities may sit in exclusion analytics, synthetic-agent testing and demand-correlation monitoring. I am not yet convinced these form one new category.

For regulators, the question is broader than whether a response is personalized. France’s competition authority has warned that control over the number of sources and offers shown by agents could allow providers to steer demand and reduce meaningful choice.

A different kind of monoculture

We should still believe in the central promise of personal agents. A system that understands the household, the budget and the moment should make many purchases better.

The distinction we have not fully appreciated is between representing one consumer well and producing a healthy market when millions of those representatives act together.

Whether the long tail survives will depend on more than technical access. It will depend on whether merchants have realistic paths into the consideration set, whether agents preserve some room for exploration, and whether the market can learn from unfamiliar products rather than only reinforce what it already knows.

The internet made an enormous range of products available. The conditions necessary for agents to keep that range meaningfully within reach are still an open question.

---

🚀 Major Announcements & Funding News

• Natural raises $30 million for AI-agent payments infrastructure: Natural secured a $30 million Series A to develop payment systems designed for autonomous agents. Its focus extends beyond transaction processing to the harder control layer: linking an agent to a verified principal, defining spending permissions, enforcing merchant and amount limits, and preserving an auditable record of delegated purchases. The company is positioning itself against incumbent payment infrastructure that was built primarily for direct human authorization rather than software acting under a mandate (TechCrunch)

• Stripe reportedly explores an OpenRouter acquisition: Stripe is reportedly discussing an acquisition of OpenRouter at a valuation of approximately $10 billion. OpenRouter aggregates access to multiple AI models and charges a routing margin, creating a business structurally similar to payment processing: metered requests, variable supplier economics, high-volume authorization, and transaction-level take rates. A combination could allow Stripe to route both inference and money for software agents executing commercial tasks (Axios)

• Atoms raises $1.7 billion for industrial AI and robotics: The industrial automation company founded by Travis Kalanick reportedly secured $1.7 billion from investors including Andreessen Horowitz and Uber. Its focus on warehouse and industrial operations places it downstream from digital agentic commerce, where automated purchasing must eventually connect with inventory movement, fulfillment capacity, and physical execution (The Economic Times)

• Cleeng expands its subscription platform with AI agents and global payment capabilities: Cleeng’s first-half product release adds agent-supported retention workflows, broader payment capabilities, and infrastructure intended to help subscription businesses respond to churn and billing issues more dynamically. The development extends agentic commerce beyond acquisition and checkout into recurring revenue management, where an agent may need to diagnose payment failures, select an intervention, and preserve the customer relationship (Cleeng)

• Lianlian DigiTech and UnionPay International partner on AI-agent payments for global procurement: The companies plan to introduce a human-in-the-loop payment system that allows an AI agent to support supplier discovery and initiate cross-border procurement transactions while keeping the business buyer involved in authorization. The model applies agentic execution to a workflow with established purchase mandates, approval hierarchies, and payment controls (Electronic Payments International)

• Pilot Protocol emerges from stealth with $4.5 million in seed funding: Pilot is building infrastructure through which agents can identify one another, establish identity and reputation, communicate, and coordinate economic activity. The investment reflects growing demand for an agent-native service layer that sits between model orchestration and payment execution (Business Wire via Yahoo Finance)

---

🛡️ Security & Fraud

• A verifiable event layer is proposed for autonomous commerce: New research introduces canonical transaction-event schemas, deterministic batching, Merkle commitments, blockchain anchoring, signed fraud markers, and tamper-evident dataset lineage. The prototype reportedly builds a Merkle tree for 50,000 events in 47 milliseconds while keeping inclusion proofs logarithmic in size. The design targets a core operational problem: proving what an agent knew, was authorized to do, and actually did across systems controlled by different parties (arXiv)

• Suspected North Korean operators reportedly used AI-assisted identities to seek access through remote hiring: More than 700 suspicious applicants allegedly targeted a UK bank using polished résumés, technically credible responses, and coordinated identity signals. While this is an employment-security incident, it is directly relevant to commerce-agent governance: machine-generated identities can attack organizations through legitimate workflows, and behavioral correlation may become as important as document verification (The Times)

• Experian and Fastly push agent-verification decisions to the network edge: A July 27 analysis examines Fastly’s participation in the Experian Agent Trust ecosystem, which is designed to evaluate agent identity, delegated authority, intent, and payment credentials before requests reach a merchant’s applications. The architecture could let businesses distinguish authorized commercial agents from scraping, abuse, or impersonation without treating all autonomous traffic as hostile (Sourcing Journal)

---

📈 Consumer & Market Insights

• The operating cost of autonomous agents is becoming less predictable than conventional SaaS: Agents can invoke external models, premium datasets, browser tools, payment services, and downstream agents without a fixed per-seat cost structure. That changes budgeting from license management to continuous, policy-constrained resource allocation and creates pressure for real-time unit economics at the agent, workflow, customer, and outcome levels (TechRadar Pro).

• Agentic commerce increasingly depends on banking infrastructure that can represent delegated authority: Cross River’s July analysis highlights unresolved liability when agents act incorrectly, the role of stablecoins in compressing international settlement times, and the strategic value of merchant-controlled cards and financing. The common requirement is an account layer capable of distinguishing the customer, the agent, the mandate, the merchant, and the authorized transaction boundary (Cross River)

• Zip finds AI has become a mainstream pre-purchase research tool: New research released July 21 reports that approximately two-thirds of U.S. shoppers use AI before purchase decisions, particularly for research and comparison. The findings also indicate that consumers want transparency and continued control, suggesting that recommendation tools may achieve broad adoption before fully autonomous purchasing does (Business Wire)

• Retailers face a new product-discovery optimization layer: Pacvue argues that AI shopping systems increasingly compress discovery, comparison, and recommendation into one mediated interface. That changes the merchant’s optimization target from ranking individual product pages to supplying structured, current, machine-readable information about price, availability, attributes, reviews, policies, and fulfillment (Pacvue)

• Customer service is becoming part of the agentic transaction loop: Gladly’s analysis extends agentic commerce beyond the moment of purchase to exchanges, order changes, delivery problems, returns, and ongoing customer relationships. This matters because a useful commerce agent must preserve context across pre-purchase advice, payment, fulfillment, and post-purchase resolution rather than handing each stage to an isolated system (Gladly)

• AI-generated search is shifting from an optional feature to a default discovery layer: Similarweb data reported by TechCrunch indicates that Google AI Overviews appeared in 43% of searches, compared with 15% one year earlier. Visits to Google AI Mode increased from 126 million in June 2025 to 279 million in May 2026, reinforcing the shift from link-based discovery toward synthesized, conversational product research (TechCrunch)

• Schnucks’ nutrition assistant shows how retailer-owned agents can combine product data with customer objectives: The regional grocer’s forthcoming assistant will provide nutrition guidance, meal ideas, and product recommendations through its rewards app and website. The deployment illustrates an emerging retailer strategy: place an agent inside an authenticated customer relationship rather than ceding discovery and personalization to a general-purpose assistant (Modern Retail)

• Retail competition may shift from winning clicks to influencing an agent’s decision set: A July 27 analysis argues that shopping agents will increasingly compare products, monitor prices, apply payment preferences, and execute orders without sending consumers through conventional storefront journeys. Retailers will therefore need to compete through machine-readable product data, reliable fulfillment information, differentiated offers, and direct agent integrations (Business Standard)

---

🎯 Strategic Hiring Highlights

Product / Platform

• Mastercard — Director, Merchant Cloud Data + Agentic Commerce Strategy — Dublin, Ireland — Salary not listed (Director level) — Simplify

• Checkout.com — Product Manager, Agentic Commerce — London, UK — Salary not listed — Ashby

• Wesfarmers OneDigital — Agentic Commerce Product Manager — Melbourne/Sydney, Australia (hybrid) — Salary not listed — Livehire

• Sana Commerce — Product Manager – Agentic Commerce — Dubai, UAE — Salary not listed — Bebee

• JD Sports Fashion — Agentic Commerce Senior Product Manager (9‑month FTC) — Bury, UK — Salary not listed — Bebee

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

• Amazon — Senior PMT, Agentic Commerce Experiences — Seattle, WA — $151,200–$204,600 base — US Chamber

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

• 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 — Growth Tech & Agentic Commerce Advisory Senior Manager (all genders) — Germany / Austria / Switzerland (multiple locations) — €80,000–€100,000 (estimated range in external listing) — Accenture

• Accenture Song — Growth Tech & Agentic Commerce Advisory Consultant (all genders) — Germany (Munich/Berlin/Frankfurt/Hamburg/Kronberg) — Salary not listed — Accenture

---

📖 Articles Worth Reading

• Building Trust in Autonomous Commerce: A Verifiable Global Event Timeline and AI-Ready Fraud Intelligence Layer: A concrete proposal for cross-system transaction provenance using canonical schemas, deterministic ordering, Merkle commitments, anchoring, signed risk labels, and reproducible data lineage. The design is particularly useful for teams separating authorization, execution, fraud adjudication, and post-transaction evidence (arXiv)

• What’s Behind Stripe’s OpenRouter Move: A useful examination of the strategic similarities between model routing and payment processing. Both businesses aggregate fragmented suppliers, normalize access, meter consumption, manage reliability, and monetize flow rather than owning the underlying product (Axios)

• The Agent Problem Nobody Budgeted For: A practical argument that agent deployment creates a new governance category spanning finance, security, procurement, and engineering. The most important implication is that policy must be enforced at execution time, before agents incur costs or transmit data, rather than reconstructed through monthly reporting (TechRadar Pro)

• TechCrunch Disrupt’s Smart Money Stage: The event preview identifies AI, trust, verification, privacy, fraud prevention, and human oversight as connected financial-infrastructure questions. That framing is directionally correct: autonomous execution makes identity and evidence part of the payment product rather than adjacent compliance functions (TechCrunch)

• Agentic Commerce GPU Infrastructure: Sizing AI Shopping Agents: A technical discussion of infrastructure planning for shopping agents, including concurrent sessions, model size, tokens per request, latency budgets, and peak-load provisioning. Its most useful contribution is treating checkout performance as a distributed inference-capacity problem rather than only an application-design problem (Spheron)

• What Is Agentic Commerce? How AI Is Redefining the Shopper Journey: Pacvue examines how product discovery changes when an AI system synthesizes consumer intent, catalog data, pricing, inventory, and competitive alternatives before presenting a narrower set of choices. The article is particularly relevant to commerce teams deciding how product feeds and retail media strategies must adapt for agent-mediated discovery (Pacvue)

• What Is Agentic Commerce — and Why It’s Not Just About Shopping Agents: Gladly broadens the definition of agentic commerce to include support, fulfillment changes, returns, and relationship continuity. The operational implication is that customer context must remain portable across systems after checkout, especially when an agent is expected to resolve problems rather than merely recommend products (Gladly)

• AI Shopping Agents for Health and Wellness Brands: Alhena AI offers an operator-oriented view of shopping-agent design in a category where recommendations depend on detailed product attributes, customer preferences, recurring purchases, and sensitive claims. The article illustrates why category-specific data models and escalation rules are likely to outperform generic conversational layers (Alhena AI)

• Agentic Commerce on Shopify: What It Is and How It Works: This guide examines how Shopify merchants can prepare storefront data, automation, recommendations, and checkout processes for agent-mediated purchasing. It is most useful as a checklist of the practical integration work required below the conversational interface (AiTrillion)

• Why Agentic Commerce May Settle on Retailer-Controlled Checkout: ShopAppy argues that retailers will resist architectures that surrender transaction control, first-party customer information, and post-purchase relationships to external agents. The strategic tension is not whether agents participate in shopping, but where recommendation authority ends and merchant-controlled execution begins (ShopAppy)

---

🧭 Looking Ahead

Berkeley Agentic AI Summit 2026

• Date: August 1-2, 2026

• Location: Berkeley, CA

• Focus: Agentic AI research, infrastructure, interoperability, governance, and academic-to-industry transfer

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.