Liam J.E Gerada writes about ecommerce payments and how online stores can serve software agents that shop for people. His posts cover payment costs, subscription billing and checkout systems, and describe Krepling Pay as work on payments infrastructure for ecommerce.
1.7K followersAgentic commerceBased on 10 posts · latest Aug 2026
Most useful for merchant-level payment diagnostics and arguments about how commerce systems should adapt for shopping agents.
Best for
A payments infrastructure brand or conference panel on agentic checkout and payment legibility.
Not for
Consumer lifestyle or mass-market fintech short-form video campaigns aimed at broad audiences.
Key takeaways
A $110 million merchant reportedly paid an effective rate above 4.3% versus quoted 2.9%.
Agents abandon checkouts that assume a human is present, including human-shaped payment forms.
Subscription billing declines can be mistaken for customer cancellations when payment and storefront data sit separately.
Check first: Confirm claims are verified, product ties do not affect work, and request audience demographics and engagement.
Main ideas
The quoted processing rate is not the rate merchants pay · Agents abandon checkouts that assume a human is present · Agent purchasing moves consent from the transaction to the relationship
Liam J.E Gerada writes about ecommerce payments and the changes needed when software agents shop on people’s behalf.
Their posts link practical merchant problems—processing fees that exceed quoted rates, subscription cancellations caused by billing failures, and checkout flows agents cannot complete—to a broader argument that payment infrastructure must work for automated transactions.
They identify Krepling Pay as the product they are building.
Gerada turns abstract payment claims into concrete tests.
In one post, they describe a merchant processing $110 million a year whose effective processing rate was reportedly above 4.3%, despite believing it was 2.9% plus 30 cents.
They offer a simple check: divide total fees by total processed volume.
In another, they explain how agents can abandon checkouts built around human approval, then ask whether a purchase can complete without intervention, product and payment data are machine-readable, and the merchant can distinguish agent traffic.
They also connect these problems to consent and incentives: when an agent buys without a person present, authority must be set in advance, and infrastructure providers may have conflicts if they control both the agent and the checkout.
Their coverage also includes subscriptions: a shorter post explains how billing declines can be mistaken for customer cancellations when payment and storefront data sit in separate tools.
Follow their work for merchant-level diagnostics and broader arguments about how commerce systems should adapt.
The examples are their own claims and have not been independently verified.
Liam J.E Gerada writes about a narrow set of ecommerce payment problems: what merchants actually pay, what breaks when software agents buy instead of people, and where subscription revenue leaks between tools.
Read more
A consistent priority is legibility: making fees and transaction data visible enough to act on. They argue with specific figures and observed failure patterns, then draw a structural conclusion, often closing on Krepling Pay, the payments product described as their own work.
The quoted processing rate is not the rate merchants pay
Gerada argues that the headline rate on a pricing page is marketing, not a merchant’s true cost.
Read more
Cross-border assessments, FX spread, gateway fees, chargeback fees and downgraded interchange all add up, yet most operators cannot quote their effective rate even when they know customer acquisition cost to the dollar. The post works through published fee schedules for Stripe, PayPal, Shopify and Adyen, then gives a one-line diagnostic: divide total fees by total processed volume.
A merchant processing $110M a year believed they paid 2.9% + 30c, but their blended rate was north of 4.3%—about 140 basis points, or roughly $1.54M a year compounding against gross margin. The post applies the same test to Krepling Pay: at a $71 average order value, the 30c becomes 42 basis points, making the real effective rate 2.92% rather than the advertised 2.5%.
Agents abandon checkouts that assume a human is present
Checkout flows are built around getting a person to approve a transaction in the moment, through 3DS, CVV prompts and confirmation screens.
Read more
Autonomous agents stall at the same friction points—interstitial pop-ups, forced account creation and payment forms that presume a human operator—but, unlike people, they do not hesitate. They route to a merchant whose infrastructure they can transact with, turning payment plumbing from a cost centre into a demand-capture asset, Gerada says.
Their team watched autonomous agents attempt purchases across mid-market stores and saw a consistent failure pattern, Gerada says.
The post offers three questions merchants can ask about their own stack: can a transaction complete end to end without a human in the flow; is payment and product data machine-legible rather than buried in a UI an agent must reverse-engineer; and can the merchant distinguish agent sessions from human ones in analytics?
Gerada says almost no merchant can answer the last question.
Agent purchasing moves consent from the transaction to the relationship
Existing controls answer whether a person approved a specific purchase.
Read more
When an agent buys, no one is present at checkout; the question is whether a person authorised that agent to make purchases like this one.
Incumbents are issuing tokens that flag transactions as agent-initiated and adding agentic checkout within their own ecosystems, Gerada says.
But these retain the old shape: a party between agent and merchant, taking a fee and deciding which agents get through.
Card networks issuing agent-initiated tokens and large platforms adding agentic checkout within their own ecosystems are reasonable moves, Gerada says. But they preserve the old structure: a fee-taking intermediary between agent and merchant.
Neutral settlement is argued from conflict and volume
Gerada expects the incumbent middle-party model to fail for two reasons.
Read more
First, conflict: a company that owns both the agent and checkout has an incentive to point the agent at its own inventory and take margin on both sides.
That pushes merchants and agent developers toward a settlement layer with no stake in what gets bought.
Second, volume: a per-transaction toll that is rounding error at human purchase frequency becomes a real tax when agents transact constantly—and real taxes get engineered around.
Gerada describes Krepling as building a commerce platform for the open web that picks no winner.
Its first piece, Krepling Pay, lets a store connect once and become something an agent can find and buy from, whether it runs on Shopify, WooCommerce, Magento or a custom stack.
The post acknowledges the incumbent moves as reasonable, then argues they do not hold.
Subscription revenue leaks in the seams between tools
One-time sales give a clean pass or fail at checkout; subscriptions run for months and lose useful signal when the storefront and processor record different things.
Read more
Involuntary churn—expired cards, declines and fraud flags rather than customer intent—drives a large share of cancellations, but merchants may mistake it for customer intent. Payment-method mix also affects retention and is invisible when subscription cohorts live in one tool and payment data in another.
When a WooCommerce store sees ‘cancelled’ while Stripe sees ‘declined,’ merchants send win-back emails for what is really a billing problem. Gerada puts involuntary churn at 20-40% of subscription cancellations and says Krepling Pay now puts subscriptions, checkout and payments on one stack, so merchants can see where recurring revenue leaks.
It answers the basic question newcomers ask: what actually goes wrong when software buys from an online store?
Read more
Start here for a plain-language statement of the agentic-checkout thesis and a three-question test to apply to any store.
This is the clearest short statement of the problem Gerada returns to: checkout flows are built for people, so autonomous agents stall at the same points humans find annoying — interstitial pop-ups, forced account creation and payment forms that assume a human operator.
The difference is what happens next.
A person hesitates; an agent routes to a merchant whose infrastructure it can use, turning payment plumbing from a cost centre into a demand-capture asset.
Gerada offers merchants three questions to ask about their stack: can a transaction complete end-to-end without a human in the flow; is payment and product data machine-legible, rather than buried in a UI an agent has to reverse-engineer; and can the merchant distinguish agent sessions from human ones in analytics?
The post says almost nobody can answer the third question because the tooling barely exists.
It closes by naming Krepling Pay as the work.
For brands planning a campaign
A payments company pitching agentic checkout can use the three questions as a diagnostic framework for a webinar, sales call or conference panel. The third question — separating agent traffic from human traffic — is the one Gerada says merchants cannot answer, making it a natural entry point for a vendor offering measurement or settlement tools.
It explains why payment pricing is now a strategic question, not just a utility bill, using numbers readers can check against their own statements.
Read more
A good second read after the agentic-checkout thesis, it shows the commercial consequences of machines reading fee schedules.
Gerada starts with a gap many operators face: they can quote customer acquisition cost to the dollar but give an effective processing rate taken from a pricing page.
The post follows a merchant processing $110M a year who believed they paid 2.9% + 30c but was actually paying north of 4.3% once cross-border assessments, FX spread, gateway fees, chargeback fees and downgraded interchange are counted — roughly 140 basis points, about $1.54M a year.
The details make the case: Stripe's domestic rate becomes 5.4% + 30c for an international card with currency conversion, an 86% increase disclosed in a footnote, with billing and tax adding more; PayPal's 3.49% + 49c carries a 3-4% spread over mid-market baked into the exchange rate rather than shown as a line item; Shopify charges 2% on Basic for transactions processed elsewhere, falling to 0.2% at $2.3k/month; Adyen prices honestly on interchange++ but sits behind minimums most merchants never clear.
Gerada argues that opacity is a pricing strategy that works only on humans: agents calculate the true cost per transaction and route accordingly.
The diagnostic is simple: divide last month's total fees by total processed volume.
Gerada applies the same test to Krepling Pay, noting that at a $71 average order value the 30c becomes 42 basis points, making the real effective rate 2.92% rather than the advertised 2.5%.
For brands planning a campaign
This is the strongest single asset for a conference panel on payment legibility: it offers a concrete, checkable consumer-facing metric (blended rate versus quoted rate), a five-minute diagnostic merchants can run and a clear position that legible pricing becomes a liability rather than a moat once software audits the fee schedule. A payments vendor can build a panel around the 140-basis-point example without needing Gerada to endorse a particular provider.
It answers the design question newcomers quickly face: if no human is present, who is accountable for the purchase, and what does consent mean?
Read more
Read it third: it builds on the agentic-checkout premise established in the earlier posts and explores how settlement should be structured.
This post extends the agentic argument into trust and settlement design.
Existing controls — 3DS, the CVV prompt, the confirmation screen — get a human to approve a transaction as it happens.
When an agent buys, nobody is present, shifting the question from whether a person approved the purchase to whether they authorised the agent to make purchases like it.
Gerada expected incumbents to rethink this from first principles; instead, they have mostly added a party between the agent and the merchant to take a fee and decide which agents get through.
Two reasons explain why that model will not hold.
First, conflict: a company owning both the agent and the checkout is incentivised to direct the agent to its own inventory and take margin on both sides, pushing merchants and agent developers towards a settlement layer with no stake in what gets bought.
Second, volume: a per-transaction toll is easy to ignore at human purchase frequency but becomes a real tax when agents transact constantly — and real taxes get engineered around.
Krepling Pay is described as the first piece of a commerce platform that picks no winner, connecting a store once so an agent can find and buy from it across Shopify, WooCommerce, Magento or a custom stack.
For brands planning a campaign
For a panel or thought-leadership piece on payment legibility for autonomous agents, this moves the discussion from fees to governance: who gets to decide which agents transact, and what happens when the same company sits on both sides. It gives a payments vendor a principled case for positioning itself as neutral infrastructure rather than another gateway.
It explains where recurring revenue leaks and how tool boundaries, rather than customer intent, cause the problem.
Read more
Read it last: it is the least connected to the agentic-checkout thread, but shows that the payment-cost and visibility argument also applies to subscriptions.
A shorter post on another part of the stack.
One-time sales give a clear pass or fail at checkout; subscriptions play out over months, with useful signals lost between tools.
Gerada points to involuntary churn — expired cards, declines and fraud flags driving 20-40% of subscription cancellations — and a failure mode in which the storefront sees "cancelled" while the processor sees "declined", so win-back emails target what is really a billing problem.
Payment method mix matters too: auto-updated cards, manually entered cards and wallet tokens retain at different rates, but that split is invisible when cohorts live in one tool and payment data in another.
The post says Krepling Pay now puts subscriptions, checkout and payments on one stack so merchants can see where recurring revenue leaks.
For brands planning a campaign
Useful for a payments company selling to subscription merchants rather than one-off sellers: it turns an abstract integration story into a specific operational failure — win-back spend aimed at a billing decline — and offers a concrete segmentation idea in payment-method-level retention. It also shows Gerada covers recurring revenue, not just agentic checkout.
How the arguments are built: quantified teardown, observed failure, then incentives4
Headline payment rates versus what merchants actually pay1
Checkout built for a buyer who is present1
Consent moves from the transaction to the relationship1
Subscription revenue leaks where billing data and cohort data are split1
Headline payment rates versus what merchants actually pay
Gerada argues that the rate on a processor's pricing page is rarely the rate a merchant pays.
Read more
In one case, a merchant processing $110M a year believed it paid 2.9% + 30¢, but cross-border assessments, FX spread, gateway fees, chargeback fees and downgraded interchange pushed the blended rate above 4.3% — 140 basis points, which they put at roughly $1.54M a year compounding against gross margin.
They walk through the published mechanics of Stripe (1.5% added for international cards, 1% for currency conversion, plus billing and tax add-ons), PayPal (cross-border fee plus a 3–4% spread baked into the exchange rate rather than shown as a line item), Shopify (2% on the Basic plan for transactions processed elsewhere) and Adyen (interchange++ behind minimums most merchants never clear).
Their conclusion is that opacity is a pricing strategy that only works on humans, and that software initiating transactions will compute true cost per transaction and route around it.
The post ends with a five-minute diagnostic: divide total fees by total processed volume on last month's statement.
Cost teardown using one named merchant case and published fee schedules, followed by a self-run diagnostic calculation.
Adyen is presented as the one provider that prices transparently, but its minimums put that legibility out of reach for most merchants. The alternative Krepling Pay rate of 2.5% + 30¢ in the same post is a product claim, not an independently verified price.
For brands planning a campaign
Directly usable for a conference panel on payment legibility for autonomous agents: the blended-rate versus headline-rate argument, the four-processor comparison and the statement diagnostic are all concrete panel material. A brand should decide up front whether the discussion is vendor-neutral or includes their own Krepling Pay positioning, since the post argues the problem and advertises the fix in the same breath.
Gerada's central claim is that online checkout trusts a human to approve each transaction at the moment it happens — 3DS, the CVV prompt and the confirmation screen all exist for that.
Read more
When an agent buys instead, nobody is there to say yes.
They say the team watched autonomous agents attempt purchases across mid-market stores and found a consistent failure pattern: agents stall where humans do — interstitial pop-ups, forced account creation, payment forms that presume a human operator — but unlike humans, they abandon without hesitation and route to a merchant whose infrastructure they can transact with.
They frame this as inverting twenty years of competitive logic: stores competed on ads, SEO and top-of-funnel conversion, while agents evaluate whether a transaction can complete.
That turns payment infrastructure from a cost centre into a demand-capture asset.
The post closes with three questions for a merchant's stack: can a transaction complete end to end with no human intervention; is payment and product data machine-legible or buried in UI an agent must reverse-engineer; and can you distinguish agent sessions from human ones in analytics, which they say almost no merchant can answer?
Observed agent purchase attempts across mid-market stores, turned into a three-question stack diagnostic.
Gerada attributes the inability to separate agent from human sessions to missing tooling rather than operator failure.
For brands planning a campaign
The three questions are ready-made prompts for a panel on what stores must change for agent-initiated purchasing, and the 'agents don't hesitate' framing gives a payments brand a sharp opening claim. The agent observation comes from the creator's team; ask for supporting data if the panel needs evidence beyond firsthand experience.
Consent moves from the transaction to the relationship
Gerada extends the agentic-checkout argument: when an agent buys, the question stops being 'did this person approve this purchase' and becomes 'did this person give this agent authority to make purchases like this one.' They note that card networks are issuing tokens that flag a transaction as agent-initiated and large platforms are adding agentic checkout inside their own ecosystems, but argue both keep the old shape — a party sitting between agent and merchant, taking a fee and deciding which agents get through.
Read more
They expect that shape to fail for two reasons: conflict, because a company that owns both the agent and the checkout has an incentive to point the agent at its own inventory and take margin on both sides; and volume, because a per-transaction toll that was rounding error at human purchase frequency becomes a real tax once agents transact constantly and gets engineered around. The post describes Krepling as building a commerce platform for the open web that picks no winner, with Krepling Pay as its first piece: connect once and a store becomes something an agent can find and buy from, on Shopify, WooCommerce, Magento or a custom stack.
Structural incentive argument built on conflict and volume, rather than a cost teardown.
The neutrality claim is Gerada's product position; the post acknowledges that incumbents' moves are 'reasonable' before arguing they do not hold.
For brands planning a campaign
Useful for a panel segment on whether agentic payments need a neutral settlement layer, and for a campaign explaining where transaction volume breaks per-transaction fee models. Since Gerada is arguing for their own company's position, a brand should present the conflict and volume arguments as a thesis to test with other panelists, not as settled industry consensus.
Subscription revenue leaks where billing data and cohort data are split
Gerada contrasts one-time sales, which give a clean pass/fail at checkout, with subscriptions, where they say most useful signal is lost in the seams between tools.
Read more
Involuntary churn — expired cards, declines and fraud flags rather than customer intent — drives 20–40% of subscription cancellations, according to them.
When the WooCommerce store records 'cancelled' while Stripe records 'declined,' merchants end up sending win-back emails for what is actually a billing failure.
Payment method mix compounds the problem: auto-updated cards, manually entered cards and wallet tokens retain at different rates, but that split is invisible when subscription cohorts live in one tool and payment data in another.
The post closes by describing Krepling Pay as putting subscriptions, checkout and payments on one stack so merchants can see where recurring revenue leaks.
Separating voluntary from involuntary churn and treating payment-method retention as a hidden variable rather than a single subscription metric.
The win-back campaign is the concrete example of what goes wrong when a billing decline is misread as a churn signal.
For brands planning a campaign
For a subscription-billing or retention brief, this is the clearest available statement of the problem Gerada works on; the involuntary-churn framing translates directly into a webinar or panel segment. It is also the most product-led of the four posts, so it works better as a demonstration of the problem than as vendor-neutral commentary.
How the arguments are built: quantified teardown, observed failure, then incentives
Across the four posts, a consistent method emerges.
Read more
Gerada opens with a specific operator-level claim or number (a $110M merchant's blended rate, a team's agent purchase attempts, a 20–40% involuntary-churn range), works through the mechanics with named providers or failure points, then generalises to a structural claim about where the market is heading and closes on Krepling Pay as the response.
The subjects are narrow and deliberate: payment cost legibility, agent-initiated checkout, neutral settlement and subscription billing.
The posts are text-only LinkedIn essays, with no visuals or video referenced.
Numbers first, then mechanism, then a market-level implication, then a product tie-in.
The teardown style means claims arrive with figures attached. Ask for supporting data behind the agent-testing observations if the brief requires published evidence.
For brands planning a campaign
A brand booking Gerada for a panel should expect a quantified argument rather than inspirational commentary, and allow time for the fee breakdowns and diagnostics, the most portable parts of the material. Since every post ends on the same product, brands should decide whether the segment is a neutral explainer or a vendor-framed position.
Gerada's posts describe one business relationship: Krepling Pay, the payments product described with “we built” and “we're building” language.
Read more
Other brands named in the posts, including Stripe, PayPal, Shopify and Adyen, appear as examples in payment-cost analysis rather than as disclosed sponsorships, clients or partnerships.
Krepling Pay
Owned business
The author writes as part of the team behind Krepling Pay, using “we built” and “we’re building” to describe the payments product and linking to its pricing page.
Read more
Ask for confirmation of their formal title and ownership stake.
Gerada's material is narrow and practical: what payment processing actually costs a merchant, why checkouts built for a human buyer break when software agents buy, and where subscription revenue leaks when billing data and subscription data sit in different tools.
Read more
Start with the fee teardown post, which works through published Stripe, PayPal, Shopify and Adyen pricing and ends with a five-minute diagnostic readers can run on their own statement.
Follow with the two agentic-checkout posts on consent moving from the transaction to the relationship and on neutral settlement.
The subscription post is the most product-led of the four and works best as a demonstration of the involuntary-churn problem.
The creators below cover the same agentic-commerce shift from different positions: one focuses on retention and payment-failure signals in subscription businesses, while two explain the emerging agent payment standards for retailers.
Burke writes about the same failure as Gerada's subscription post: a declined or failed payment being misread as a customer decision.
Read more
Burke puts the cost at 5-7% of ARR lost annually to involuntary churn in B2B SaaS, describes monitoring account health before billing events and flagging at-risk accounts 30-60 days before renewal, and argues in a post aimed at finance leaders that many NSF declines stem from payment timing and issuer behaviour rather than a customer liquidity problem.
Burke focuses on account health and retention inside a subscription business, measuring recovery rate. Gerada approaches the same leak from the payment stack, arguing that subscriptions, checkout and payments should run on one system so a declined card is not recorded as a cancellation.
Gai covers the same shift Gerada argues for from the payments side: agents beginning to buy rather than just help shoppers discover.
Read more
She lays out the competing agent payment protocols, including Google's Agent Payments Protocol, Mastercard's Agent Pay, Visa's Trusted Agent Protocol and PayPal's agentic commerce platform, and describes the layers of agentic commerce and its move through eBay, Amazon, social commerce and short video.
Gai addresses retailers who need to become agent-ready, focusing on storefronts, APIs and product data that agents can understand and transact with. Gerada focuses less on readiness and more on the economics: headline pricing is illegible to machines, and whoever sits between the agent and the merchant has a conflict of interest.
Shivers explains the agentic commerce standards that Gerada's posts assume, including OpenAI and Stripe's Agentic Commerce Protocol and Google's Universal Commerce Protocol.
Read more
Both write for retailers trying to work out how agents will reach them, and both argue that the structured product feed, not the website, now determines whether a sale is possible.
Shivers' posts are standard-by-standard explainers with implementation checklists: provide a structured feed, implement checkout endpoints, support delegated payments, and fix product feeds and schema.
Gerada focuses on the money question, showing what a merchant actually pays and arguing that neutral settlement matters.
Ask for more detail on how a retailer implements a given protocol.
Liam J.E Gerada's LinkedIn audience is small, with 1,681 followers observed on 8 October 2026.
Read more
Their posts target ecommerce operators and payments decision-makers, discussing effective processing rates, subscription billing leaks, and autonomous-agent checkout.
That makes them a plausible content-fit candidate for a payments infrastructure campaign or a conference panel on payment legibility for autonomous agents, but audience fit is unverified: the posts provide no data on followers’ roles, company sizes, geographies, or buying authority, and include no impressions, engagement rates, or audience survey results.
Ask for LinkedIn audience demographics (industry, seniority, geography) or a recent audience survey, plus engagement data for posts on comparable payment and checkout topics over the last six months.
Without that evidence, treat audience reach as narrow and rely on the content argument.
Gerada’s posts contain no sponsorship, affiliate or paid-partnership offer, and no media kit, rate card or partnership terms page.
Read more
The commercial activity shown is their own company: posts use “we” and “our team” to describe building Krepling Pay and direct readers to its pricing page.
That is a product they are behind, not a sponsorship format a brand can book.
The posts do not say sponsorships are never accepted; they simply publish no terms.
A brand considering them for a panel or paid placement should ask about deliverables, disclosure, exclusivity, lead time and rates, and confirm terms before committing spend.
A payments infrastructure company wants to understand the shift toward agentic checkout and book a conference panel on payment legibility for autonomous agents.
Read more
Decision: strong content fit, with two conditions to settle before booking.
Reasons: Liam J.E Gerada's four posts are already on this subject.
One argues that checkout flows assume a present human, while agents stall at interstitials, forced account creation and human-shaped payment forms before routing to stores whose infrastructure they can transact with.
It gives three stack questions, including whether a merchant can distinguish agent sessions from human ones.
Another examines a merchant processing $110M a year whose blended rate ran north of 4.3% against a quoted 2.9% + 30c.
It breaks down Stripe, PayPal, Shopify and Adyen fee mechanics, then offers a five-minute diagnostic: divide total fees by total processed volume.
A third argues that consent moves from the transaction to the relationship, and that a per-transaction toll becomes a real tax when agents transact constantly, pushing toward a neutral settlement layer.
The fourth covers subscription billing leaks where storefront and processor record different reasons for the same cancellation.
Together, these offer four distinct panel-ready segments with named providers, figures and a testable merchant exercise—not general commentary.
Conditions: the posts are product-led—all four close on Krepling Pay, their own payments product—so settle whether the panel is vendor-neutral or explicitly framed as a product position.
The agent-purchase observations and merchant example are reported firsthand, not independently verified; present them as this creator's account.
Reach is narrow: the only audience figure published in the posts here is 1,681 LinkedIn followers, observed 8 October 2026, with no engagement, demographics or composition data.
The partnerships section lists no sponsorship, media kit or rate card; ask for paid panel terms directly.
When to choose a specialist
A consumer lifestyle brand or mass-market fintech app wants a short-form video campaign across TikTok and Instagram, aiming at a large general audience and measurable sign-ups from broad demographics.
Read more
Decision: poor fit based on what Liam J.E Gerada's posts say.
Reasons: all four posts are text-only LinkedIn essays, with no video, visuals or short-form format.
The subject matter is narrow and operator-facing: payment cost legibility, agent-initiated checkout, neutral settlement and subscription billing leaks.
The writing targets ecommerce and payments practitioners, not a general consumer audience.
The only audience figure published in the posts here is 1,681 LinkedIn followers on 8 October 2026; no engagement rates, impressions, demographics or platform mix indicate mass reach.
The posts mention no sponsorship, affiliate or paid-partnership offer, and publish no media kit, rate card or deliverable specification.
To assess the campaign, ask for short-form video examples, disclosure practices, lead times and performance data from another channel and format.
A worked campaign example
Worked campaign angle — conference panel: 'Who pays the toll when software buys?' Format: 45-minute panel plus a 15-minute audience diagnostic.
Panelists: Gerada; a neutral payment infrastructure operator; a merchant with recurring revenue; moderated by the payments brand's own content lead.
Opening segment (10 min): frame the problem in Gerada's terms—checkout trusts a present human, agents stall where people stall, and unlike people they abandon and route to infrastructure that can complete the transaction.
Anchor the discussion on three stack questions: can a transaction complete end to end without human intervention; is payment and product data machine-legible; can the merchant separate agent sessions from human sessions?
Middle segment (15 min): cover payment legibility using the $110M merchant example.
Put the quoted rate and blended rate on screen, then walk through the fee mechanics for Stripe, PayPal, Shopify and Adyen.
Close with the five-minute diagnostic and ask the merchant panelist to run it live against a sample statement.
Governance segment (10 min): explore how consent moves from transaction to relationship, along with two structural arguments: conflict when one company owns both agent and checkout, and volume when a per-transaction toll that was a rounding error at human frequency becomes a real tax.
Test neutrality by asking each panelist who should be allowed to decide which agents get through.
Product-handling: Gerada's posts close on Krepling Pay.
Agree beforehand whether the panel is a neutral explainer or a vendor-framed position, and disclose the product interest on screen if discussed.
Treat the agent-purchase testing and merchant example as this creator's reported firsthand account, not audited industry data.
Take-away asset: a one-page checklist combining the three stack questions with the statement diagnostic, usable by merchants after the session.
Why this works as a campaign angle: it gives the payments brand a practical, numbers-first panel with a live exercise, draws on the creator's reusable material, and turns a product-led corpus into a segment that can be framed neutrally or explicitly, depending on the brand's disclosure choice.
Choose Burke when the panel centres on subscription revenue retention rather than agent-initiated checkout: preventing involuntary churn, spotting at-risk accounts before renewal and treating failed payments as a payment-performance problem rather than a customer-liquidity problem.
Read more
Burke’s posts give the payments infrastructure company a practical angle on retention and decline diagnosis for a merchant or CFO audience.
Burke’s posts focus on B2B SaaS account optimisation and NSF/decline patterns, not agentic checkout, agent-initiated payments or fee transparency.
They offer a sharper retention and billing-operations discussion, but lack Gerada’s merchant statement example and blended-rate breakdown.
Choosing Burke means trading the agent-commerce framing for that focus.
Choose Gai when the audience is retailers who need the agentic commerce landscape explained: which protocols are emerging, how AI agents discover and buy products, and what brands should do to become agent-ready.
Read more
Gai’s posts name Google’s AP2, Mastercard’s Agent Pay, Visa’s Trusted Agent Protocol and PayPal’s agentic commerce platform, and connect agentic shopping to answer-engine discovery. That makes Gai a better fit for a retailer-facing panel on protocol readiness and discovery than for one built around merchant payment-cost diagnostics.
Gai’s material offers broader education on agentic commerce and retail readiness, with less detail on fee mechanics, blended processing rates and the merchant statement exercise.
It lacks Gerada’s argument for neutral settlement and discussion of conflicts and volume problems in agent payment intermediation.
Choosing Gai means a more accessible overview of protocols and discovery, but a less practical session on payment legibility and who pays the toll when software buys.
Choose Shivers when the panel needs implementation detail on agentic checkout standards.
Read more
Shivers compares OpenAI and Stripe’s Agentic Commerce Protocol with Google’s Universal Commerce Protocol, explains delegated payments and cart building inside ChatGPT, and lists concrete retailer actions such as structured product feeds, ACP checkout endpoints and feed-readiness checks. That makes Shivers a stronger choice for a working session on preparing merchant data and checkout infrastructure for agent-led purchasing.
Shivers’s posts give less attention to processing-rate opacity, cross-border fees and the merchant statement diagnostic at the heart of Gerada’s legibility argument.
They also give less attention to consent moving from the transaction to the relationship and to the case for neutral settlement.
Choosing Shivers means a more technical protocol and feed-readiness panel, but a thinner discussion of fee transparency and the commercial incentives around agent payments.
This profile rests on four complete LinkedIn posts by Liam J.E Gerada, all text-only and concerning ecommerce payments, agent-led checkout and subscription billing.
The posts carry no publication dates, so their order and any development over time cannot be established; internal references such as "last month" and "over the past quarter" are the only timing signals.
The one audience figure available is 1,681 LinkedIn followers on the profile, observed 8 October 2026, with no engagement, impressions or demographic data.
Similar-creator comparisons draw on two posts each by Jeremy Burke, Sharon Gai and Tom Shivers.
No interviews, videos, transcripts, third-party coverage or sponsorship terms appear in the posts here.
Signature work such as books, shows or earlier career projects that would explain searches for this name does not appear in the posts here, so the overview treats it as a gap rather than a complete account.
Merchant figures, churn ranges and agent-testing observations are reported as this creator's own claims, not independently verified facts.