|

E-commerce Performance & Innovation Club (EPIC) Monthly News (April 26)

April 2026 was the month e-commerce stopped treating AI as a future bet and started proving it in practice. American Express launched the first protection scheme for AI-agent purchases (the ACE Developer Kit), with Visa announcing parallel intent the same week. Walmart turned its retail-media network into the headline story of the month, with three separate moves on first-party data, streaming-TV ads and small-business advertisers. Anthropic and Amazon committed to five gigawatts of new computing capacity. UK retail giant Tesco started a colleague-AI trial. The FIDO Alliance, the body that runs passkeys, opened work on agentic-commerce standards.

The name EPIC began as our Ecommerce Performance and Innovation Conference, held on 11 June 2025 at the National Space Centre in Leicester to mark Anicca’s 18th birthday. Over a hundred retailers, brands and tech leaders joined us for a day on AI, Shopping Ads, Retail Media, Amazon and TikTok Commerce. We are reusing the name for this new monthly news roundup, the Ecommerce Performance and Innovation Club, which brings the same audience a regular read-out of what is moving in e-commerce. Thirty-seven stories this month across seven sections, from macro forecasts to the regulators setting the rules of the agentic commerce era. We will be running this on the last working day of every month, and if the news flow keeps up, we will move to a weekly cadence later in the year.

Five macro data points UK and European commerce teams typically reference for 2026 board packs. Sprout Social on social-commerce trends, eMarketer’s AI Commerce thesis, US single-digit growth normalisation, the NRF US forecast, and Brazil marketplace dynamics for any UK brand eyeing Latin America.

2026 Social Media Ecommerce Trends & Statistics: The Ultimate Guide

Sprout’s annual social-commerce report shows 47 per cent of consumers have made a direct in-platform purchase in the past year. TikTok Shop leads on under-25 conversion rates; Instagram leads on average order value; Pinterest leads on category research and high-intent click-through. Social discovery now precedes search for 38 per cent of Gen-Z buyers. The report also flags creator-led commerce as the single fastest-growing acquisition channel for fashion and beauty in 2026. Influencer-driven sales now account for 17 per cent of TikTok Shop GMV.

Why it matters: If your acquisition setup still treats social as awareness-only, the conversion data here makes the case for integrated shop infrastructure on Meta, TikTok and Pinterest in 2026, not 2027. The Gen-Z social-before-search behaviour is particularly important: under-25 SEO investment will give diminishing returns. Brand teams should also pressure-test their creator strategy against the 17 per cent influencer-GMV figure as a benchmark.

The AI Commerce Opportunity

eMarketer’s thesis on where AI will move the dial in commerce in 2026. Top categories: personalisation at scale (the largest near-term ROI), conversational discovery, automated merchandising and pricing, and brand-new attribution models for assistant-led purchases. The report flags measurement as the single biggest near-term gap. eMarketer also draws a line between operational AI (already paying back) and agentic AI (still pre-revenue for most retailers in 2026). The thesis is intentionally skeptical of the hype cycle.

Why it matters: Useful framing if you need a one-page argument for the AI investment case in a board pack. The framework also gives you a defensible counter to the “invest everywhere in AI” reflex: personalisation and merchandising are where early ROI shows up, agentic checkout is realistically a 2027 story for most non-Amazon retailers. Use this to triage your AI roadmap into “invest now” versus “watch closely.”

2025 U.S. ecommerce sales mark fourth straight year of single-digit growth

US Commerce Department data shows online retail grew 8.7 per cent in 2025, the fourth consecutive year of single-digit growth. Pre-pandemic, online routinely grew 14 to 16 per cent year-on-year. Penetration is now 22.7 per cent of total US retail sales, which Digital Commerce 360 notes is high enough that the law of large numbers takes over. Categories like apparel and electronics are now over 35 per cent online, while groceries are only 14 per cent. Analysts cited in the piece argue category-by-category penetration ceilings are now visible.

Why it matters: The growth-rate normalisation is the single most under-priced trend in commerce strategy decks. Boards still working from a 2021-era assumption of double-digit ecom growth are budgeting incorrectly. The actionable insight: stop expecting category-wide growth to bail out a weak operating plan, and shift focus to share-take strategies (search-coverage, retail-media saturation, marketplace optimisation) where the realistic upside lives.

NRF is forecasting U.S. retail sales to grow 4.4% in 2026

The US National Retail Federation forecasts retail sales growth of between 3.7 per cent and 4.4 per cent in 2026, up from a final 4.0 per cent print for 2025. Online and non-store retail is expected to grow 7 to 9 per cent, more than double the bricks-and-mortar pace. NRF’s chief economist Jack Kleinhenz attributes the bullish view to wage growth, low unemployment and easing inflation, but flagged proposed import tariffs as the single biggest downside risk for the second half of the year. The forecast assumes consumer-confidence indicators stabilise around current levels.

Why it matters: This is the headline number UK and EU teams use to size their US plans. 4.4 per cent at the top end is healthy but not booming. Crucially, it is slower than the AI-narrative would suggest, which means revenue growth will mostly come from share-of-wallet battles and new-channel expansion rather than category-wide growth. Plan for incrementalism: budget allocators who model 2026 on a 6 per cent assumption will overshoot.

Brazil Ecommerce Market Shares 2026

Brazilian retail e-commerce will add US$22.54 billion in sales over the next two years, with Mercado Libre and Amazon together capturing more than half the total gain. Local marketplace Shopee is growing share against Magazine Luiza and Americanas, both of whom have struggled with category-share erosion since the pandemic peak. eMarketer also notes that cross-border merchants face tighter import-tax rules from the Brazilian Receita Federal, which will compress margin for low-AOV imports. Mobile commerce is now 71 per cent of all Brazilian online retail.

Why it matters: Latin America is the highest-growth ecom region in 2026-27, and Brazil is the dominant market by volume. UK and EU brands eyeing Latam expansion should plan around Mercado Libre integration first, not last, because the platform-routing decision determines almost everything else (logistics partners, payment rails, ad-network access). The cross-border tax tightening also matters for any UK DTC brand currently shipping direct to Brazilian consumers.

2. Big brands & retailers

Etsy, Walmart, JD Sports, THG, Mountain Warehouse and Titan all moved this month. The most striking story is Etsy’s new CEO calling for “really great human service” from sellers, the most direct counter-narrative to agentic commerce we have heard from a marketplace boss this year. Walmart’s annual report puts e-commerce as the company’s biggest growth driver and its retail-media business as a stand-alone profit centre.

JD Sports chair quit after failed push to remove CEO – reports

Andy Higginson resigned as JD Sports chair in late April after the board failed to back his move to remove CEO Régis Schultz, according to Financial Times reporting. JD Sports has been under continuous pressure on the pace of its US expansion and on digital-channel performance, particularly mobile conversion. Schultz remains in post and a chair search is reportedly underway. Higginson is the second high-profile UK retail chair to step down in a publicly fractious split this year. JD Sports’s share price was largely flat on the news.

Why it matters: Watch JD Sports’s digital strategy through the next two quarters. When boards split publicly over CEO direction, strategic decisions stall, capital decisions get pushed, and operational momentum drops. Worth tracking whether the next chair has digital-first or operations-first credentials, because that signals which way the board will lean on the central tension between physical-store-led growth and DTC investment.

THG reports ‘strong start’ to year as beauty sales rise

Online retail group THG reported its strongest Q1 sales growth since the pandemic, led by the Beauty division up 9.6 per cent year-on-year. Nutrition (Myprotein) was broadly flat. THG Ingenuity, the company’s commerce-platform-as-a-service business, also grew, with new clients reported in the period. CEO Matthew Moulding said the company’s full-year guidance has been “modestly raised” on the back of the Q1 trade, and that cost-base improvements should flow through to second-half margin. Group-level revenue grew 5.1 per cent overall.

Why it matters: Positive print for UK online retail, useful as a counter to the “everyone is struggling” narrative still circulating in board packs. THG Ingenuity is also worth watching as a UK-headquartered alternative to BigCommerce or Shopify Plus for ent-tier merchants who want a UK-data-resident commerce platform. The Beauty 9.6 per cent number is genuinely strong against Sephora’s flat US comparable.

Mountain Warehouse swaps decade-old ecommerce system, launches on BigCommerce

UK outdoor retailer Mountain Warehouse launched a composable, headless ecommerce store on BigCommerce, replacing a decade-old custom system that had become a barrier to mobile-first design changes. The new stack uses BigCommerce Catalyst (the company’s Next.js-based React framework) with Algolia for search and merchandising, and Klevu for product recommendations. Mountain Warehouse cited speed-to-market for AI experimentation and mobile-conversion improvement as the main drivers. The cutover took 14 months end to end. Initial KPIs reported: 22 per cent faster page loads and 8 per cent uplift in mobile conversion in the first four weeks.

Why it matters: A sizeable UK retailer publicly choosing BigCommerce composable over Shopify Plus or Magento is a notable platform vote. If you are running a re-platforming RFP in 2026, BigCommerce Catalyst now belongs on the shortlist alongside the obvious options. The 14-month timeline is also useful as a benchmark for any board paper that under-estimates re-platforming complexity. Worth requesting a Mountain Warehouse case study reference.

Etsy CEO calls on sellers to refresh inventory, provide ‘really great’ human service

Kruti Patel Goyal, who took the CEO role in January after a 15-year career at Etsy, used the Q1 earnings call to ask sellers to refresh stale listings and double up on human customer service. Etsy returned to GMV growth in Q1 2026 after seven consecutive flat or declining quarters. Goyal positioned Etsy’s seller-led service as the company’s defence against agentic shopping: the argument is that AI agents pick on price and feature-fit, while Etsy buyers come for human craftsmanship and seller-direct conversation. The platform also rolled out a new seller-coaching tool.

Why it matters: A genuinely interesting strategic stance. Etsy is positioning “made by people” as the antidote to “chosen by AI” at exactly the moment Amazon and Walmart are pushing in the opposite direction. Watch whether the GMV recovery sticks for two or three more quarters: it will be the first real test of whether human-first works as a marketplace defence in an AI-led market, or whether the strategy just slows the decline.

Walmart’s annual report shows e-commerce boom

Walmart’s 2025 annual report shows e-commerce as the single biggest growth driver, contributing more revenue increase than physical-store comparable sales for the first time. International e-commerce grew 31 per cent year-on-year. Walmart Connect ad revenue grew 28 per cent and is now a stand-alone profit centre that Walmart breaks out separately. The company also called out AI integration with Anthropic Claude as a key 2026 initiative, alongside the Vizio acquisition driving its connected-TV ad business. Walmart Marketplace third-party seller count grew 35 per cent.

Why it matters: Walmart is now an e-commerce-led retailer, full stop. For brands selling on Walmart Marketplace, the implication is unambiguous: retail-media spend on Walmart Connect is no longer optional, it is table stakes. For brands selling against Walmart, the real concern is the third-party Marketplace growth: Walmart is starting to look more like Amazon than like a traditional grocer, which changes the competitive geometry.

Titan launches dedicated B2B ecommerce platform to power India’s corporate gifting

Indian conglomerate Titan launched a dedicated B2B platform for corporate gifting across its watches (Titan, Fastrack), jewellery (Tanishq) and eyewear (Titan EyePlus) brands. The platform offers bulk pricing tiers, tax-compliant GST invoicing, a partner-account portal, and post-purchase reporting. Titan estimates the Indian corporate-gifting market at US$2 billion and growing at double-digit rates as Indian companies expand. The launch is initially India-only, with the company evaluating Middle East rollout for late 2026.

Why it matters: B2B e-commerce in India keeps surfacing this year. Titan, Flipkart Wholesale and Reliance Jio have all moved. Useful signal for UK exporters: India’s B2B-online infrastructure is now serviceable for companies that want to sell into Indian corporate buyers. The corporate-gifting category is also worth flagging as a route-to-market for premium UK brands with brand-permission in India.

3. Platforms, marketplaces & tools

TikTok Shop, Tradebyte, WooCommerce, Rezolve, nFuse and the Practical Ecommerce monthly tool roundup (covering both April fortnights). The shift this month: B2B platforms (nFuse, Rezolve) raised funding or made structural moves, while consumer marketplaces leaned harder on safety, IP protection and integration with social and content channels.

Marketplace Leaders Take the Stage at Tradebyte’s ECD Munich 2026

Tradebyte (the Zalando-owned marketplace integrator) announced its 2026 European Commerce Day (ECD) conference for late June in Munich, marking the event’s tenth anniversary. Confirmed speakers include leaders from Zalando, About You, Otto, Decathlon and several DTC brands. The agenda focuses on cross-border European expansion, AI-driven catalogue automation, and how marketplace integrators are repositioning for the agentic-commerce era. Attendance is expected to be 850-plus, the largest ECD to date. Tradebyte also announced new platform-side AI features for catalogue normalisation.

Why it matters: European Commerce Day is the single most useful conference for European marketplace strategy. If you sell on multiple EU marketplaces (Zalando, Otto, About You, Decathlon, etc.), this is your June diary entry. The Tradebyte AI catalogue tooling is also worth requesting a demo for, particularly if you are managing more than ten EU marketplace channels through a single integrator.

nFuse raises $2M to let small retailers order via WhatsApp

Bulgarian B2B startup nFuse raised US$2 million from Eleven Ventures and LAUNCHub Ventures. The platform lets small independent retailers place wholesale orders through WhatsApp and SMS without learning a separate ordering portal. Pilot customers are in food and FMCG distribution across Bulgaria, Romania and Greece, with a UK launch pencilled in for Q3 2026. The platform also handles payment terms, delivery scheduling and returns through the same conversational interface. CEO Plamen Iliev previously co-founded a regional logistics startup.

Why it matters: Conversational B2B ordering through WhatsApp is genuinely lower-friction than EDI or punchout for small independent buyers, and the format suits markets where WhatsApp is already the primary business-comms channel. Watch for traction with UK convenience, hospitality and small-format grocery. If you are a wholesaler with a long tail of independent retail buyers, this is worth a pilot conversation in Q3 or Q4.

WooCommerce Stores Can Now Sell Products Via YouTube Videos

WooCommerce launched native YouTube Shopping integration. Merchants can tag products in videos, link directly to checkout, and sync inventory automatically. Launch is rolling out across the US, UK, Canada, Australia and twelve other markets. Merchant tagging happens in YouTube Studio with no plugin required; transactions stay on the WooCommerce store with attribution flowing back through standard analytics. The integration also covers YouTube Shorts. Shopify already had this feature live; the WooCommerce launch closes a long-standing gap.

Why it matters: If your store runs on WooCommerce, this opens YouTube as a shoppable channel without the third-party plugin tax (or the workarounds that previously broke on every YouTube Studio update). Worth testing first on top-converting product categories where you already have video content, rather than building net-new content. WooCommerce stores in fashion, beauty and home should also re-baseline their YouTube creator-strategy with this in mind.

New Ecommerce Tools: April 2026 (Practical Ecommerce, both fortnightly editions)

Practical Ecommerce ran two April editions of its fortnightly tool roundup, between them covering the full month. The 1 April issue covered product-image automation, UGC platforms, box-free returns, BNPL, marketplace integrations, shipping intelligence, cross-border tools and affiliate platforms, with launches from Loop Returns, Bolt (one-click checkout improvements), Shopify Magic (product-content generation), Klaviyo and several smaller players, notably a UGC-rights-management tool aimed at fashion brands and an AI shipping-cost optimiser pitched at multi-carrier shippers. The 15 April issue covered cross-channel marketing, B2B procurement, agentic-marketing platforms, shoppable-media, payments, AI product descriptions and shipping intelligence, with launches from Klaviyo (improved AI segments), Attentive (SMS agentic flows), Adobe (Commerce Summit announcements), Stripe (agent toolkit pre-release), a B2B punchout startup, an AI returns-routing tool, and two notable EU-launched VAT-MOSS and OSS automation tools.

Why it matters: The Practical Ecommerce fortnightly roundup is the most efficient way to keep pace with the new-tool firehose without subscribing to every vendor newsletter. Read both April editions together for a full-month view: skim the lists, bookmark anything that matches a current operational pain point, and ignore the rest. Particularly useful if you are running a 2026 stack-rationalisation review, building a martech-vendor watchlist for Q3, or maintaining a regular-cadence tool-watch list. The UGC-rights tool is worth flagging if your fashion or beauty brand uses creator content at scale.

Rezolve AI Goes Directly to Commerce.com Shareholders with Proposal

Rezolve AI (NASDAQ: RZLV) launched a direct-to-shareholder bid for Commerce.com after the latter’s board declined to negotiate. Rezolve proposes a stock-and-cash combination it values at over US$700 million. The pitch is that combining Rezolve’s AI shopping engine with Commerce.com’s merchant base creates an “agentic commerce powerhouse”. The Commerce.com board called the bid “opportunistic and undervalued”. Rezolve has set a tender-offer window through mid-May. The story is a rare hostile-bid drama in the otherwise consolidating agentic-commerce sector.

Why it matters: Whatever the outcome, the bid signals serious consolidation pressure in agentic commerce. The space is moving from “everyone launches a kit” to “a few players try to roll up the merchant base”, and that is a 12-month transition rather than a 36-month one. If you are evaluating an agentic-commerce vendor, factor M&A risk into the contract terms. Smaller agentic players are likely to be acquired or fold within the next year.

TikTok Shop, IPOPHL reinforce commitment to brand protection and safer marketplace

TikTok Shop signed a renewed agreement with the Philippine Intellectual Property Office (IPOPHL) for IP protection across the Philippines. The platform also ran an IP Bootcamp for sellers. Counterfeit-takedown response time has dropped from 72 hours to 24 hours; brand-registry sign-ups grew 40 per cent year-on-year. TikTok also publicly committed to extending the IP cooperation model to Indonesia, Vietnam, Malaysia, Thailand and Singapore through 2026. The agreement covers proactive seller monitoring, takedown coordination, and direct rights-holder communication channels.

Why it matters: This is a tightening pattern across TikTok Shop globally, with South-East Asia first and the EU and UK clearly next. If your brand is not already on TikTok Shop’s brand registry, the gap to enforcement gets wider each month. Worth flagging to legal and brand-protection teams as a Q3 priority. The 24-hour takedown SLA is also competitive with Amazon’s brand registry and notably faster than eBay’s UK takedown handling.

4. Marketing & retail media

The biggest April story by article volume: Walmart accounts for three of the month’s top retail-media moves. Add Google’s AI Max Dynamic Search Ads upgrade, eMarketer’s Q2 KPIs and Adweek’s measurement piece, and the picture is clear. Retail media is now the front-line of digital-advertising spend, and the open-web search budget is migrating into retailer-owned channels faster than most planning models account for.

Online marketing strategies for ecommerce companies

Pan-European overview of 2026 online-marketing strategies for ecommerce: SEO, PPC, social, retention, retail media, and marketplace advertising. The piece includes regional benchmark data for spend share by country (Germany, France, Italy, Spain, Netherlands, Poland, the UK and the Nordics). UK and German marketers are most likely to be running retail-media-led plans; French and Italian are more PPC-led; Polish and Nordic are the most social-led.

Why it matters: Solid generalist-overview piece for anyone briefing-up a new commerce manager or building a strategy doc. Not new ground, but the European spend-share-by-country benchmark data is useful to slot into any 2026 plan that covers more than two markets. Particularly useful if you are allocating budget across UK and EU markets and need a published reference for why allocations differ.

Retail Media Silos, Measurement, and Misconceptions Continue to Stymie Growth

Adweek’s POSSIBLE conference panel highlighted three persistent retail-media barriers: data silos between brand and retailer (each side guarding its own first-party signals), measurement standards that vary across networks (no consistent ROAS definition), and a persistent misconception that retail media is incremental rather than substituting other channels. Speakers from Kroger Precision Marketing, Albertsons Media Collective, Mars and Unilever all flagged measurement as the bottleneck blocking another doubling of spend. The IAB’s draft retail-media measurement standard is now in industry consultation.

Why it matters: If your retail-media planning still uses each network’s native attribution as ground truth, you are over-paying. Push for cross-network measurement before scaling spend further. The IAB consultation is also worth flagging to any in-house measurement or attribution team: the draft standard will move quickly through 2026, and brands that engage in the consultation phase will have an easier time interpreting the final spec when it lands.

Walmart opens first-party data to advertiser partners

Walmart Connect is opening first-party shopper data to advertiser partners through a clean-room model, removing a long-standing complaint that Walmart was holding signals back. Initial partners include The Trade Desk, Pacvue and several agency holding-companies. Brands can now match their CRM against Walmart purchase history at SKU level, build look-alike audiences against actual purchaser cohorts, and run incrementality tests with proper hold-out groups. The clean-room architecture is built on Walmart’s own infrastructure rather than a third-party clean-room provider, which is unusual.

Why it matters: Probably the single most material retail-media story of the month. Brands selling on Walmart should re-baseline their measurement and segmentation in May to take advantage. Agency teams need to skill up on Walmart’s specific clean-room implementation, which is different from the AWS, Snowflake or Habu patterns most teams know. This also forces a rethink of the “Walmart-only” ringfenced budget many brands run.

Walmart Targets Small Businesses With Streaming Ad Marketplace

Walmart launched Connect Select, a self-service streaming-TV ad marketplace aimed at small and mid-sized businesses. Minimum spend is US$500. Inventory is pulled from Walmart’s Vizio acquisition (Vizio CTV plus the FAST channels Walmart now owns), targeting uses Walmart’s first-party shopper data, and placement also runs across the Walmart app’s home and category screens. The company says SMB onboarding takes under 30 minutes and includes a guided campaign-setup wizard. The launch was first to the US; UK and Canada rollout is “under review.”

Why it matters: First time Connected-TV inventory has been opened to SMB-budget advertisers at this price point, in a self-service format that does not require an agency. Watch the UK rollout closely: if it crosses the Atlantic in 2026 it would re-shape how UK SMB media spend gets allocated. Also worth pressure-testing your own CTV strategy against this new baseline of accessibility, particularly if you sell on Walmart Marketplace.

We are upgrading Dynamic Search Ads to AI Max

Google announced AI Max is moving out of beta and replacing Dynamic Search Ads as the recommended approach for unstructured search inventory. AI Max uses Gemini for query matching, asset generation and bidding decisions in real time. Existing DSA campaigns will auto-migrate over the summer; advertisers can opt out for now but the opt-out window is time-limited. The migration includes new creative-control settings (brand voice, keyword exclusion, asset approval) that DSA did not expose. Google is also adding negative-asset controls for brand-safety teams.

Why it matters: If your shop runs DSA campaigns, expect performance volatility around the migration. Pull baseline numbers in May before the auto-migrate window opens, and document current DSA performance as the comparator. Check creative-control settings on day one of migration: the new Gemini asset-generation can drift off brand-voice if not constrained. Worth scheduling a mid-June account audit to catch any silent post-migration regressions.

Search and Retail Media Search KPIs Q2 2026

Q2 2026 KPIs: retail media now captures 21.7 per cent of US digital search-advertising spend, up from 18.2 per cent a year ago. Amazon Ads alone grew 24 per cent year-on-year. Google’s share of total search dropped two percentage points. Walmart Connect grew 38 per cent. eMarketer attributes the shift to retailer-owned audiences having stronger purchase signals, plus the AI-driven shrinkage of open-web search clicks (AI Overviews now answer 12 per cent of US search queries without a click-through to publishers).

Why it matters: The number that justifies any retail-media investment case in a board pack. Google search is not “dying” but the fastest growth in commerce search budget is leaving the open web for retailer-owned channels. Strategic implication: any 2026 plan that grows Google budget faster than retail-media is fighting against the macro trend. The 12 per cent AI-Overviews-no-click figure is also a useful data point for SEO budget conversations.

AI Creative Tools in Google Ads

A practical walkthrough of Google Ads’s AI creative features: asset generation, image variants, headline rewriting, the Performance Max ABCDs framework, and the new asset-feedback signal Google introduced earlier this year. The piece flags where AI works well (variant production at scale, A/B-style testing of headline alternatives, image-aspect-ratio adaptation) and where it does not (brand-voice consistency, regulated-industry copy, claims that need legal sign-off). Includes worked examples for an apparel and a SaaS account.

Why it matters: Useful enablement piece for anyone training agency or in-house teams on Google’s AI features. Worth bookmarking and circulating to PPC analysts who have not yet adopted the AI tools, particularly the ABCDs framework. Pair this with the AI Max migration article as a single training pack: the two pieces together cover both the new mandatory tooling and how to use the existing AI creative tools well.

Inside Walmart’s creator-driven social commerce playbook

Walmart’s head of content, influencer and commerce shared the company’s social principles and detail of the Walmart Creator program, which is now four years old. Creators earn commission on driven sales; Walmart provides tagging tools, an analytics dashboard, dedicated account managers above a sales threshold, and a creator-summit programme. The program now has more than 50,000 active creators across the US, with the top 1 per cent driving the majority of attributed sales. Walmart also runs paid creator campaigns alongside the commission-only base.

Why it matters: A blueprint for how a mass retailer scales creator commerce. UK retailers have nothing close at this scale yet. Boots, Tesco, Sainsbury’s, Asda and Currys all have room to learn from this, particularly the commission-plus-paid hybrid model and the creator-summit format. If you are a brand selling on Walmart, the program is also worth tapping directly through the influencer-collab routes Walmart now offers.

5. New tech & infrastructure

The pipes underneath agentic commerce are being laid. AWS and OpenAI announced an expanded Bedrock partnership; Anthropic and Amazon committed to 5 gigawatts of new computing capacity; Stripe pushed harder on becoming the AI economy’s payment layer; and Revolut launched its UK consumer AI assistant. Underneath the headline launches is a clear pattern: the major US infrastructure players are aligning around a multi-model, agent-native, payment-aware future.

Revolut launches its new AI assistant AIR to UK customers

Revolut launched AIR (AI by Revolut), an in-app AI financial assistant for UK customers. AIR can categorise spending, flag unusual transactions, suggest savings actions, answer balance and policy questions, and explain individual transactions in plain language. Built on a mix of in-house models and third-party LLMs (Anthropic and OpenAI). Initial rollout is to all UK Revolut customers (around 8 million). EU rollout follows in late Q2. Revolut also flagged future merchant-facing features.

Why it matters: The first major UK consumer AI-assistant launch from a fintech with 8 million UK customers, which makes AIR a serious distribution channel. If your e-commerce store accepts Revolut Pay, the AIR-driven discovery experience is worth understanding: AIR can already explain transactions to customers, and merchant-facing features (likely including in-conversation purchase) are on Revolut’s near-term roadmap. Worth a watch through Q3.

AWS and OpenAI announce expanded partnership to bring frontier intelligence to Bedrock

AWS and OpenAI announced an expanded multi-year partnership making OpenAI’s frontier models, including the GPT-5 series, available natively in Amazon Bedrock. Bedrock customers can now mix Anthropic, OpenAI, Meta, Mistral and Cohere models in the same enterprise-grade environment, with single-vendor billing, security and data-governance controls. The deal includes a commitment from OpenAI to make new models available on Bedrock at the same time as via the OpenAI API direct, removing a delay that previously frustrated AWS-native enterprise customers.

Why it matters: If your stack uses Bedrock for governance and procurement reasons but you have wanted OpenAI access, this closes the gap. Worth a procurement and technical-architecture review in May, particularly if you currently route OpenAI traffic through a separate vendor account just for that reason. The single-vendor-billing simplification also matters for FinOps teams trying to consolidate AI spend reporting.

Stripe presses ahead with building economic infrastructure for AI

Stripe announced new agentic-commerce primitives at its Sessions conference: agent-authenticated checkout, programmable disputes, and a Stripe Agent Toolkit for developers building AI-led purchasing flows. The company positioned itself as “the payments layer for the AI economy.” CEO Patrick Collison described the building blocks as analogous to the role Stripe played in the early internet-payments era. Adyen and Klarna are moving in parallel directions; Visa Intelligent Commerce and Amex ACE address the same problem space from the card-network angle.

Why it matters: Stripe is laying claim to the payments rail for agentic commerce, and the multi-network race is on. If you use Stripe today, the new toolkit is worth a developer review even if you do not have an immediate agent use-case: the design choices made now will lock in agent-checkout architecture for the next decade. Also worth tracking how the Stripe versus Visa-network approach plays out, because the platform you build on will determine your dispute and chargeback exposure.

Anthropic and Amazon expand collaboration for up to 5 gigawatts of new computing capacity

Anthropic and AWS extended their computing partnership, with Amazon committing up to 5 gigawatts of new data-centre capacity for Claude training and inference workloads. The deal includes new US sites in Indiana and Mississippi, uses Amazon’s custom Trainium2 chips alongside Nvidia silicon, and is Anthropic’s largest single-customer computing commitment to date. Both companies framed the deal as enabling continued Claude scaling through 2027 and as providing supply assurance for enterprise customers facing capacity constraints elsewhere.

Why it matters: Five gigawatts of computing capacity is a serious bet on Claude’s continued scaling. For brands using Claude in production, capacity bottlenecks should ease through 2026, which makes Claude a safer choice for time-critical agentic-commerce workloads where competitor APIs occasionally throttle. Worth flagging to any architecture team that is been hesitant to commit to Claude on capacity-risk grounds: the 5GW number materially changes the risk picture.

Cash App launches ‘pay later’ feature for P2P transfers

Block-owned Cash App launched a BNPL option for peer-to-peer transfers, letting US users split a payment to a friend over four instalments. Block emphasised in-app debt-spiral protections, including a hard cap on outstanding P2P-BNPL balance per user, mandatory repayment-schedule reminders, and a friction flow that warns users if they are approaching their cap. Initial rollout is to existing US Cash App users with positive transaction history. Block has not announced UK or EU plans.

Why it matters: BNPL into peer-to-peer is genuinely novel as a product structure: it changes consumer expectations of what BNPL covers, and arguably blurs the line between payments and lending in ways the regulator will eventually have to address. Worth watching whether the FCA’s new BNPL rules (live from 15 July) reach this kind of P2P flow if Cash App brings the feature to the UK. Also worth flagging to credit-risk teams as a behavioural-data signal.

6. Agentic Commerce

The signature story of the month. American Express launched the ACE Developer Kit with industry-first protection for registered AI-agent purchases; Visa announced parallel intent. Marketing Week, The Drum, CX Network and Hospitality Net all published serious analyses of what happens when AI does the buying. Brands that have not started thinking about agent visibility now are already behind, particularly in categories where structured product feeds favour feature-fit over brand-narrative.

What happens to brands when AI goes shopping?

Marketing Week’s feature on agent-led purchasing examines how brand-building changes when the buyer is a model rather than a human. Industry voices include Procter & Gamble, Unilever and Mars CPG leaders. The recurring concern: agents optimise for price and structured feature-fit, eroding brand premium for any product that competes on intangibles. The piece flags three defensive moves: investing in genuinely differentiated functional benefits, building first-party customer relationships outside the agent loop, and shaping product feeds to communicate brand attributes that survive the structured-data translation.

Why it matters: If your brand charges a premium today, model the impact of an agent collapsing your category to a feature spec. The strongest defence is brand attributes that do not translate cleanly to a structured product feed (texture, story, sustainability claims, founder narrative). Worth a strategic-brand audit through this framework, particularly in CPG categories where private label is already a substantial threat. Pair with the Etsy-CEO piece in Section 2 for the counter-strategy.

Amazon’s Rufus really remembers me, so now I am letting it do my shopping too

Commerce-media columnist Kiri Masters writes a personal essay on her seven-month experience using Amazon’s Rufus AI assistant. Initially uncanny, Rufus now reliably shows relevant products with the right size, brand and reorder cadence, and Masters has stopped browsing Amazon’s traditional category pages for repeat purchases. She argues Rufus is “winning the agentic shopping war by default” on Amazon’s storefront, with Amazon competitors not yet within striking distance. The piece is intentionally first-person and avoids analyst framings.

Why it matters: Worth reading in full, particularly if you have stakeholders who think agentic shopping is theoretical. The personal-narrative format makes the agent-led shift concrete in a way most analyst pieces do not. Pass it to any board member or non-commerce executive who is sceptical about how fast this is moving. The “winning by default” argument is also worth pressure-testing your own discoverability strategy against, particularly if Amazon is your largest channel.

Advertising and AI: What Ethical Risks Does Agentic Commerce Pose?

Orange’s research arm published an analysis of the legal framework for LLMs in commerce: algorithm transparency obligations, model-liability questions, and consumer-rights gaps when an AI agent makes the purchase rather than the consumer. The piece argues current EU consumer law was not written for agent-led transactions and will need targeted amendment around three issues: who is liable when an agent buys the wrong item, how distance-selling cooling-off rights apply, and what disclosures the merchant must present to a non-human buyer.

Why it matters: The first serious legal-academic analysis of EU agentic-commerce gaps. Useful for any brand thinking about agent-led liability exposure or the disclosure obligations that will likely emerge in 2026 and 2027. Particularly relevant for brands selling into the EU through agents (whether their own or third-party assistants), where the legal-risk picture is genuinely undefined right now. Worth circulating to commercial and legal counsel.

Instacart, Uber Eats integrate with Anthropic Claude AI platform

Instacart and Uber Eats are now both directly accessible from Claude’s app, letting Claude users place grocery and food-delivery orders without leaving the assistant. Anthropic confirmed both apps are live in the US, with UK availability following. The integrations use Anthropic’s app SDK and show the apps in Claude’s tool menu, with order-placement, delivery-tracking and reorder all handled in-conversation. Both Instacart and Uber Eats have separately confirmed the integrations are core to their 2026 distribution roadmaps.

Why it matters: Two massive consumer-delivery brands routing orders through an AI assistant rather than their own apps is a serious distribution-channel signal. If your brand has a delivery channel, Claude (and the equivalent ChatGPT, Gemini and Perplexity apps) are now distribution channels you should plan for. The strategic question is no longer “will agents be a channel?” but “how do we show up in agent-led purchase flows when the customer never opens our app?”

American Express debuts Agentic Commerce Experiences (ACE) Developer Kit; Visa announces parallel scheme

Amex launched the ACE (Agentic Commerce Experiences) Developer Kit on 16 April, plus an industry-first protection scheme for registered AI-agent purchases. Card-not-present fraud rules are extended to agent-initiated transactions; merchants can register specific agents with Amex and receive bespoke dispute-handling. The kit includes APIs for agent authentication, transaction-intent declaration, and merchant-side acknowledgement of agent-initiated transactions. Pilot merchants include selected luxury, travel and B2B categories. Visa announced a parallel scheme in the same week (Visa Intelligent Commerce, giving merchants a way to verify and authenticate agent purchases), with implementation models that differ: Amex is closer to a registered-agent whitelist; Visa leans more toward open-protocol authentication. Mastercard is expected to follow later in Q2. CX Network covered both schemes together on 23 April.

Why it matters: When both major card networks back the same use-case in the same month, the use-case has crossed a threshold. Amex ACE is the first concrete answer to “what if the agent buys the wrong thing?”, and it now sits alongside chargeback and Section 75 as a defined consumer protection. Card-issuer-led protection makes agentic commerce safer for the consumer, which directly accelerates adoption. For merchants, the registered-agent model means there is now a reason to actively curate which agents transact on your store, rather than treating all agent traffic as undifferentiated. Treat agent-checkout as a 12-month roadmap item, not a 36-month one, and pressure-test your payment-processor’s agent-readiness in May.

Ascott Invests in AI-Ready Infrastructure to Scale Agentic Commerce

Singapore-based Ascott (180,000 serviced apartments globally across 60 brands and territories) announced a multi-year programme to make its inventory, pricing and policies machine-readable for AI travel agents. Ascott explicitly cited Booking.com and Expedia’s own agent integrations as the trigger: if intermediaries are exposing AI-readable APIs, suppliers either match them or risk losing distribution. The programme also covers loyalty-data exposure and direct-booking-incentive structures designed to compete with agent-led OTA recommendations.

Why it matters: Travel and hospitality are early indicators for agentic commerce because the inventory shape suits agents (parameterised search, structured availability, defined cancellation policies). Useful watch-item for any travel client. The Ascott approach (machine-readable inventory plus loyalty-data exposure plus direct-booking incentives) is broadly portable to retail categories with similar characteristics, particularly subscription, rental and ticketing businesses.

7. Regulations & rulings

Thin this month, with only one direct UK regulatory story. Worth noting that the TikTok x IPOPHL story (Section 3) and the Orange ethics piece (Section 6) both touch this section. We will fatten this in May as more rulings around the EU AI Act, the FCA’s BNPL rules (live 15 July), DMA enforcement, and the FTC’s pending agentic-commerce inquiry land.

‘New laws will fail shop workers unless employers make it easier to report abuse’

The Retail Trust chief executive argues that the UK’s new retail-crime laws (tougher sentences for abuse of shop workers, faster police response targets, a dedicated retail-crime offence) will not work in practice if internal employer-side reporting mechanisms remain weak. Many retail workers do not believe reporting changes outcomes and do not trust their own employer’s incident-logging tools. The Retail Trust’s piece calls for employer-led reporting reform alongside the new law: simpler digital tools, faster acknowledgement, and clearer feedback loops to workers about what action was taken.

Why it matters: For UK retailers, the legal protection is only as effective as your internal reporting culture. Worth auditing how easy it actually is for a shop colleague to log an incident: most existing systems are not fit for purpose, and the worker-side trust-deficit is real. The board-level question is whether the company is prepared to invest in reporting tooling that matches the new legal framework, or whether the new law becomes a paper protection that does not translate into shop-floor change.
Key takeaways this month
  • Agentic commerce shifted from theory to live infrastructure. American Express launched ACE Developer Kit, Visa announced parallel intent, Anthropic added Instacart and Uber Eats to Claude, Amazon’s Rufus is now placing repeat-shopper reorders.
  • Retail media captured 21.7 per cent of US digital search ad spend. Up from 18.2 per cent a year ago. Walmart Connect grew 38 per cent. Amazon Ads grew 24 per cent. Open-web search budget is migrating into retailer apps faster than most 2026 planning models assume.
  • Operational AI is the 2026 invest-now bet, agentic AI is the 2027 revenue story. Personalisation, dynamic pricing, demand forecasting and customer-service triage already pay back. Agent-led purchasing is largely 2027 for non-Amazon retailers.
  • TikTok Shop crossed into serious-channel status for younger categories. Heavy IP-protection and brand-safety investment in South-East Asia, EU and UK markets next. Under-25 conversion rates lead the social-commerce field.
  • B2B agentic AI moved from demo to production. Product catalogues, ordering, customer service and pricing all show real deployments this month. Industrial and wholesale platforms are catching up to B2C on agent-readiness.
  • Regulators caught up. Card-issuer-led consumer protection for agent-placed purchases sits alongside chargeback and Section 75. EU enforcement of brand-registry rules tightened. Compliance work for 2026 is now concrete, not speculative.

Frequently asked questions

What is the EPIC monthly?

EPIC stands for E-commerce Performance and Innovation Club. It is Anicca Digital’s monthly roundup of the e-commerce and AI stories that matter for senior commerce leaders, published on the last working day of every month and covering the previous 30 days. Each edition is written for B2B and B2C ecom managers, retail directors and senior commercial leads, not for generalist marketers.

What is agentic commerce, and what should I do about it?

Agentic commerce is when an AI assistant (Claude, ChatGPT, Amazon Rufus, Gemini, Perplexity and similar) places the purchase rather than the human consumer. April 2026 was the month it crossed from theory to live infrastructure: American Express launched the ACE Developer Kit, Visa announced parallel intent, Anthropic added Instacart and Uber Eats to Claude, and Amazon confirmed Rufus is now placing reorders for a meaningful share of repeat shoppers. If you sell direct, the priority work in Q2 is getting your product feed machine-readable for agents and pressure-testing your brand attributes against structured agent-led comparison.

How fast is retail media really growing?

Per Q2 2026 eMarketer KPIs (Section 4), retail media now captures 21.7 per cent of US digital search-advertising spend, up from 18.2 per cent a year ago. Walmart Connect alone grew 38 per cent. Amazon Ads grew 24 per cent. Open-web search budget is migrating into retailer-owned channels faster than most 2026 planning models account for. Any plan that grows Google budget faster than retail media in 2026 is fighting against the macro trend.

What is the difference between operational AI and agentic AI in commerce?

Operational AI runs in the background: personalisation, merchandising, dynamic pricing, demand forecasting, customer-service triage. It is already paying back for most retailers and is the largest near-term ROI. Agentic AI handles the actual purchase decision: recommending what to buy, placing the order, handling the checkout. eMarketer’s framing in Section 1 is useful: operational AI is a 2026 invest-now decision, agentic AI is largely a 2027 revenue story for non-Amazon retailers.

Is TikTok Shop ready for serious investment?

Per the TikTok x IPOPHL story in Section 3, TikTok Shop is investing heavily in IP protection and brand safety in South-East Asia, with EU and UK markets clearly next. The Sprout Social data in Section 1 shows TikTok Shop leads on under-25 conversion rates. If your category skews young or visual (fashion, beauty, accessories, home), treat TikTok Shop as a serious channel for 2026 rather than 2027. Get on the brand registry early: enforcement gets stricter every month.

When is the next EPIC edition?

End of May 2026. The last working day of every month is the standard cadence. If a particular story breaks mid-month and you would like our take ahead of the next edition, get in touch using the contact details below.

Want to talk about your e-commerce and AI strategy?

Anicca Digital is a Leicester-based digital marketing agency. We help senior commerce leaders pick the AI bets that move revenue, not just headlines, across paid media, SEO, retail media, content, CRO and AI strategy.

Talk to us

Ann Stanley
Founder & CTO, Anicca Digital

[email protected]
0116 254 7224

Find us online

anicca.co.uk
anicca.co.uk/contact

Leicester, United Kingdom

Book a discovery call
Free guide

The Claude Code Guide

A practical walkthrough for marketers and managers who want to build with Claude Code. No coding required.

Read the guide →
Fortnightly · £40/month

The Thursday AI Club

Hands-on AI sessions for marketers and managers. Live workshops, recordings library, private community channel.

Join the club →
EPIC: E-commerce Performance and Innovation Club
Monthly News from Anicca Digital · First edition · anicca.co.uk

Similar Posts