Key Takeaways
- Australian retail giant Woolworths is making its loyalty-app AI fully agentic, with eight automated "judge" agents vetting every response before it reaches customers
- Expedia declared AI a "growth engine" at its Explore 2026 event, embedding agents across booking, marketing, operations, and the partner ecosystem
- AI-driven "discovery replacement" is advancing in both travel and retail, surfacing new questions around visibility, trust, and fragmented consumer protection
Today's Top Stories
Woolworths Makes Its Loyalty-App AI Fully Agentic, Backed by Eight "Judge" Agents
Woolworths has disclosed plans to upgrade Mandy - the AI assistant embedded in its Everyday loyalty, mobile, and insurance products - to a fully agentic system.
letsdatascience.comWoolworths, Australia's largest retailer, has revealed plans to rebuild Mandy, the AI assistant inside its Everyday loyalty app, into a fully agentic system that completes tasks autonomously. Its sibling assistant Olive, already live for customer service, now handles more than 70 percent of contact-centre interactions and moves from staff preview to consumer availability in July.
What stands out is the operational design for quality assurance. Woolworths built eight dedicated agents it calls "agentic judges" that automatically check every Olive response before it reaches a customer. A "number cruncher" recalculates any math, a "product detective" verifies legal and food-safety compliance, and a "goal judge" confirms the agent actually completed its task. This layered review suppresses hallucinations. Both assistants run on Google Cloud's Gemini Enterprise for Customer Experience (GECX).
The hardest part of putting AI agents into production is not accuracy but governance. Woolworths offers a concrete architecture in which one AI vets the output of another, providing a template for retailers weighing agent deployment.
Full article: Woolworths Goes Fully Agentic: How Eight 'Agentic Judges' Vet Every AI Response in Production
Expedia Declares AI a "Growth Engine" at Explore 2026, Embedding It from Booking to Distribution
Expedia Group Explore 2026 shows AI reshaping travel industry workflows, boosting bookings, marketing, operations and partner performance globally.
www.travelandtourworld.comExpedia Group used its annual partner event, Explore 2026 in Las Vegas, to position AI explicitly as a corporate "growth engine." The announcements went beyond traveller-facing conversational tools to reach into the operations of partners such as properties and advisors.
Its Partner Central console gains several agents that help run a business, automating content creation and inventory distribution. Expedia also built B2B infrastructure and an MCP server that lets external AI agents retrieve its products, a design that anticipates a "business-to-agent" flow in which AI participates directly in booking. At the same time, the company addressed the trust questions that arise when AI plans trips on a traveller's behalf, noting its move toward responsible AI governance.
The notable shift is that a booking-intermediary platform is starting to redefine itself as something AI agents "use." For travel and ecommerce operators, structuring their own inventory and product data so AI can handle it is becoming a near-term competitive requirement.
Agentic Commerce
WhatsApp Becomes Oman's New Shopfront as AI Turns Messaging Apps into Full Stores
A growing number of Omani businesses are turning the messaging apps already on customers' phones into fully fledged online stores.
www.omanobserver.omIn Oman, a growing number of businesses skip dedicated ecommerce sites and apps, using the WhatsApp already on customers' phones as their shopfront. This conversational commerce, where product questions, orders, and payment guidance all happen inside a chat, is spreading especially among small businesses.
AI is accelerating the trend. By letting generative AI handle automated replies, stock checks, and order collation, even small operators without dedicated staff can offer round-the-clock service and lose fewer sales. Not requiring customers to install a new app also speeds adoption.
In the Middle East and emerging markets, layering AI service onto a messaging app that already functions as everyday infrastructure often beats building a standalone store. It shows that the starting assumptions for platform choice differ by region.
Travel Commerce
AI Travel Search Risks Making Small Operators "Undiscoverable," New Zealand Tourism Warns
New Zealand tourism faces AI disruption as discovery shifts to prompts, risking SME visibility, overtourism, and reliance on digital data structure.
www.travelandtourworld.comNew Zealand's tourism industry is warning of a "prompt layer" visibility crisis driven by the rise of AI travel search. As destination discovery shifts from Google search to prompts in tools like ChatGPT, operators that do not appear in AI-recommended results risk disappearing entirely.
The core issue is which sources AI chooses to cite. Operators with structured data, third-party mentions, and trust signals are more likely to be surfaced in AI answers, while smaller players get buried. The result could skew recommendations toward a few famous spots, accelerating both overtourism and an opportunity gap.
This is not unique to travel; it is a shared challenge for every business in the AI-search era. Designing information to be cited by generative engines, known as GEO (Generative Engine Optimization), is emerging as the next frontier for protecting visibility, succeeding traditional SEO.
Mobility and Transport Booking
Indonesia Revamps Travel with the Whoosh High-Speed Railway and AI Assistant MaiA
Indonesia combines high-speed rail connectivity with AI-driven travel services to improve visitor experiences.
www.travelandtourworld.comIndonesia's government is pairing Southeast Asia's first high-speed railway, Whoosh, with MaiA, an AI travel assistant developed by the Ministry of Tourism, to revamp the travel experience. Connecting Jakarta and Bandung, Whoosh cuts a road trip that once took around three hours to roughly 45 minutes, making weekend and multi-city itineraries practical.
Unlike conventional tourism sites that simply display information, MaiA understands traveller preferences and generates personalised itineraries within seconds. By surfacing destinations, cultural experiences, and local cuisine in one place, it aims to help first-time visitors explore an archipelago of thousands of islands. The government is also training one million people in digital skills in parallel.
Integrating transport infrastructure with AI-driven planning is an early example of mobility commerce that seamlessly links movement and booking. The pattern of railways and booking services connecting to AI assistants is instructive for other countries, including Japan.
Global E-Commerce Trends
South Korea's KISA Tackles a Surge in Peer-to-Peer Disputes, Filling a 30,000-Case-a-Year Protection Gap
As peer-to-peer secondhand transactions emerge as a blind spot in consumer protection, the Korea Internet & Security Agency accelerates its response framework.
www.asiae.co.krThe Korea Internet & Security Agency (KISA) is stepping up its response to a surge in disputes over peer-to-peer (C2C) secondhand transactions on resale platforms. According to officials, each dispute may involve only a small sum, but they total 20,000 to 30,000 a year, and protection mechanisms have not kept pace with the market's explosive growth.
Peer-to-peer trade has long been a "blind spot" in existing consumer protection. Because buyer and seller hold equal status, consumer protection rules rarely apply, and oversight is split across different ministries and legal principles. KISA is unifying dispute-resolution standards that were previously scattered across organisations and pairing them with an autonomous resolution system in which platforms mediate first.
As AI agents move toward transacting on people's behalf, the institutional design of who bears responsibility and how disputes are resolved is a shared global challenge. Efforts in C2C, the least-protected segment, are instructive for thinking about the trust foundations of agentic transactions.
AI Commerce Tools
AI Forces Affiliates to "Innovate or Die" in the Zero-Click Era
AI is forcing commerce publishers to innovate, which in turn is providing better affiliate partners for ecommerce merchants.
www.practicalecommerce.comAffiliate operators, the commerce publishers who guide readers to purchases through reviews and comparison articles, are being forced to overhaul their model by AI. The long-winning recipe of ranking on Google and funnelling intent-driven searchers to merchants for a commission is breaking down as AI search and AI shopping tools spread.
The biggest blow is a double loss of traffic and attribution. When AI answers shopping questions directly in results and compares and recommends products without sending shoppers to an article, no referral or measurement happens even if AI drew on that article. In response, some publishers are using AI agents to unify fragmented reporting tools, making visible which pages actually generate revenue and informing decisions to update, promote, or consolidate.
AI is both a threat and a productivity weapon, through product-data ingestion and content scaling. For merchants, a selection-and-upgrade dynamic emerges: publishers who survive become higher-quality affiliate partners.
Conclusion
Today's news shows AI agents moving from the stage of "transacting on our behalf" to the stage of implementation and institutional design: how they are operated and who bears responsibility. Woolworths' judge-agent governance, Expedia's partner-facing agent infrastructure, New Zealand's visibility warning, and KISA's C2C dispute work are all new rule-making around trust and discovery. From here, the question to watch is how concretely each operator can translate this operational governance and GEO readiness into working systems.
