Microsoft Copilot launches Agentic Storefronts as Google Gemini enables Gap purchases

Key Takeaways

• Shopify merchants can now sell products directly within ChatGPT, Microsoft Copilot, and Google Search AI through Agentic Storefronts.
• Gap is the first major fashion retailer to enable direct purchases within Google's Gemini AI, utilizing Google Pay.
• OpenAI has enhanced ChatGPT with richer visual shopping experiences, allowing users to browse and compare products more effectively.
• Marketeam.ai launched RL-KPI at NVIDIA GTC, an AI training method built on NVIDIA NeMo RL, to optimize for real business outcomes like ROI and customer acquisition cost, potentially boosting ROI up to 6x.
• The US AI governance crisis is linked to a broader decline in democratic infrastructure, with concerns about industry influence over regulation.
• An MIT-led team is developing

  Shopify merchants can now sell directly in ChatGPT

  Shopify is enabling millions of merchants to sell products directly within ChatGPT, Microsoft Copilot, and Google Search AI starting this week. This new feature, called Agentic Storefronts, allows customers to discover and purchase items without leaving the AI chat interface. Merchants can manage these sales centrally through their Shopify Admin, with customizations and brand experience carrying over. Orders will show ChatGPT referral attribution, and merchants retain customer ownership and data. This move integrates Shopify's commerce infrastructure with major AI channels, aiming to make shopping more seamless.

  Gap first major fashion retailer to sell via Google Gemini

  Gap is becoming the first major fashion company to allow customers to buy products directly within Google's Gemini AI. Shoppers can purchase items from Gap's brands without leaving the Gemini platform, using Google Pay for checkout. Gap controls the product information shown to customers, ensuring accuracy and retaining customer data. This partnership allows Gap to be more relevant in conversational searches and provides a competitive edge in the retail market. Gap is also launching an AI-powered sizing tool called Bold Metrics.

  ChatGPT now offers richer visual shopping experiences

  OpenAI has enhanced product discovery within ChatGPT, allowing users to browse products visually, compare options side-by-side, and get detailed information. The platform now uses the Agentic Commerce Protocol to provide more relevant and up-to-date product details. This upgrade aims to turn online shopping from a fragmented process into a seamless experience, helping users make faster decisions. These improvements are rolling out to all ChatGPT users this week.

  Marketeam.ai launches RL-KPI for AI training at NVIDIA GTC

  Marketeam.ai introduced RL-KPI at NVIDIA GTC, a new AI training method that extends deterministic reward learning to non-deterministic business outcomes. Built on the NVIDIA NeMo RL library, RL-KPI allows AI models to optimize for real business metrics like delayed conversions and multi-objective trade-offs. This approach addresses the gap where AI is optimized for proxy metrics instead of actual business value. Industries like marketing and sales can now develop AI solutions that better align with strategic goals and improve metrics such as customer lifetime value and revenue growth.

  Marketeam.ai's RL-KPI boosts ROI and cuts costs with new AI training

  Marketeam.ai unveiled RL-KPI at NVIDIA GTC, a new AI training method that optimizes for real business outcomes like Return on Ad Spend and customer acquisition cost. This approach allows AI models to learn from actual campaign data and KPIs, moving beyond theoretical performance. Marketers can achieve up to 6x ROI and significantly lower customer acquisition costs within 6 to 8 weeks. RL-KPI handles complex business challenges such as delayed conversions and balancing multiple objectives, marking a new paradigm for AI in business.

  US AI governance crisis linked to democracy decline

  A new analysis suggests that America's struggle to govern artificial intelligence is a symptom of a broader crisis in its democratic infrastructure. The White House's proposed federal AI policy is criticized for preempting state laws and favoring industry-led standards over meaningful regulation. Critics argue this approach, similar to tactics used by the tobacco and gun industries, creates regulatory gaps that benefit tech companies. The article contends that the dismantling of democratic processes over decades has enabled powerful industries to overwhelm political systems and avoid accountability.

  MIT team develops 'humble' AI for medical diagnosis

  An MIT-led research team is developing artificial intelligence systems for medical diagnosis that are designed to be more humble and collaborative. These AI systems will indicate when they are uncertain about a diagnosis or recommendation, encouraging doctors to seek more information. The goal is to prevent AI from making overconfident, incorrect decisions and instead function as a coach or co-pilot for clinicians. This framework aims to instill values like curiosity and humility in AI, fostering a partnership between doctors and AI systems.

  Apple delays new hardware due to AI software issues

  Apple is reportedly holding back the launch of new smart home hardware because its AI features, particularly Siri, are not yet ready. This marks a shift from Apple's usual hardware-first approach, as competitors increasingly integrate AI into their products. The company appears unwilling to release new 'smart' devices without robust AI capabilities. While the hardware may be finished, the software delays are causing a noticeable absence in Apple's recent product cycles. New AI upgrades are expected to be ready for a potential launch between now and September.

  Intercom shifts to outcome-based pricing for AI services

  Intercom has moved from traditional seat-based pricing to an outcome-based model where customers pay only for successful AI resolutions. This new model aligns revenue directly with product performance, encouraging faster improvements. If Intercom's AI agent, Fin, cannot resolve a customer conversation and it needs escalation to a human, the customer does not pay. This approach puts revenue at risk for the company, driving accountability and faster development. This shift requires new billing tools to manage cost volatility and supports scalable growth.

  Middle East AI infrastructure faces geopolitical challenges

  The Middle East has rapidly become a major hub for AI infrastructure, driven by significant sovereign investments. However, escalating geopolitical tensions are now creating volatility for this expansion. Governments view data centers and AI platforms as critical for future economic growth, with capacity expected to triple in five years. While technology architecture is designed for resilience, regulatory frameworks and data localization requirements pose challenges. Disruptions in the region could also impact the global semiconductor supply chain, affecting AI development worldwide.

  Ex-Blizzard president defends AI use in games

  Former Blizzard president Mike Ybarra stated that gamers need to accept AI in video games, following an apology from Crimson Desert developers for using undisclosed AI-generated assets. Ybarra argued that AI will be integrated into all forms of technology, including games, and that developers should not apologize for its use. He believes gamers should 'man up' and accept this reality. The controversy arose after players discovered AI-generated assets in Crimson Desert, which the developers have since promised to replace.

  IBM Consulting embraces AI for client transformations

  IBM Consulting is adapting its approach to client services in the age of AI, viewing it as a fundamental shift in how work gets done. The firm is integrating AI software and intelligent agents into its operations, creating a model of 'services as software.' IBM emphasizes that real AI impact requires embedding it into core systems with proper governance and skilled personnel. Having implemented AI across its own operations, generating significant productivity gains, IBM Consulting now applies these tested methods to help clients navigate their AI transformations.

  Raunchy AI fruit videos go viral on TikTok

  A new genre of AI-generated videos featuring anthropomorphic fruits and vegetables engaged in dramatic love affairs is gaining popularity on TikTok. These 'AI slop' videos often feature nonsensical plots and mispronounced words, drawing comparisons to reality TV and soap operas. Examples include a strawberry having an affair with an eggplant, leading to an 'aubergine baby.' The trend highlights the bizarre and often illogical content being created with AI tools.

  Sources

  • Millions of merchants can sell in AI chats
  • Gap becomes the first major fashion company to offer checkout within Google's Gemini
  • Powering product discovery in ChatGPT
  • Marketeam.ai Unveils RL-KPI at NVIDIA GTC: Breakthrough AI Training Method Extends Deterministic Reward Learning to Non-Deterministic Business Outcomes
  • Marketeam.ai Unveils RL-KPI at NVIDIA GTC: Breakthrough AI Training Method Extends Deterministic Reward Learning to Non-Deterministic Business Outcomes
  • America’s AI Governance Crisis Is a Democracy Crisis
  • How to create “humble” AI
  • Apple’s AI issues are stopping new products from launching
  • In an AI-Driven Economy, What Are Customers Actually Paying For?
  • AI infrastructure investment in the Middle East enters a new geopolitical reality
  • Ex-Blizzard president says gamers need to 'Man up' over AI usage in games after Crimson Desert devs issue apology
  • Consulting’s AI moment: Why the old playbook no longer works
  • A.I. Slop Videos of Adulterous Fruit Are Taking Over TikTok