# The New AI Agents That Shop, Plan, and Gossip Recap: Every Talk and Launch Explained _This recap of The New AI Agents That Shop, Plan, and Gossip covers the launches and all 4 talks from a16z speakers._ **Published:** 2026-11-04 **Source:** https://www.startuphub.ai/ai-news/technology/2026/the-new-ai-agents-that-shop-plan-and-gossip-recap-every-talk-and-launch-explained --- The New AI Agents That Shop, Plan, and Gossip is an event published by a16z as a video with the same title. The video brings together speakers from a16z to discuss agent products. Speakers from a16z joined for 4 talks. This recap covers the launches and then every talk in order with what was announced and argued. ## What launched at The New AI Agents That Shop, Plan, and Gossip - **Wabi**: launched, multi-thread life management agent interface - **Tommo Interest Groups**: feature connecting users into group chats by shared interest via agents - **Way Joe MCP**: released, provides ability to connect AI assistants to smart home hardware ## a16z: Categorizing AI agents and building trust "the number one fear that people talk about, 22% of the messages are talking about trust as the main fear." a16z said a key concern preventing widespread adoption of financial and personal AI agents is user trust. a16z broke down AI assistant products into general, work-oriented, multiplayer, and travel categories. The survey of group chats with roughly 3,600 messages found trust was the main fear in 22% of messages. A trust framework called Picnic uses bounded boundaries or fences for AI agents. In 1997, 60% to 90% of internet users believed putting credit cards online was preposterous. Hark ran a promotion offering $30 of free pizza in exchange for access to users' Google Workspace accounts. ## a16z: Benchmarking AI agents on flower ordering and travel 14 out of 15 tested AI agents allowed users to export their memory context files to transfer to competitor agents. a16z said testing revealed distinct trade-offs in how different AI agents balance cost efficiency versus speed or luxury. Instinct, Kicker, and Season performed best at balancing cost and delivery speed in a flower ordering benchmark. Kicker utilizes a local Chrome browser installation to enhance its web browsing capabilities. In a travel planning benchmark, Grokbot prioritized time efficiency, DOTS prioritized cost efficiency, and Instinct took a balanced approach. Grokbot was the only agent in a multi-round test to incorporate insights from competing agents into its itinerary. Only 3% of 354 surveyed AI power users use DOTS as their primary agent. Amazon retail was the second most popular purchase category among surveyed users, accounting for 35% of activity. "Instinct really did a phenomenal job. And it is something it's it's interesting how news cycles come and go." ## a16z: Direct shopping and emerging agent product designs [Shopify](https://www.startuphub.ai/startups/shopify) is integrating features to allow AI agents to browse direct-to-consumer store inventories. a16z said agents allow consumers to bypass traditional search aggregators and buy directly from long-tail brands. Wabi recently launched features allowing users to manage multiple life topic threads with AI support. Scout provides a swipe-based UI for learning user shopping preferences. ## a16z: Agent infrastructure scaling, local agents, and agent-to-agent interaction Infrastructure reliability and agent-to-agent communication are emerging focus areas for agent builders. Many current AI agents experience severe performance degradation when scaling up to hundreds of thousands of users. [Thinking Machines](https://www.startuphub.ai/ai-news/artificial-intelligence/2026/thinking-machines-lab-interaction-models-mira-murati-2026) is developing tools to allow users to customize their AI models locally. Tommo released an Interest Groups feature that pairs users studying for exams like the LSAT via agents. Agents interacting in social environments create serendipitous connections and new social opportunities for their human users. Way Joe released Model Context Protocol support to connect AI assistants to smart home hardware. Connecting agents to home appliances like vacuums or AC units provides contextual signals about personal preferences. Non-technical users can adopt AI assistants to perform complex real-world tasks like finding specific dog breeders and negotiating. "can it help you become the human that you've always wanted to be but have struggled achieving yourself? and you start to consider it your secret weapon in your pocket." --- Original analysis from [startuphub.ai](https://www.startuphub.ai), the #1 AI startup directory. © StartupHub.ai. All rights reserved. You may not republish this article in full without a license. Search engines and AI research tools may crawl and summarize for reference. Bulk reproduction or model training on this content requires a license. See https://www.startuphub.ai/terms.