#11: Microsoft, GitHub, Honda, Baidu, and the New Era of Multi-Agent Systems

AgentPulse: Your Weekly Dose of AI Agents News.

In partnership with

Welcome back, AI Agents Enthusiasts!

In today’s Agent Pulse:

  • 📢 The top AI Agent news you shouldn’t miss

  • 🚀 A roundup of the hottest trending AI agents

  • Featured AI Agents & Agencies

  • 🎓 Free courses to level up your AI Agent game

Submit your AI Agent to the leading AI Agents Directory & Marketplace.

Request Custom AI Agent from trusted AI Agent builders

📢 Top Stories This Week

Microsoft AI Agent and Copilot Summit Reveals Impressive Adoption: At the Microsoft AI Agent and Copilot Summit this week, partners showcased innovative use cases combining Copilot Studio with Azure AI Foundry. Microsoft revealed that its business-focused agents have already attracted over 2.1 million monthly users, demonstrating significant enterprise traction

Anthropic Releases Claude 3.7 Sonnet with Stronger Reasoning: This model was positioned as having particularly strong reasoning capabilities, especially suited for business use cases involving coding and developer tools, an area where Anthropic has historically focused. Notably, Anthropic acknowledged that the model exhibits weaker performance on mathematical tasks compared to its reasoning prowess.

GitHub Copilot adds autonomous “Agent Mode” in VS Code: Copilot now can act like a self‑driving dev, executing multi-step coding tasks end‑to‑end. In agent mode it can run terminal commands, refactor across multiple files, and even diagnose runtime errors automatically – all powered by large models like Claude 3.7 or GPT‑4o​.

Azure AI Foundry gains agent orchestration and Red‑Team tool: Microsoft enhanced Azure’s AI Foundry platform to better support agentic workflows. Its Semantic Kernel agent framework is now GA for coordinating many AI agents in parallel​, and it added an “AI Red Teaming” agent (preview) to automatically probe LLM apps for vulnerabilities​.

Honda demos Waigaya‑style multi-agent LLM system: In an ICLR 2025 paper, Honda showed that multiple LLM “agents” acting like its internal Waigaya discussion process can solve problems better than a single model. Four expert agents with different specialities debated and coordinated, yielding more accurate engineering solutions than one agent alone​

Baidu debuts “Xinxiang” multi-agent collaboration app: At its Create 2025 conference, Baidu introduced Xinxiang, a “general super agent” that orchestrates multiple AI sub‑agents to handle tasks like planning trips, analyzing data and more. Xinxiang currently supports 200 task types (e.g. knowledge work, travel planning) and is set to expand to 100,000 types, all coordinated from a single prompt​.

Baidu cuts AI costs with new Ernie Turbo models: Baidu rolled out Ernie 4.5 Turbo and Ernie X1 Turbo LLMs that dramatically lower inference costs. The Ernie 4.5 Turbo matches the capabilities of Ernie 4.5 at roughly 20% of the price (and only ~0.2% the cost of GPT‑4.5)​, making high‑end AI much cheaper for businesses.

Atlassian brings Rovo AI agents to Jira and Confluence: Atlassian’s new Rovo platform (announced at Team ’25) is now available to users​. It includes pre‑built and custom Rovo “agents” – AI teammates integrated into Jira, Confluence and Service Desk – and a new Rovo Studio app for building bespoke agents without code. Rovo aims to let teams tap enterprise knowledge with chat, search and assistant features at no extra cost.

Slack expands “Agentforce” digital assistant in Slack chat: Slack made Agentforce (its Salesforce‑powered chatbot) more interactive​. Soon users can simply @mention Agentforce in a channel to get AI‑generated answers in threads, and Slack has improved the agent’s answer formatting for clarity. These enhancements bring AI agents into regular Slack conversations and workflows.

Learn AI in 5 minutes a day

This is the easiest way for a busy person wanting to learn AI in as little time as possible:

  1. Sign up for The Rundown AI newsletter

  2. They send you 5-minute email updates on the latest AI news and how to use it

  3. You learn how to become 2x more productive by leveraging AI

🚀 Trending Agents this Week

View Leaders:
🥇 TheLibrarian.io: WhatsApp AI Personal Assistant
🥈 Agentverse: Search and Discover AI Agents
🥉 PlayAI: Voice AI Agent

Upvote Leaders:
🥇 TheLibrarian.io: WhatsApp AI Personal Assistant workflows
🥈 PPE Kit Detection Agents: AI Based PPE Kit Detection System
🥉 DigitalEmployees: Digital workers for DepOps & QA
 
Review Leaders:
🥇 TheLibrarian.io: WhatsApp AI Personal Assistant
🥈 DoozerAI: Digital workers
🥉 Chorus AI: Create AI Music using Choruz AI

Check out Dynamic AI Agents Leaderboard

 Featured Agents & Agencies

Agents

  • SEObotAI: fully autonomous «SEO Robot» with AI agents

  • Twig: Automate Tier 1 support with AI Agents

  • DoozerAI: AI-driven digital workers that transform business operations

  • TeammatesAI: Autonomous AI Teammates

  • TensorStax: Autonomous AI Agents for Data Engineering

  • TheLibrarianAI: WhatsApp AI Personal Assistant

Agencies

  • DoozerAI: Custom AI Agent development that delivers immediate business value

🎓 FREE AI Agents Courses

 Microsoft has just launched a free new course — Generative AI for Beginners!

This series breaks down the fundamentals of generative AI — from how it works, to how it's built, and how you can start creating your own projects. It's beginner-friendly, well-structured, and packed with real examples.

SPONSOR US

Get your product in front of over 7k+ readers

Our newsletter is read by VCs, founders, engineers, managers and tech professionals.

FEEDBACK

Did you like the newsletter?

Login or Subscribe to participate in polls.

Reply to this email, I would LOVE to hear from you!

I appreciate your support.

OP & Team