News

Microsoft and Google Turn to Autonomous AI Agents

• By David Ramel
• 10/21/2024

Cloud giants Microsoft and Google are both moving quickly in the trending field of autonomous AI agents as the GenAI space matures from simple query-and-response chatbots and content creators.

Autonomous AI agents can perform tasks or solve problems independently with minimal, non-continuous human intervention. They operate based on predefined goals, using AI techniques such as machine learning, natural language processing, or reinforcement learning to interact with their environment, gather information, and make decisions.

Examples include:

• Self-driving cars: These vehicles navigate roads, make driving decisions, and avoid obstacles without human intervention.
• Robotic Process Automation (RPA): Autonomous software agents can handle routine business processes, such as data entry or email responses.
• Virtual Assistants: AI-driven personal assistants like Siri or Alexa execute commands, answer questions, and interact with users autonomously.
• AI in Video Games: Non-playable characters (NPCs) that adapt to player behavior, learning and modifying their tactics over time.

Today, both Microsoft and Google announced AI agent initiatives. Here's what's up.

Microsoft
The company announced two major developments in autonomous AI agents:

• Public Preview of Autonomous Agents in Copilot Studio: Starting next month, Microsoft will make the ability to create autonomous agents in Copilot Studio available in public preview. This feature allows organizations to "reimagine" critical business processes using AI. These agents can leverage data from various Microsoft platforms to support functions like IT help desks, employee onboarding, and sales assistance.



• Introduction of 10 New Autonomous Agents in Dynamics 365: Microsoft is launching 10 new autonomous agents designed for sales, service, finance, and supply chain teams. These agents aim to drive business value across different organizational functions. Examples include:
  • Sales Qualification Agent: Helps prioritize sales opportunities and guide customer outreach.
  • Supplier Communications Agent: Optimizes supply chains by tracking supplier performance and detecting delays.
  • Customer Intent and Knowledge Management Agents: Improve customer service by learning to resolve issues and creating knowledge-based articles.

"As agents become more prevalent in the enterprise, customers want to be confident that they have robust data governance and security," Microsoft said in an Oct. 21 blog post. "The agents coming to Dynamics 365 follow our core security, privacy and responsible AI commitments. Agents built in Copilot Studio include guardrails and controls established by maker-defined instructions, knowledge and actions. The data sources linked to the agent adhere to stringent security measures and controls -- all managed in Copilot Studio. These include data loss prevention, robust authentication protocols and more. Once these agents are created, IT administrators can apply a comprehensive set of features to govern their use."

Google
Honeywell and Google Cloud announced a strategic partnership to accelerate autonomous operations in the industrial sector using AI agents. The deal integrates Google Cloud's AI platform, Gemini on Vertex AI, with Honeywell Forge, an Internet-of-Things (IoT) platform for industrial applications. Goals include increasing operational safety, reducing maintenance costs, upskilling the industrial workforce and improving productivity by providing actionable insights from vast enterprise-wide data.

The partnership will produce AI solutions tailored to industrial needs, such as automating tasks and resolving maintenance issues, with solutions set to be available in 2025.

The collaboration also includes enhanced cybersecurity measures by integrating Google Threat Intelligence into Honeywell's existing cybersecurity products. Future advancements will explore leveraging Google's Gemini Nano model to improve the intelligence of Honeywell's edge AI devices, enabling them to operate more efficiently without constant internet connectivity.

"These AI-powered tools will also help automate tasks for engineers, warehouse workers, and technicians for Honeywell and its customers," Google said in an Oct. 21 blog post. "For example, AI agents can troubleshoot equipment, suggest design improvements, and offer preventative maintenance insights, such as, 'How did this unit perform last night? 'or 'Why is my system making this sound?'"

What Else Is Going On with Agents?
Today's dual dueling AI agent announcements came shortly after Salesforce and Google Cloud last month said they were joining forces to deliver autonomous AI agents that collaborate across Salesforce and Google Workspace.

A Sept. 17 news release explains the pact was made "to create Salesforce Agentforce Agents that help people collaborate securely across Salesforce Customer 360 and Google Workspace apps. Salesforce's launch of Agentforce extends the existing bidirectional integrations between Salesforce and Google Workspace, allowing mutual customers to easily deploy autonomous agents that can take action and work seamlessly in the apps they use every day, all backed by robust privacy and user data protections already in place across Salesforce and Google Workspace."

And earlier that month, ServiceNow said it was going to unlock 24/7 productivity at massive scale with AI agents for IT, customer service, procurement, HR, software development, and more.

Also last month, Oracle announced more than 50 role-based AI agents within the Oracle Fusion Cloud Applications Suite designed to help successfully execute frequent, repetitive tasks and allow employees and managers to focus their time on more strategic tasks and initiatives.

Earlier this month, Aisera unveiled its vision for "Agentic AI Era."

Also earlier this month, Accenture and NVIDIA announced AI agent news with these highlights:

• New Accenture NVIDIA Business Group launched with 30,000 professionals receiving training globally to help clients reinvent processes and scale enterprise AI adoption with AI agents
• Accenture AI Refinery platform helping companies jump-start their custom agentic AI journeys using the full NVIDIA AI stack
• Network of Accenture AI Refinery Engineering Hubs serving 57,000 Accenture AI practitioners to open in Europe, Asia and North America, supporting large-scale operations, agentic architecture and foundation model development with NVIDIA AI
• Deployment of autonomous agents built in AI Refinery achieves early outcomes in Accenture's marketing function

Obviously agentic AI is one of the hottest areas of AI development right now, so stay tuned for further developments.

Meanwhile, to catch up in the space, here are some recent articles on the topic from sister publication Pure AI:

• The AI Agents Are Coming (For Better or Worse)
• AI Agents Will Change the Internet
• AI Agents at Forefront of Microsoft's Copilot Vision
• Researchers Take AI Agents to the Next Level with the AutoGen Framework
• OpenAI's Swarm Framework: An Experimental Tool for Multi-Agent AI Networks