Microsoft has unveiled the Agent Store, a centralized marketplace where users can “shop” for AI agents tailored for specific tasks.
Integrated into Microsoft 365 Copilot, the Agent Store serves both developers and non-technical users who want to use artificial intelligence (AI) agents to streamline workflows, automate tasks and enhance productivity across the workplace.
“The Agent Store is your one-stop shop for the next generation of AI assistants,” Principal Product Manager Siffat Hingorani and Senior Product Manager Olive Hu wrote in a blog post.
Through the marketplace, users can browse, install and try agents from Microsoft and its partners and customers.
OpenAI has also introduced a similar marketplace for GPTs that offer bots designed to perform specific tasks. Users can pick from a selection of GPTs or make their own.
What’s different between the two is that Microsoft’s agents are embedded within the user’s work environment, with the user’s data as context. OpenAI’s GPT marketplace offers more general tasks to the public.
Microsoft’s Agent Store launched with more than 70 agents, which it said would grow over time. Developers can build their own agents and publish to the store using Microsoft Copilot Studio and Microsoft 365 Agents Toolkit.
Read more: Microsoft’s Radical Bet Turns Windows Into an AI Playground
Anthropic’s Claude Introduces Voice Mode
Following in the footsteps of ChatGPT and Gemini, Anthropic’s Claude now lets users converse through voice, not just text, for all its users.
But unlike ChatGPT, Claude voice integrates with Google Workspace to access Calendar, Gmail and Docs so it can do such things as audibly tell users about upcoming meetings, important emails and relevant news.
The Workspace integration is for paid subscribers, with Docs accessibility for Enterprise plan users only.
Gemini voice also offers integration with Workspace.
Voice mode is currently available in English on Claude mobile apps. For free users, Claude has a daily limit of 20 to 30 voice messages. Paid plans have much higher limits depending on the type of plan.
Users can choose among five voices. Voice mode is different from dictation in that Claude can talk back and converse with users.
Last week, Anthropic released its most advanced models yet: Claude Open 4 and Claude Sonnet 4. These models can reason and also use tools like doing web searches. The startup also released four new capabilities for the Anthropic API to build AI agents: code execution tool, MCP connector, Files API and the ability to cache prompts for up to one hour.
Read more: Anthropic Snags $2.5 Billion Credit Line as AI Demand Rises
See also: Anthropic Says DOJ Proposal to Monitor Google’s AI Investments Would Harm Competition
For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.
