18 easy starter agents

Welcome back to Building AI Agents, your biweekly guide to everything new in the AI agent field!

Or maybe trade school. How else would they learn to use tools?

In today’s issue…

• CrewAI’s new hardware for enterprise agents

• 18 easy starter agents

• The rise of the AI manager

• Why do agent systems fail—and how to prevent it

• 5 top business use cases for agents

…and more

🔥 IN CASE YOU MISSED IT

Readers’ favorite items from last week

1. How to know a business process is ripe for agentic AI

2. No one knows what the hell an AI agent is

3. 5 thoughts on agent engineering

4. Actual LLM agents are coming. They will be trained

5. As AI agents multiply, IT becomes the new HR

📰 NEWS

CrewAI launches hardware for enterprise agents

As part of its Enterprise Agent Week (which includes our Building AI Agents’ own Chicago meetup!), agent framework provider CrewAI announced a partnership with Hewlett Packard Enterprise (HPE) to build hardware fine-tuned for Crew’s multi-agent software.

Browser Use raises $17 million

The maker of a tool which enables AI agents to browse the internet—and was the backbone of the recently viral Manus agent—raised $17 million from a consortium of investors including YC co-founder Paul Graham.

If you find Building AI Agents valuable, forward this email to a friend or colleague!

🛠️ USEFUL STUFF

Claude’s simple but brilliant “think” tool

Anthropic recently discovered that adding a tool which simply allows a model to take more time to think when necessary significantly raised Claude’s performance—and it should be easy to implement for any other tool-enabled LLM as well.

18 easy starter agents

A goldmine of simple agent demos by database provider MongoDB providing an easy starting point for building a variety of agentic systems.

Zapier MCP integrates agents with thousands of APIs

Zapier Model Context Protocol (MCP) servers enable agents to connect to over 7,000 apps through a common interface, letting users create agents that can integrate with Slack, Google Calendar, Jira, and more.

💡 ANALYSIS & RESEARCH

The rise of the AI manager

Contrary to popular narratives about AI taking jobs, this piece argues that it will actually empower workers by turning everyone into a manager of armies of AI employees.

Why agent systems fail

This paper comprehensively studies the ways in which LLM agents can fail, dividing them into 14 unique modes across 3 categories, and evaluates several possible solutions to them.

5 top business use cases for AI agents

A roundup of the 5 main areas in which AI agents are already making a significant impact within large enterprises.

How (not?) to pivot to AI agents

UiPath is one of the robotic process automation (RPA) companies aggressively re-orienting their product offerings towards AI agents, but massive costs, competition, and internal culture clashes are leaving investors skeptical that it can pull it off.

Reducing uncertainty in agentic workflows

Rather than trying to eliminate uncertainty from agent systems, the author of this article makes the case that they should be built to assume that uncertain situations will arise and handle them reliably.

Thanks for reading! Until next time, keep learning and building!

If you have any specific feedback, just reply to this email—we’d love to hear from you

Follow us on X (Twitter), LinkedIn, and Instagram