An agent company worth $4.5 billion

Plus: How the greatest agentic apps were built, and using LLMs to simulate the economy

Author

Michael Cunningham
October 31, 2024

In partnership with

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

As I’ve assured you before, I write every word of Building AI Agents without assistance from large language models. I guess the next time there’s a heat wave, you know who to blame

In today’s issue…

  • Sierra raises $175 million at $4.5 billion valuation

  • Breakout agentic apps and how they were made

  • AI agents take over hiring

  • How AWS used LLMs to simulate the economy

…and more

📰 NEWS

Sierra

Sierra, a startup by Silicon Valley legend Brett Taylor, has raised $175 million at a $4.5 billion valuation to grow its business building AI agents for customer service.

The CRM giant’s much-hyped enterprise agent platform, announced earlier this year, has been released to general availability.

LinkedIn has launched a new AI agent called Hiring Assistant designed to take on a wide range of recruiting tasks for business customers. The system was reportedly trained on data from 1 billion users across 68 million companies.

Release of the company’s promised multimodal agentic assistant—covered in the very first issue of Building AI Agents back in May—has been delayed until next year at the earliest.

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

🤝 PRESENTED BY AI TOOL REPORT

There’s a reason 400,000 professionals read this daily.

Join The AI Report, trusted by 400,000+ professionals at Google, Microsoft, and OpenAI. Get daily insights, tools, and strategies to master practical AI skills that drive results.

🛠️ USEFUL STUFF

LangChain

A series of deep dives published by LangChain on some of the most successful agentic applications, such as Perplexity and Replit Agent, and how they were built.

A detailed map, broken down by sector, of nearly 300 recent agent startups and projects.

A tutorial on how to use Arcade AI’s user authentication system in LangChain to verify users’ credentials, allowing agentic apps to integrate with real-world systems.

💡 ANALYSIS

Flickr

A new report by tech consulting firm Gartner predicts that, by 2026, over 100 million humans will work with virtual colleagues, but that the genAI field will go through a difficult shakeout as it passes through the trough of disillusionment.

An look at the changes that agents are bringing to the hiring process, including the surprising finding that most job candidates like their chances better when facing an AI recruiter than a human.

A talk by Microsoft’s Jared Spataro giving a case study in achieving significant time and cost savings by using AI agents to process emails from prospective clients.

While massive, trillion-parameter LLMs receive the most attention, this piece argues that smaller, more efficient models will ultimately win out for enterprise applications.

🧪 RESEARCH

Amazon Web Services

This research by AWS used Llama models to roleplay as market actors such as energy companies and utilities, finding that they were capable of successfully reproducing known market dynamics.

The authors of this paper demonstrate a new attack method, termed “foot-in-the-door”, which compromises ReAct agents by exploiting their tendency to re-use the same tools repeatedly.

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