How Hunt Club Accelerates Executive Recruiting with Letta

Automating Executive Recruitment: How Hunt Club Built Hunter with Letta

From tedious manual processes to a streamlined AI assistant, Hunt Club transformed their executive recruitment workflow by building an agent named Hunter with Letta. Their journey shows how the right agent framework can automate knowledge-intensive tasks while maintaining high standards for executive placement.

The Challenge: Manual Knowledge Work in Recruiting

"In executive recruiting there's a lot of unstructured data from a lot of different sources, including direct feedback and conversations logs. It's a mess how you get from this unstructured data to a really crisp idea of the perfect candidate for a given job."

James Kirk, VP of AI at Hunt Club, faced a challenge common to knowledge workers: time-consuming, manual processes that were ripe for automation but required sophisticated handling of unstructured information.

Hunt Club, an executive placement firm, helps companies find the right executives for leadership positions. The process involves gathering information from various sources — job descriptions (sometimes just PDFs, sometimes only emails), conversations with clients, and candidate profiles — and distilling it all into structured "scorecards" that outline the criteria for an ideal candidate.

"That exercise is really tedious," Kirk explained. "We have a lot of people doing this task. Everybody has different mental models about how they come up with the scorecards – when humans do the work, it is hard to be consistent across the organization."

Adding to the complexity, Hunt Club needed to integrate with Atlas, their subsidiary that maintains a vast database of professional relationships and candidate information. Any solution would need to access this rich source of data through API calls.

The team realized they needed an agent that could automate these knowledge-intensive tasks, freeing up recruiters to focus on what they do best: building relationships with clients and candidates.

The Search for the Right Agent Framework

When Kirk and his colleague Steve Corby, the design head, began building their AI agent Hunter, they had clear criteria in mind. They needed a framework that could:

1. Manage a fleet of agents without requiring them to build their own management tools
2. Support custom tools to access their extensive Atlas databases
3. Handle memory and summarization for stateful, persistent conversations
4. Be secure enough to handle protected information

"We didn't want to have to build our own tooling to manage agents," Kirk explained. "Every individual at Hunt Club has their own copy of Hunter. That is a unique agent within Letta. Also every chat room that people pull people into is another copy of Hunter. So we needed to be able to manage this entire fleet of Hunter agents, which Letta does out of the box.”

The team explored various agent frameworks, including OpenAI's own framework, open source projects like Agno, and even considered building their own agentic loop.

"We looked at OpenAI's built-in solution, of course. We just didn't find it flexible enough for the kinds of things we wanted to do and also for fleet management tasks," Kirk said. They also considered the LangChain ecosystem but had concerns about its production-readiness having used it for prior projects.

The ability to easily add custom tools was another crucial factor. "A lot of the other agentic frameworks we've talked to support no custom tooling whatsoever. Or maybe they support some custom tooling, but they don’t support MCP," Kirk explained. "Some companies will say they'll send you some forward deployed engineers to make custom tools. Absolutely not. We are not playing that game."

Building Hunter with Letta

Kirk's approach to building with Letta was hands-on and self-guided. "The ADE (Agent Development Environment) was great for me to just explore how Letta actually worked. I'm the type to not crawl through the documentation until I need to. I just want to actually get a thing running and break it and poke around it."

This experimental approach allowed Kirk to quickly understand how Letta worked and start building. He started with creating the base agent architecture (persona and behavior), then adding tools, and finally integrating it with Slack, where recruiters would interact with Hunter.

One of the most interesting aspects of their implementation was how they handled tools. Hunt Club already had an API service called Atlas AI which managed their AI-related functions, information retrieval, and interfaces to Elasticsearch.

To connect the Hunter agents running on Letta to Atlas AI, they simply needed to expose Atlas AI as an MCP server. This approach allowed them to rapidly add about 50 tools to Hunter, including tools that interact with their database, web browsing capabilities, and file management functions.

From Prototype to Production: Driving Real Results

Hunt Club started small, with about 10 people using Hunter, but plans to expand to around 50 users within weeks. The initial focus has been on the scorecarding workflow, where Hunter can take a PDF job description or email text and generate a standardized scorecard.

"Now those folks can just drop the PDF, paste the email text and just tell us like, 'Hey, Hunter, please make a scorecard for this role.' And it'll just dissolve all that," Kirk explained. Hunter generates the scorecard and pushes it into their CMS, where it becomes accessible as a rendered document in the browser.

The value proposition is clear: "We are automating that process with an agent that has been designed to follow a practice of how to get there. So it gives us the value of cutting down time," Kirk noted. "And it gives us a benefit of standardization."

This standardization creates a "rising tide" effect — as they improve Hunter, all of their processes improve. "Hunter behaves consistently and that means that we can improve Hunter, and when we do that it improves the processes across all our teams," Kirk explained.

Looking Forward: Expanding Hunter's Capabilities

With the initial scorecarding workflow in place, Hunt Club is looking to expand Hunter's capabilities to cover more of the recruitment process. The team is also considering expanding beyond Slack as an interface.

"A lot of our users have already stated that they want to be able to interact with Hunter via email or via other methods or potentially a standalone UI," Kirk said. Letta’s API-centric approach makes it easy to allow the same Hunter agents to work seamlessly across different platforms.

For Hunt Club, the goal is clear: "Just cutting down time. It's a lot of people spending a lot of the pie chart of their hours doing exercises that we know could full well be fully automated, letting them focus on the things that are much higher value-add like drumming up new business, managing relationships, managing candidate relationships, making sure everybody's happy. Less time in PowerPoint decks and more time making sure clients and candidates are happy with the result."

Try Letta today, or request a demo for your team.