Why Juicebox Spent 2 Years Solving Just One Problem

David Paffenholz, co-founder and CEO of Juicebox, interviewed by Nakul Mandan (Knuckle Up with Nakul)

David Paffenholz is the co-founder and CEO of Juicebox, an AI recruiting platform he started out of college at 25 with co-founder Ishan Gupta, first as PeopleGPT. The host says about 5,000 companies, from Cursor and Ramp to the Fortune 100, recruit through the product, which has reached more than 3 million candidates. It reached $10 million of ARR with 16 people and now has about 90 employees and a London office. Paffenholz grew up in Germany.

Summary

Paffenholz says Juicebox bet that LLMs would change search, then spent two years on search alone while competitors tried to build a full AI recruiter. An agent runs until a final decision, such as whether to contact a candidate, and a human approves by default. He sees LinkedIn as locked into a seat-based model, and he thinks Juicebox's moats are an index of 800 million profiles and data on candidate intent. His company runs on speed and intellectual honesty, five days a week in person, and he has never lost an employee. These signals are from an interview that Nakul Mandan did with David Paffenholz on Knuckle Up with Nakul.

1. Search First. Paffenholz says almost every recruiter pain point comes down to search: finding the right talent and bringing new people into the funnel. The first version of Juicebox was only a search box, where you could type a query, see profiles and click through to LinkedIn. That is the opposite of mapping a workflow and recreating it, and for two years the team focused on making search as good as possible. Over half of engineering capacity still goes to search. ▶ 2:46