The company has raised $30 million in new capital led by Sequoia, betting that big language model–powered search can shrink the time and cost of hiring. The startup’s LLM-native engine, PeopleGPT, will flush out top candidates from anywhere on the public web using plain-language queries and nuanced inference instead of mere keyword matching.
Juicebox has now raised $36 million in total funding with the round. Its product has already become a must-have for talent teams that are desperate to gain an edge in a market that’s hypercompetitive when it comes to AI and engineering talent, claim its co-founders (both of whom started the company after going through Y Combinator).
An LLM-Native Approach to Search for Talent
Conventional recruiting platforms search résumés and profiles for specific terms and titles. PeopleGPT instead interprets intent. A recruiter can write “staff-level ML engineers who’ve shipped RAG systems in healthcare and led small teams,” and the model goes through public profiles, personal sites, conference talks, open source contributions, company pages to figure out who’s right — even if that combination of words isn’t there.
After candidates are discovered, Juicebox’s agent can draft outreach, personalize messages and even schedule an initial call. The company says its system combines LLM reasoning with structured filters and verification steps to minimize hallucinations and maintain traceability — important for enterprise adoption where compliance and auditability are key.
Quick Adoption and Initial Revenue Signs
Since releasing PeopleGPT in late 2023, Juicebox says it has more than 2,500 customers and more than $10 million in annual recurring revenue to date. Early adopters range from fast-growing startups to larger tech companies, with users including those at Cognition, Ramp and Perplexity using the tool to help them find scarce AI talent.
Importantly, Juicebox has grown with very little in the way of sales infrastructure, and that’s a sign the product is spreading through practitioner word of mouth. Sequoia’s internal recruiting team tested the software internally, which underscores the firm’s belief that LLM-native search could become default infrastructure for early-stage hiring the way developer tools and payments platforms did in prior cycles.
Why Speed Has Become the Ultimate Hiring Advantage
Time-to-hire is the new competitive moat for teams in a race to ship AI features. SHRM’s Human Capital Benchmarking data has consistently put the average U.S. time-to-fill at a month or more, with cost-per-hire averaging about $4,700. LinkedIn’s studies also reveal that 70% of the global workforce is composed of passive talent — those great candidates who are not applying. They need to be found and persuaded.
Juicebox’s goal is transforming the sourcing process into a natural-language search problem, and thus reducing hours spent on manual profile review and squeezing the first-contact calendar window. The company contends that small advantages — a few days eliminated here from sourcing or scheduling — add up to a quantifiable advantage in winning the kind of competition for top candidates who frequently receive several offers at once.
Where Juicebox Plans to Allocate the New Capital
Juicebox will use the funds to grow its engineering team and strengthen investment in model evaluation and safety, integration with applicant tracking systems including Greenhouse, Lever and Workday. Look for beefed-up data provenance, bias testing against established frameworks such as the NIST AI Risk Management Framework and admin controls in enterprise rollouts.
On the product side, that means richer company and project-level inference (e.g., understanding candidates who have built vector databases at scale despite not having that line on their résumé) and more seamless multi-agent workflows that handle search handoff to outreach to scheduling with less and less human in the loop. It’s not recruiters the tool is meant to replace, but rather to help recruiters so they are able to spend much more time relationship-building and closing.
AI Recruiting Is Getting Increasingly Competitive
Incumbents are responding. And companies such as Eightfold AI, SeekOut and LinkedIn Recruiter have been incorporating generative AI, and semantic search in their products. Juicebox’s pitch is that LLM-native from day one enables deeper reasoning and faster iteration on prompts, evaluation datasets and feedback loops designed for talent search, instead of grafting LLMs onto old filters.
The issue will be maintaining accuracy and scale across noisy, rapidly changing public data while complying with regimes such as GDPR and CCPA. Companies will examine false positives, representation across demographics and geographies and how the system explains its recommendations — all of which frequently drive whether a pilot becomes a standard contract.
What to Watch Next as Juicebox Scales Its Platform
Some of the key measures of success will be a decrease in time-to-source, response-rate lift compared to traditional outreach and percentage of hires from LLM-discovered matches. Depth of ATS integrations and international breadth will also become increasingly relevant as customers expand usage.
If Juicebox can deliver natural-language search in repeatable hiring outcomes, it could prove out the model and develop a kind of default tool for startup operating stacks that could act as a real alternative to encrusted platforms. The Series A provides it with the runway to prove that LLMs can do more than summarize résumés — they can help you find the right people, faster, at the one moment when speed matters most.