CHAPTERS
Alex in one sentence: an autonomous AI recruiting partner
Nicolas introduces Aaron Wang and John Rytel and the company’s recent $17M Series A. Aaron gives the crisp product definition: Alex is an AI recruiting partner that can run key recruiting workflows end-to-end.
- •Announced $17M Series A led by Peak XV
- •Positioning: “AI recruiting partner,” not just a tool
- •Goal: interview everyone and hire the best people
- •Handles recruiting workflows with autonomy
What Alex automates: sourcing, screening, interviewing, scheduling, ATS updates
The founders explain what Alex does day-to-day for recruiters and staffing firms. The agent connects to existing recruiting systems and executes repetitive tasks that consume recruiter bandwidth.
- •Conducts phone screens and video interviews
- •Schedules candidates automatically
- •Sources and reaches out to talent
- •Updates ATS/HRIS and keeps systems of record current
- •Designed to work with “your favorite recruiting tools”
Why recruiting is broken: application volume surge and slow time-to-hire
Aaron argues the hiring market has become dramatically less efficient: more applicants, longer cycles, and limited recruiter capacity. Alex targets both the bandwidth bottleneck and the matching problem using data and scalable interviews.
- •Applicant volume has tripled in ~3 years
- •Time-to-hire around 60 days (all-time high)
- •Core bottlenecks: recruiter bandwidth + poor matching
- •Approach: interview at scale + use data for better matching
- •Promise: better candidate insights from structured conversations
Scale and traction: thousands of interviews per day, tens of thousands of open roles
They quantify usage and deployment maturity since launch. Alex is already running large volumes of interviews daily and supporting hiring across many active requisitions.
- •Thousands of interviews conducted each day across customers
- •Helped hire thousands of people since launching last year
- •Actively filling tens of thousands of job roles
- •Suggests production reliability at high throughput
Who uses Alex and why staffing agencies were the wedge
Aaron explains the go-to-market: staffing agencies pulled hardest because incentives align and pain is concentrated. They also support very large employers across diverse job families.
- •Recruiting market is huge because every employer hires
- •Initial pull came from staffing agencies (“hair-on-fire” problem)
- •Aligned incentives: better screening → more placements/revenue
- •Also serves some of the world’s largest employers
- •Roles span technical to blue-collar (e.g., nuclear welders)
Role depth and customization: testing skills and learning from ATS + intake notes
The conversation digs into how Alex can screen for niche qualifications (like nuclear welding) and assess technical skill early. Alex uses internal company context (ATS history, job descriptions, hiring manager notes) and can be tuned per role.
- •First-round interviews include skill testing
- •Learns employer context via ATS/HRIS + hiring manager inputs
- •Fine-tuning for niche roles and domain-specific questioning
- •Aims to act as an “employer brand ambassador”
- •Can “dig deep” to qualify candidates beyond surface keywords
Origin story and timing: candidate-first insight and GPT-4 Turbo enabling voice agents
They trace the idea to their experience as candidates and earlier hiring-tech experimentation at Brown. Building became feasible when conversational AI/voice latency improved in late 2023, leading to a 2024 build and launch.
- •Founders’ advantage: long-time candidates, focus on candidate experience
- •Met at Brown; built earlier hiring-tech project
- •Inflection point: GPT-4 Turbo (late 2023)
- •Started building early 2024; launched around April 2024 (YC batch)
- •Early customer traction soon after launch
Making AI interviews feel natural: low latency, orchestration, guardrails, and naming
John describes the technical work to escape the “uncanny valley”: orchestrating multiple models for low-latency voice, improving transcription and voice quality over time, and adding guardrails so interviews stay structured. Aaron shares a practical reason for the name “Alex”: transcription reliability.
- •Key hurdle: voice agent latency low enough for comfort
- •Built orchestration platform before modern voice tooling matured
- •Continuous improvements: latency, transcription, voice quality, reliability
- •Prompting + guardrails to keep interviews on track and complete
- •Name “Alex” chosen partly because speech-to-text recognized it reliably
Adversarial candidates: prompt injection, AI cheating, and deepfake detection
They discuss how candidates try to “hack” interviews, especially engineers, and how the product defends against abuse. The arms race includes mass applying with AI tools, using real-time AI “overlays,” and even live deepfakes in video interviews.
- •Attempts include prompt injection via XML/Markdown and edge-case probing
- •Market dynamic: candidates are incentivized to game scarce jobs
- •Rising trend: AI tools to mass-apply and to cheat during interviews
- •Built cheat detection and models for deepfake detection
- •Video interviews run on a custom-built conferencing platform
Fixing the job hunt: everyone gets a first interview and ghosting drops
Aaron and John outline a candidate-centric future: applications should reliably yield an initial conversation. They argue AI can democratize access to opportunity, reduce ghosting, and help candidates get updates and answers throughout the process.
- •Vision: applying should mean getting a first-round interview
- •Goal: opportunity less tied to pedigree; more to demonstrated skills
- •Candidates can ask questions during interviews
- •Ongoing support via texting Alex (policy questions, status updates)
- •Improved transparency and reduced “ghosting”
Proof it works: resurfacing hidden talent, stack-ranking, and retention improvements
They give a concrete example where Alex mined an existing ATS database to find scarce COBOL candidates and quickly placed them. Success metrics include whether Alex’s rankings match or beat recruiters, plus downstream retention and staffing revenue outcomes.
- •Example: found 11 COBOL candidates “lost in the database” and placed them
- •Alex can auto-reach out, interview, and submit qualified candidates
- •Pilot metric: quality of stack-ranking vs human recruiters
- •Uses interview quotes/evidence to justify scores
- •Observed improvements: retention and staffing placement economics
Series A, rebrand, and building trust with enterprises
Aaron explains raising now due to market pull and the need to be best-in-class for enterprise HR buyers who’ve been burned by past tech. They also discuss the rebrand from Apriori to Alex because customers already used that name and it felt more approachable.
- •Use of funds: expand team and product leadership in AI recruiting
- •Enterprise HR requires transparency and confidence-building
- •Non-obvious focus: go-to-market, brand, and relationships in enterprise sales
- •Rebrand rationale: customers already called them Alex; easier to pronounce/share
- •Positioning: approachable “partner,” not abstract AI software
Five-year outlook: recruiters supercharged, humans focus on relationships; founder lessons
They argue AI won’t replace recruiters but will remove repetitive admin work so recruiters can focus on closing candidates and partnering with hiring managers. They close with hiring plans and advice: hold strong vision while staying flexible as models and tools evolve quickly.
- •AI removes repetitive questions, scheduling, and ATS busywork
- •Humans remain critical for relationship-building and closing candidates
- •Company hiring across engineering, design, PM, and go-to-market
- •Founder lesson: high conviction on the end state; flexibility on the path
- •Stay versatile—tech changes fast; learn new tools quickly
