Staffing Industry Spotlight: Aaron Wang
Staffing Industry Spotlight is an interview series featuring leaders shaping the staffing industry, sponsored by Ascen, a leading back-office and employer-of-record for staffing agencies. This edition features Aaron Wang, Founder and CEO of Alex, a YC alumnus and former AI researcher at Meta.
Aaron shares his journey of building domain-specific AI for the staffing sector, focusing on how autonomous 24/7 phone and video screening can dramatically increase placement efficiency. The conversation covers critical modern challenges, including the rise of AI-assisted candidate fraud, the strategic shift by recruiters toward relationship-driven selling, and Aaron’s outlook on why advanced technology will continue to drive recruitment firms' market value upward.
Francis Larson:
Aaron, thanks for being on the Staffing Industry Spotlight. I'm super excited to chat about AI stuff today. You are the founder and CEO of alex.com, a YC founder, and you recently raised a $20 million in Series A. Most importantly, I want to know how on earth you got into AI for recruitment.
Aaron Wang:
Thanks for having me on, Francis. We got started because my background is in computer science and AI. That is an important aspect of my background, as I am not a recruiter and have never worked in staffing. I have been a job seeker for longer than I have been an employer. You need to adopt a very humble mindset when working with customers, asking them to share their pain points so we can work together to solve them.
The job market has always been particularly interesting to me. It is the largest market in the world, and humans are what make the world go round. People say that with all this AI, it will replace people. I can't think of a more interesting problem than saying, "Hold on, what if we use AI to put more people back to work and connect them with the right opportunities?" Alex is great for that. Staffing firms use Alex to interview and screen everyone, ultimately placing the best people for their clients.
Francis Larson:
You had done other jobs before this. I saw in your background that you were an AI researcher at Meta for a while, so you came into this with a deep technical background. You started in 2024, and AI has developed a lot even since then. How has the project progressed from what you started building to what you are doing today with Alex?
Aaron Wang:
A few things have changed. When I was at Facebook AI Research, now known as MSL (Meta Superintelligence Lab), we focused heavily on domain-specific models. I worked on computer vision for augmented reality glasses, like the Ray-Ban glasses you see. Today is very different because you have these generalized models, like GPT-5.4. It does a lot of work, and that has been particularly interesting because you can take that generalizable model and make it domain-specific. Alex is a domain-specific AI for staffing. The way staffing firms use AI differs inherently from the way law firms use it. A major spark for us in 2024 was the advent of low-latency models, which enabled a human to have a real-time conversation with AI. That is very useful for screening calls and interviews. You can screen thousands of people over phone calls in a day because the AI has the low latency required for those conversations.
Francis Larson:
When you say domain-specific, do you mean you have a lot of recruitment-specific context, or are you actually fine-tuning on top of the models?
Aaron Wang:
Both are true, and I will offer a third: modality. If we look at the recruiter versus law firm example, the way a lawyer interacts with clients is different from how a recruiter interacts. Law firm lawyers are typically emailing and texting, so it is mostly text-driven. A recruiter is often on phone calls or Zoom meetings. That is a different modality involving video and audio. When I think about domain specificity, I think of context, fine-tuning, and modality.
Francis Larson:
Alex does a lot of things. What is Alex, what are the feature sets, and how can companies use it?
Aaron Wang:
Alex is essentially your AI recruiter. We work with some of the largest staffing firms and employers in the world—three out of the four largest recruiting firms. Alex conducts autonomous phone screens, video interviews, and emails or texts candidates 24/7. Ultimately, if you open a requisition and get 1,000 applicants, Alex calls them all the second they apply. It whittles that list down to a shortlist of the top three or four people who are actually worth your time. The other 997 people might not have made it because they were based in the wrong location, were not work-authorized, were not within the pay range, or were a better fit for a different role. Alex uncovers all of that with a quick phone call or interview.
Francis Larson:
So, someone applies, Alex reaches out to schedule, does the phone or video interview, and rapidly whittles the list down without human interaction. How much input does the recruiter or the company have to put into the interview to ensure it's good?
Aaron Wang:
It is up to the firm, but typically, it is minimal involvement. If I am a busy staffing firm, time is of the essence, and I need to outcompete the rest of the market. If I am an IT staffing firm hiring for a .NET developer, it turns out the AI knows a lot more about .NET than I do. Alex does a great job at crafting that interview guide and scoring those interviews using the job description and any available data in the applicant tracking system. It builds a holistic guide for what is important in that interview. It can then go out and interview thousands of people and tell the recruiter, "These are the top five people. Feel free to chat with all of them, but you should probably submit them to your client as quickly as possible."
Francis Larson:
Do you see your customers doing any manual follow-up interviews, or are they often just relying on Alex's results and sending them to the client?
Aaron Wang:
Our best customers get great results with a fully autonomous process. It is as if you have an AI employee or colleague working for you 24/7. A lot of people haven't figured out how to buy or interact with an AI colleague yet, or what a performance review for an AI employee looks like. Those have been interesting problems to solve. One of our customers, a massive IT staffing firm, increased the number of offers given by 30% quarter-over-quarter. They are almost completely automated. They are essentially beating their competitors to the candidate and then submitting them to the client.
Francis Larson:
So they don't even have to do a follow-up human interview to screen them again?
Aaron Wang:
Many customers start with an SMS screener, then move to a phone call, and finally to a video interview via Google Meet or Zoom. If all that looks good, you might have a quick call with the top one or two candidates to guide them through the client-specific process and continue to "sell" the candidate. AI will never do a better job at selling than you will, and that's not the point. The point is to make sure you are selling to the right people. Our humans should spend more time on the relationships that matter and affect the top line. That has been an interesting change in how recruiters spend their time, becoming more strategic.
Francis Larson:
Candidates are starting to experience more AI interviewers. I'm assuming you're seeing attempts from candidates to use AI voice bots or context-fed tools to cheat. Are you seeing that kind of fraud yet?
Aaron Wang:
It is a mix of both. We have seen a huge rise in AI-supported cheating, whether through deepfakes or other software. About one in four candidates today are flagged for some form of cheating, and we make that information available to our customers. We suspect the number is actually higher, but we are relatively lenient because we don't want to flag something that shouldn't be. My guess is the number is at least double that in reality. The cost of submitting an application is zero, and they can boot up their own AI to apply to thousands of jobs a day. As a staffing firm, you need to react to that. If 200 people apply, your acceptance rate is very low. Out of those 200, there are two people who are incredible fits. The problem is I now have to go through 198 people who aren't a fit to find those two. In reality, that number is closer to 2,000, and it's getting higher.
Francis Larson:
How can you tell that someone is AI-assisted in these video or audio contexts? Are there any methods you can share for how you flag that?
Aaron Wang:
A lot of it is based on what humans are doing today. For us, we essentially give Alex eyes and ears. For example, I can tell that you have a monitor up and an iPad, and it looks like you're taking notes. I'm not going to flag you for cheating there. But there is a big difference between that and a candidate who pauses for ten seconds every time I ask a question, while I hear typing on a keyboard before they respond. We record the interview, and it marks timestamps where these things happen, so the recruiter can click and see it for themselves. We do a lot of incredible stuff under the hood to ensure we catch people who aren't interviewing with integrity.
Francis Larson:
There is interviewing, scheduling, and fraud control. What else is on the roadmap?
Aaron Wang:
Our Alex Verify product has been incredible for the cheating and fraud detection we talked about. Because we sit at the top of the funnel, we might as well tell you whether they are legitimate. Alex also screens resumes. If a role is based in Boston and they say they are in India, it might not be worth the AI's time to interview them. Our customers are also really excited about our Talent Match product. Before I even post a job on expensive job boards, what if Alex could look into my existing database, find the best people, and give them a call first? I've already paid to have those candidates in my database. If Alex finds 100 people who could be a good fit, let's call them first, so I don't have to post on Indeed or LinkedIn.
Francis Larson:
Do you see scope for cold outreach with AI anytime soon?
Aaron Wang:
I think likening recruiting to sales is useful as an analogy for product and incentives. I'm sure someone will do cold outbound. It is something we technically offer our customers today, but we do not go to market with it because you are better off targeting the warmest, high-intent areas to find candidates. Whether it's inbound or folks already in your database, you have the materials you need. We find high-intent channels to be the best.
Francis Larson:
Because you're able to deal with volume—interviewing a thousand people for a very low cost—are you seeing Alex being picked up in certain high-volume industries specifically, or is it across the board?
Aaron Wang:
In staffing, it has been really across the board. We are seeing strong uptake in white-collar roles such as IT, engineering, and healthcare. Even blue-collar roles, like light industrial, have been incredible because a lot of those folks don't even have LinkedIn profiles. It is very hard to assess those candidates online, so you have to call them. Those recruiters are often on back-to-back phone calls all day, so Alex can do that for you and let you focus on the best people.
Francis Larson:
AI is advancing so fast. Models like Opus 4.6 and Gemini 3.1 are coming out, and the scaling laws seem to be continuing. How do you stay on top of the explosion of competition and the ability of other companies to produce features?
Aaron Wang:
I don't think the world of business changes dramatically. Ultimately, you offer a good product, you take care of your customer, and the score takes care of itself. We can get into the weeds on tactical elements, but we do a great job of working in person and hiring world-class talent in Silicon Valley. Our engineers are some of the best in the world. We are well-capitalized and can invest in our customers and products. Ultimately, it's your ability to consistently deliver for the customer. If you can do that, you'll do a good job whether it's 2026 or 2036.
Francis Larson:
How do you think the role of a staffing and recruitment company will shift, given that a lot of the candidate side is being solved? What does it mean to be a staffing company in 2029?
Aaron Wang:
If you look back at history, are there more or fewer staffing firms today than 100 years ago? Obviously, there are more today, or at least the market value is higher. There were more 50 years ago than 100 years ago. The market cap of the recruiting industry is on one large J-curve. As technology improves and reaches more people, does that make a person more or less valuable? Steve Jobs said the Mac was like a bicycle for the mind. Well, what is ChatGPT? People will become more leveraged, more useful, and more valuable as technology improves. If people become more valuable, so too will the services that place those people. Because you are so valuable, I want to make sure you are placed in the right spot. In 2029 or 2049, the recruiting industry will only grow because people will be more powerful. The problem of recruiting them becomes much more valuable.
Francis Larson:
You recently wrote a book called 10X Recruitment. How did that come about, and what is it about?
Aaron Wang:
It's called The 10X Recruiter. You can find it on Amazon, or if you email me, I can send you a free copy. We wrote it to educate the market. We have learned a lot in building an AI recruiting agent. It is very difficult, and there are many "unknown unknowns" that are incredibly important, from compliance aspects to hidden costs. What happens when OpenAI goes down? How do you decide which model to use? What happens when countries pass new AI legislation? We wanted to make sure we weren't gatekeeping that information. The industry deserves to know what the best practices are for onboarding an AI agent. It has quickly become a very viral piece of writing in the industry, which we've been humbled by. Long-form content is really useful for clarifying thoughts.
Francis Larson:
It's a great introduction for people who don't even know where to start with AI in recruitment.
Aaron Wang:
That's right. We almost never talk about our company in the book. It's meant to be a guide for the things you should be thinking about. Our goal is to raise the bar for the entire industry. A rising tide lifts all boats.
Francis Larson:
Aaron, this has been super illuminating. I'm definitely going to look at the book; it sounds like something everyone in recruitment needs to read.
Aaron Wang:
This was great, thanks for having me.
