In the May 1, 2018 Ask The Headhunter Newsletter, new technology levels the playing field between employers and job applicants. Enter the job seeker’s best new friend! Interview ON!


For over 20 years the Internet and job boards have made it possible to maximize our chances of landing a new job because those jobs databases enable us to apply for 1,000 jobs instantly with the press of a key. Some “intelligent job agents” will even retrieve and e-mail me hundreds of matching jobs without my having to do anything but sign up.

But no one has solved the problem of how to actually get in front of loads of employers all at once to interview with them at high speed and in large numbers. It doesn’t matter how many jobs you can apply for. The real challenge is to be able to respond to all those recruiters who contact you, and to have lots of interviews quickly. Is there anything on the horizon?

Nick’s Reply


Interview ON.

Hold on to your seats — it’s here. A new start-up company has created a new way to help job seekers navigate the job market at incredibly high speed. The technology is called Hank, and he enables you to interview for a job with as many as 1,500 companies in a single work day. He sits for screening interviews on your behalf with potential employers at a rate that would take most job seekers months to match.

He even sends customized follow-up e-mails back to his interviewers.

Perhaps more attractive — as far as his job-seeking clients are concerned — is that Hank works full-time for free and never yells at annoying recruiters.

Your job-hunting avatar does your interview

Hank’s secret: he’s not human. Officially known as “Avatar Hank,” the master job applicant is an artificially intelligent software technology that uses machine learning, allowing him to refine his conversational skills with more practice. At the moment, Hank is being used by several hundred job seekers to simplify the ongoing hunt for new jobs, according to Alex Kotts, who co-founded Avatar Hank with several partners in 2017.

“We wanted to create something that functioned like Uber for job seekers, but instead of calling a car, a person would be able to call a pool of companies to get a job,” said Kotts. “Right now, we have several hundred job seekers using Avatar Hank, which means Hank is doing about 50,000 interviews a day.”

Hank has the ability to speak at different speeds and sound like a man or a woman, depending on his job seeker’s preference. Kotts said the software is most effective for job seekers who apply to large numbers of blue-collar jobs, such as sales clerks, baristas, and construction workers, “which is where all those new jobs you read about are actually happening.”

How Hank works for you

The process starts when job seekers provide Hank with their LinkedIn profile and with titles and descriptions of the jobs they want. They can even specify which companies they want to work for but don’t have time to contact. Then Hank does the rest.

Get this: The avatar has an API (Application Programming Interface) that is linked to the leading ATSes (Applicant Tracking Systems) that employers use to interact with job applicants. These include CareerBuilder, Indeed, LinkedIn, Taleo, ZipRecruiter and half a dozen others.

Hank submits thousands of job applications directly to those systems per day. Recruiters have no idea where those applications are actually coming from — they assume it’s a human. When a recruiter responds, Hank intercepts the e-mail, “reads” it, and instantly generates whatever follow-up information recruiters demand — resumes, cover letters, references, salary requirements. The real magic is in the API access — Hank also fills out those pesky online job application forms that recruiters demand. (Talk about the job seeker’s revenge!)

Hank talks

When a match occurs and a recruiter actually wants to talk with the job seeker, the call is routed to Hank, who handles the conversation. This is where the technology kicks it up a notch.

In the interest of full disclosure (and of legal requirements) he says “Hi, my name is Hank, and I am an avatar. I will answer all your questions about Nick factually and completely, as if you’re talking directly to Nick. Are you still looking to fill this position?”

Kotts says, “If the answer is yes, Hank can handle the entire screening interview over the phone or by video interview. Our analysis shows that recruiters usually ask very few questions, and they’re simple, because recruiters don’t really know anything about the jobs they’re filling. We’ve programmed Hank to exploit this. Just like job postings are designed to lure the maximum number of applicants, Hank tells recruiters what they need to hear to increase his hit rate — the frequency of requests for in-person interviews.”

While the average phone screen typically lasts about eight minutes, Hank can talk for 16 minutes if necessary. He is also often able to ask pre-determined questions on behalf of the job applicant. Call analysis reveals recruiters respond best to the question, “Have I answered all your questions?”

Accuracy is good enough

Right now, claims Kotts, the software is able to respond accurately 82% of the time, a number the company expects to increase to 85% in the next few months. “That’s good enough for now,” said Kotts in one article, “because recruiters’ software is less than 20% accurate when picking job applicants to call.”

After the phoner, Hanks analyzes the typical interview in less than 900 milliseconds and passes promising job opportunities directly to the human job applicant in the form of a detailed report. The human, of course, makes the final decision about a job.

Saves time and frustration

The technology’s primary benefit is that it saves job applicants time. Kotts said human job seekers waste hours filtering through job postings that are no longer available. He said job seekers often must answer 100 phone calls from recruiters just to find one job that’s actually a fit for them and pays their desired salary. (I think we can all corroborate that!)

One article about the Avatar Hank technology quotes a job seeker: “Recruiters waste my time. This was what drove me to try the new approach and use Avatar Hank. Now I have my own weapon!”

Kotts said Hank’s inventors have been surprised that recruiters often prefer to conduct interviews with Hank than with human applicants. “I think they feel they’re getting more accurate answers because they feel the algorithms will tell the truth.”

Kotts said, “What I see is that job seekers will begin managing AI more and more and using it as a tool to avoid wasting their time with all the tire-kicking recruiters who constantly contact them about the wrong jobs. Hank gives job seekers automation to respond to the automation used by employers and recruiters. Hank levels the playing field.”

Will employers interview Hank?

Kotts is circumspect about how employers in general will react when Hank is rolled out in three months to Facebook’s 2.2 billion members worldwide. Will recruiters interview Hank instead of a human?

“What are they going to do? Complain about automation?”

How will HR deal with automation in the hands of job applicants? I asked a top HR exec at a Fortune 50 company that question.

“Oh, my Gawd,” she said. “Payback is a bitch, I guess.”

Do I have your attention?

What if Avatar Hank were real? He’s not, of course — but only because job seekers can’t afford to spend the billions of dollars HR dumps every year into “recruiting automation.” Otherwise, HR technology companies would create him.

Unfortunately, there is no “job seeker’s revenge technology” to match the ATSes and goofy “algorithms” that HR sics on job seekers. I made it all up, but there is truth to Avatar Hank — a lot of truth.

robotHank’s evil sister

I made this all up for a reason. I stole the story of Hank from Peter Holley’s April 25, 2018 article in The Washington Post, Want to work for Ikea? Your next job interview could be conducted by a Russian robot.

But in Holley’s account, there’s a real robot named Vera that reportedly interviews about 50,000 job applicants a day, enabling HR departments to nap while job applicants sweat out 8-minute phone calls with a cartoon. Nobody’s making that up. Employers are paying to use Vera on real people.

That’s 833 personnel jockeys dozing eight hours a day (without time off for lunch) while 50,000 suckers are required to talk to the robot hand if they want a chance at a real job interview.

Now, what happens when nuclear HR weapons are put in the hands of — gasp — job applicants?

Hey, HR!

So my evil purpose in this week’s column — I don’t think I’ve ever fabricated a whole column before — is to wonder out loud how HR would like it if we deployed Avatar Hank against employers the way employers deploy Vera and robo-recruiting avatars like ZipRecruiter, LinkedIn, CareerBuilder, Indeed, Taleo and their ilk.

Apologies to The Washington Post and Peter Holley for satirizing their article to make a point, but thanks to them for shining a light on Vera, the spawn of HR technology. We know she’s not real because if she were, she’d start every phone interview with, “#MeToo!” There is no one named Alex Kotts and no robot named Avatar Hank. You’ll have to read Holley’s article to learn who’s behind Vera.
How would HR like it if “the talent” refused to appear in person — like recruiters and hiring managers refuse to appear in person — until the employer talked to the cartoon hand first, and filled out the forms, and got diddled digitally?

How would HR like it if the next 50,000 job applicants it called to conduct phone interviews were robots? Would job seekers’ robots be any less legit than Vera?

Hey, HR, can our robots have phone sex and produce skilled offspring to do your jobs?


Hey, Boards of Directors

Vera and Hank tell us one thing: It doesn’t take any brains to interview 50,000 job applicants or to interview for 1,500 jobs.

HR, employers, corporations invest billions of dollars every year avoiding using their brains — they spend it on what’s plainly stupid, laughable, and counter-productive “technology” that they’d never abide if subjected to it themselves. (See HR Technology: Terrorizing the candidates.)

Managers have destroyed any chance of matching the best workers to the jobs they need to fill because they refuse to show up. They deal in avatars, robots, algorithms, HR technology. They deal in keywords, automated job applications and programmed applicant “assessments.” They’re trying to wash their hands with rubber gloves on, to recruit without recruiting, and to identify the best candidates by rote. (Contrast: Smart Hiring: A manager who respects applicants.)

How do we point out the real problem with hiring?

It doesn’t take any more than flipping around Peter Holley’s account of Vera technology. We subject employers to fake job applicants, like they subject job applicants to fake “selection processes” via robo-forms and algorithmic judgments. We deploy cartoons to apply for jobs and to “show up” to be phone screened by recruiters.

The boards of directors behind these companies reveal that they are the truly unskilled and clueless stewards of industry. Would you have lunch with a cartoon character to talk about the future of your business?

You deploy a talking cartoon character to judge whether a person is worth interviewing for a job — then you report to your investors that there’s a talent shortage?

Go ahead. Look us in the eye and say, HR technology — then realize you and your robots are talking to our robots.

Interview: ON.

How does HR learn a lesson from the stupid HR technology it foists on job applicants? Can job applicants turn the tables and make HR eat its own high-tech dog food? Will a tech company create Avatar Hank and make recruiters talk to the robo hand? What can job applicants do to even the playing field — do they have to dumb the game down to HR’s current level, or is there a way to raise the ante and the standards?

: :

  1. You had me going with “Hank” for awhile. Good story.

  2. In the future, when all interviews are conducted with AI interviewers talking to AI applicants, everyone will be out of work and companies will complain about having 0 qualified applicants.

  3. Actually, all joking aside, companies should be scared of using AI for HR purposes.

    AI algorithms are, in part, only as good as the training data you feed them. If you introduce a systemic bias into the training data, you’ll get that in the AI. I remember reading about how some AI trained to classify an image came to rely on the fact that all the backgrounds of the training images were similar and ended up using that, not the item itself.

    If they’re not careful, bias for or against a particular gender/race/national background/veteran status/etc. will end up in the algorithm. And when said system conducts 50,000 interviews, you’ll have a very large population set for analysis. Then, when a certain sub-population is, say, 20% of the applicants but makes up only 10% of people offered jobs, you’ve got a pretty good argument for discrimination.

    Yeah, companies can (and do) do it now, but if you only interview 15 people, it’s hard to prove.

    Finally, AI can have unintended consequences that satisfy what you’re trying to accomplish but in a very sub-optimal way. For example, AI in a particular application was used to try to build a mechanical design that would traverse a distance in the shortest amount of time. Know what it did? It ended up building a giant “beam” that simply fell over to cross the finish line.

    Look for something similarly goofy to happy with HR systems sooner or later.

    • More worser than that is that “AI” (it’s not AI, it’s crap) cannot tease apart the subtle cues that tell a trained human which applicants should make the first cut. Anyone with knowledge of signal processing will tell us that poor pre-processing results in corruption of any output — that is, you wind up with questionable candidates and you likely reject some of the very best.

      “More worser” would likely be flagged on my resume and get me rejected. What else do they ever learn about me?

  4. I am not getting this. Isn’t the point of ATH to encourage the job seeker to find one company that you most want to be a part of, due your due diligence to get the interview and then get the job?
    Not sure what I am missing about the shotgun approach to getting hired? Then again, I haven’t had coffee yet. That’s probably it. sorry, my bad.

    • @Tony: You probably missed the part where I said, “I made this all up for a reason. I stole the story of Hank from Peter Holley’s April 25, 2018 article in The Washington Post, Want to work for Ikea? Your next job interview could be conducted by a Russian robot.”

      Getcher caffeine, boy :-)

  5. Could only wish! Give HR a hefty dose of its own medicine.

  6. I have learned from my experiences with part of the HR world, that many companies will minimize their HR staff (budget constraints..?) and yet still try to accomplish the “goal” acquiring qualified job candidates. Combining this with the current mentality of “I bet there’s an app/software that could handle that task” the HR system now seems to be less about effective human interaction and more toward just processing data, requiring minimal effort, from the comfort and anonymity offered from behind (in front of) the screen…

  7. Oh man, this shared the sh*t out of me for a minute.

    But, did y’all know there’s someone that (sort of) did this already?! It’s amazing.

    • Kimberlee – thanks for that link. Great story! But it also reveals how so many smart people in tech get deluded into thinking the challenge is not to get in the door by being the right candidate. They’re all focused on getting in the door by beating the algos. They miss the point. This guy seems to come to an epiphany – that it’s about talking to people.


    • That’s fascinating! This guy’s experiment really shines a light on the inadequacy of ATS and the “common wisdom” approach for job hunting. I’m confident that the only reason I was able to get a job where I’m at now was because they didn’t use ATS.

  8. Chris is correct: The spawn of Hank and Vera would be an evolving system of bots that are designed to tell each other what they want to hear. As the HR bots optimize to find the purple cow with the least amount of effort, the Job App bots will learn how to adjust in milliseconds to provide exactly the right information in exactly the right format to give the HR bots “technical ecstasy.”

    In the meantime, the rest of us can re-learn how to meet people face-to-face and how to do the real jobs we need to do. Because I think people have it backwards: Many people are worried that automation will take away jobs. True. But I believe that if a job can be automated and done by a mindless, soulless machine, then it is beneath human dignity to make people continue to do that job.

    • @Michael: I wasn’t kidding about a tech firm creating Hank. It would be no harder than creating Vera, or a digital one-armed bandit.

      Would Hank work? Who cares. Hank would create havoc. My bit about giving Hank to all FB users is my real fantasy…

  9. I checked the date to see if I was one month off on this one. Great story.
    I can see Hank coming. If companies who don’t want to recruit intelligently post their jobs on dozens of boards, then thousands of applicants will apply. Vera is needed when HR can’t screen all the resumes effectively. And a tiny fraction of those who apply ever hear back, so anyone who applies to a thousand jobs to get a shot needs a tool to handle a dozen interviews.
    I bet Vera just adds one more layer of interview, since I doubt anyone is going to bring in a candidate interviewed only by machine. SAG (the actors union) requires that actors get paid after a certain number of auditions for one job. Maybe the general industry needs the same thing.

    • @Scott, yah would have been nice if Holley’s article had appeared end of March so we could have an April Fool column…

      The spiral you describe is kinda like mining for bitcoin. Cost goes way up but supply is fixed (like jobs).

    • @Scott: That’s what I was thinking! This would have made a great April Fool story.

      I read the story about Ikea and Vera, and although we don’t have a Hank yet, I wonder how long it will be before someone develops Hank and offers it to job hunters. If the fee to use Hank is low enough, job hunters could truly wreak havoc on employers, HR, and all of the companies to which they’ve outsourced recruiting and hiring. Revenge!

      Yes, I think you’re right–Vera is just one more layer, more distance, between you and hiring manager. Just like ATSes, HireVue, and others. And it is another way for employers to avoid meeting, talking to, and dealing with prospective employees. None of this makes any sense.

      • There IS a Hank (but not called Hank). Its creators have decided not to let him into the wild for fear of the enormous disruption it would cause. Heh-heh. I say bring it on. I’m in discussions with them. They’re not hackers. They’re in the career business and built their Hank as an experiment.

  10. AI is moving from the lab to the workplace, with profound implications for business and society, per the November 2017 McKinsey Report.

    We are going to be psychoanalyzed by the algorithms and get labeled all for the wrong reasons – this makes some employment lawyers nervous about lawsuits.

  11. Reminds me of a colleagues take on voice mail when it came into its own. Two people crossing paths and one saying “have your machine call my machine & set up a meeting”

  12. For the record, I knew the minute I started reading this you were joking.

  13. You know, it would be nice if they would automate the interview and application process at least. That way, when we get rejected for not having the “right fit’, not disclosing salary, too many gaps, not having the right keyword, etc, we could at least not waste 45 minutes – 2 hours filling in those forms.

    That way, we can automate the applying and the rejection and save lots of time on human interaction!

    (BTW, I was being facetious, not serious.)

    • BTW, in all seriousness, I did love the humor of your article. I was just mocking the idea of HR automating things and suggesting that we do the same, without having read the full article, which happened to be about just that.

      However, it might be interesting if we DID create an automated thing and sent in fake data with bots sending in loads of resumes all over the place to see if people really are being discriminated against for national origin, age, etc.

      • Paul: “Hank” has in fact been created and tested by a team of tech folks I know. He works. The firm is reluctant to release Hank into the wild because he would create havoc.

        I’m trying to get them to demonstrate the concept. Or unleash Hank altogether. The HR tech in place now needs an antidote, no matter how painful.

        • They definitely should….and combine it with recent developments that allow people to create fake videos of others. People could map their faces to Hank’s so they appear to be actually doing the interview.

        • I agree, they should just to show how silly the whole process has become even as a proof of concept.

        • Also, the phrase “Unleash Hank!” is pretty funny, especially if you say it with the same dramatic flair as “RELEASE THE KRAKEN!”

        • @Nick

          If by “havoc” you mean that, when the dust settled, real humans would actually read real resumes and talk to other real humans about real jobs …

          • @L.T.: No, by havoc I mean that, while the dust is swirling, the real humans in HR will have no idea how to deal with all the “qualified” candidates their ATSes will “identify” among the Hank-generated applications, which will be perfect. HR won’t know who is what or whom to interview. Their systems will explode.

  14. Nick, your dystopian parallel dimension seems so real that … oh yeah, it IS real.

  15. You might think this is never gonna happen. Well. Think again

    Google is already working on it. In fact Google Duplex gets quite close. Bit of tweaking and it might just be able to do the interviews for us.

  16. Man, I was excited about a robot who could fill out the ATSs for me. :(