Interview with a chatbot part 2

These past weeks I’ve been deep into the #HRTech world, tweeting frequently into the void, trying to learn more about increasingly opaque data used in smart HR platforms. Throughout the process I’m documenting the variety of hiring technology on the market, from smart platforms to machine learning for automated resume screening to virtual assistants. Along the way I’ve stumbled on loads of chatbots trying to claim a place for themselves in the hiring process. I’ve got a bit of a crush on chatbot technology so I’ve been trying them out. Two weeks ago I pined for Mya but settled on an interview using TalkPush. Today I found Paradox.ai and gave it a go.

Once again, the intro starts easily enough:

Then I was immediately asked contact details. Mind you, I’m just browsing here, not actually ready to apply. I don’t know if it’s me or what but I’m a bit irritated each time I’m asked for contact details right away (side note: I signed away my LinkedIn data for Wade&Wendy access only to be told post-data exchange that I’m on a waitlist, so maybe I’m just tired of having to give up data to engage). But this is all pretend anyway, so I gave them my phone number and then we moved on to my interests.

I thought here that we might talk about what positions are available but the onus was put on me to define what I want. I actually wasn’t sure what to answer. I like that it’s framed that way but wonder how other job seekers perform when asked this question. Admittedly it caught me off guard and I wasn’t sure what to write. I had to think about it which then sent me down a mini-spiral wondering if they evaluated me on how long it took for me to answer.

Moving along:

This is where it got interesting. They’ll find my profile (and I wonder if they’ll find my other social profiles) for me, so I don’t have to submit anything.

I think there was a hiccup when I shared my non-existent most recent role as the interview ended abruptly. I can’t tell if it’s because a. it’s a bot. b. I’m not a fit. or c. this wasn’t an interview.

Then onto the questions from me. The Q&A started a bit rocky but got better:

So throughout this whole process I wondered: how do you know the difference between a bot that helps you explore job opportunities and one that evaluates you? I misunderstood this bot from the beginning.

I went into it thinking the bot was helping me explore options at the company. But it quickly moves into interview territory by asking my experience level. Then it forces me to automatically apply to the position(s?) we discussed as they pull my LinkedIn profile. What if I hadn’t updated my LinkedIn?

Also, what if I’ve already provided my real name and contact, but wasn’t prepared to discuss my experience, what do I do? If I abandon the convo, and return, how does that affect my evaluation? Am I more desirable because I’m returning? Or am I penalized because I couldn’t answer the prior questions?

I’m so curious what happens on the backend when a recruiter receives the data.

While this experience is certainly efficient it’s hard to get a feel for company culture during these interactions. I was generally curious about the companies that they partner with but didn’t get traction there. Asking about the workplace and getting a canned response about “best talent” and “superstars” doesn’t offer much. If Olivia instead shared a video from the team, or a blog post about a day in the life of a marketer at Paradox, or even a personalized message from the founder that wasn’t full of “superstar” startup speak, it’d instantly provide more value. It’d at least add a personal touch.

Interacting with bots has me wondering how we define candidate engagement within the context of chatbots. Olivia engaged with me but she wasn’t engaging (though she was definitely better than previous bots I’ve engaged with). When the novelty of interacting with recruiting bots wears off (it’s still so very new), I wonder how candidates will view the experience.  If there’s a war for talent, how do you expect someone to chose your company if you can’t show off goods? Do bots play a role in wooing candidates? Or are they just there to expedite the hiring process for HR?

And if candidates are expected to show their soft skills, how do employers expect to identify them when the majority of HR tech aims to take humans out of the selection process?

Ok, McKinsey’s Future of Work podcast is actually pretty good

I’ll admit that listening to consultants talk doesn’t strike me as good podcast content. My podcast list is overflowing with no shortage of new recommendations. Anything I add has to compete with mighty fine podcasts like 2 Dope Queens, On the Media, Note to Self, The Read, Reply All, and Teaching and Learning in HigherEd. So I was torn when I learned that McKinsey puts out a Future of Work podcast. Grant it, this is my favorite professional subject. But there’s so much fluff in future of work circles and not enough meat. Fun fact: being a futurist doesn’t mean you have to be right. You just need research chops, a regular content production schedule, a brand with the phrase “future of work”, and an audience who will listen. It’s not rocket science.

So I was skeptical. But the McKinsey Global Institute puts in the hard work that you’d expect for a top global consulting firm. Their reports on the future of work are insightful and meaty. Their podcast is no different. I was pleasantly surprised. And by pleasantly surprised I mean I was taking loads of notes and couldn’t stop listening. It’s not terribly entertaining and feels a bit like watching CSPAN. But the podcast brings their valuable research on the future of work to life. It also broadens their research (hopefully) to audiences beyond MBA students and upper management. Anyone who is curious about how their career is going to shift should give it a shot. It pairs well with public transit rides.

I listened to their most recent episode, How Will Automation Affect Jobs, Skills, and Wages?, and could have quoted the whole damn podcast. I held back. Here are some of my favorite meaty bits.

On lifelong learning from Susan Lund, a partner McKinsey Global Institute:

It’s something that has been a bit of a mantra in the educational field. Everyone is going to have to be a student for life and embark on lifelong learning. The fact is right now it’s still mainly a slogan. Even within jobs and companies there’s not lifelong training. In fact what we see in corporate training data at least in the United States, is that companies are spending less. As we know right now people expect that they get their education in the early 20s or late 20s and then they’re done. They’re going to go off and work for 40, 50 years. And that model of getting education up front and working for many decades, without ever going through formal or informal training again is clearly not going to be the reality for the next generation.

Honestly I could quote so much from this podcast. Instead of the common “robots are going to take our jobs” narrative, they dive deeper into the subject, discussing how occupations will shift and what that means for workers. I’ll just quote this entire response on acquiring new skills, again from Susan Lund:

“We categorized 800 occupations into 58 categories. This is our shorthand way of showing how work might shift between them. For instance there’s a whole classification around customer interaction jobs. And that includes cashiers, call service representatives, etc. By grouping occupations into these categories we can start talking about which ones are growing and which ones are declining. So that number of somewhere between 75 million and 375 million people [around the world] may need to switch occupational category, means that they’re in a set of occupations that are actually shrinking in number. Some of those people are going to have shift to one of the growing occupational categories.

This is a big shift. It’s different from saying I’m one type of specialty nurse and now I need to be a different type. That would be a shift within an occupational category. Here, the changes we are talking about are very significant. It’s about somebody who may have been working in trucking or manufacturing learning to do something entirely different. Possibly a job in construction or healthcare or other types of things. This will require more than simply applying for that job. It will require some level of formal training to learn the new skills to become qualified to get that new job. This will be the defining challenge of our generation, is creating the programs and tools and opportunities for someone who is mid-career with a mortgage, with children who can’t afford to go back to school for two years to get an associates degree or four years to get a bachelors, but helping that person get the bare minimum of skills they need to get their foot in the door in an entirely different occupation and start off on a career ladder in an entirely new direction.”

You have to teach people how to become lifelong learners. You have to change the old mindset. You have to teach them how to make occupational shifts. You have to prepare them with practical advice and skills.

This is why I founded FutureMe School. We have to reinvent old career narratives and train people to adapt to multiple occupational changes over a lifetime. Which is exactly what we’re doing at FutureMe School.

Stay tuned.

Using Google Chrome makes you a better employee?!

Just need to park a few nuggets somewhere until I can write a full post about this topic. I’m currently researching the use of new talent signals, data scraping, and machine learning in the hiring process. These excerpts come from the Journal of Industrial and Organizational Psychology, in an article titled, New Talent Signals: Shiny New Objects or a Brave New World?

On the use of big data in the workplace: 

So long as organizations have robust criteria, their ability to identify novel signals will increase, even if those signals are unusual or counterintuitive. As an example of an unlikely talent signal, Evolv, an HR data analytics company, found that applicants who use Mozilla Firefox or Google Chrome as their web browsers are likely to stay in their jobs longer and perform better than those who use Internet Explorer or Safari (Pinsker, 2015). Knowing which browser candidates used to submit their online applications may prove to be a weak but useful talent signal. Evolv hypothesizes that the correlations among browser usage, performance, and employment longevity reflect the initiative required to download a nonnative browser (Pinsker, 2015).

On using social media to evaluate candidates: 

“People’s online reputations are no more “real” than their analogue reputations; the same individual differences are manifested in virtual and physical environments, albeit in seemingly different ways. It is therefore naïve to expect online profiles to be more genuine than resumés, although they may offer a much wider set of behavioral samples. Indeed, recent studies suggest that when machine-learning algorithms are used to mine social media data, they tend to outperform human inferences of personality in accuracy because they can process a much bigger range of behavioral signals. That said, social media is as deceptive as any other form of communication; employers and recruiters are right to regard it as a rich source of information about candidates’ talent—if they can get past the noise and make accurate inferences.”

On the use of video interviews for voice profiling:

“Moreover, through the addition of innovations, such as text analytics (see below) and algorithmic reading of voice-generated emotions, a wider universe of talent signals can be sampled. In the case of voice mining, candidates’ speech patterns are compared with an “attractive” exemplar, derived from the voice patterns of high performing employees. Undesirable candidate voices are eliminated from the context, and those who fit move to the next round. More recent developments use similar video technology to administer scenario-based questions, image-based tests, and work-sample tests. Work samples are increasingly common, automated, and sophisticated. For example, Hirevue.com, a leading provider of digital interview technologies, employs coding challenges to screen software engineers for their software writing ability. Likewise, Uber uses similar tools to test and evaluate potential drivers exclusively via their smartphones (see www.uber.com).”

On new technologies barreling ahead without theoretical backing

“The datification of talent is upon us, and the prospect of new technologies is exciting. The digital revolution is just beginning to appear in practice, and research lags our understanding of these technologies. We therefore suggest four caveats regarding this revolution. First, the new tools have not yet demonstrated validity comparable with old school methods, they tend to disregard theory, and they pay little attention to the constructs being assessed. This issue is important but possibly irrelevant, because big data enthusiasts, assessment purveyors, and HR practitioners are piling into this space in any event.”

I’ve said it before but the candidate process is about to get far more opaque.

Interviewing with a chatbot

I’m pretty obsessed with HR tech. Right now I’m examining HR chatbots. Chatbots are becoming ridiculously popular in HR because they save recruiters valuable time. Candidates are comfortable with them too. SHRM reported that 57% of survey respondents confirmed they were fairly or extremely comfortable interacting with AI bots in the recruiting process. In an ideal world the chatbots deliver a premium candidate experience, giving everyone the opportunity to engage with the company.

I’ve been trying to get a look at Mya, the chatbot that interviews you, analyzes and scores your responses, and does the heavy lifting for recruiters. I really really really want that damn bot to interview me in part because I have a crush on its conversational design technology and because I secretly want to be a part time conversation designer.

Natural language processing Mya

But I can’t get Mya to interview me yet as I’m not exactly qualified for their jobs, so no bot access.

Luckily I stumbled on a chatbot for TalkPush, which is a HR chatbot company that makes it easier for recruiters to source. Here’s their pitch:

TalkPush is the first conversation-driven Candidate Relationship Management (CRM) system. On Talkpush, recruiters spend more time talking to qualified candidates, which translates into a better candidate experience and huge reductions in cost-per-hire and time-to-fill”

TalkPush has received about $1 million in funding. Compared to Mya, which has received $34 million and on its Series B funding, TalkPush hasn’t gotten much of that sweet 2 billion dollar plus HR tech funding pie.

Regardless I’m on a mission to try out hiring chatbots and TalkPush made it easy. You can access their recruitment bot from the homepage but you’ll be required to give access to your FB page, an offputting ask since we literally just met (what are you doing with that data?).

However their jobs page gives you access to their bot without having to give over your FB access.

The interaction starts simply enough.

HRChatbotinteraction

HRChatbotInteraction

I was a bit put off by the questions about contact info before I even got to see the jobs. Then again I’d submit this info on a resume, so I suppose it evens out in the end.

There were also a few typos in the exchanges. I thought it reflected poorly on the company since in the US, typos result in the immediate death of your candidacy for employment. Then I found out the company is based in Hong Kong and China, so the conversation is likely written by a non-native speaker. Makes sense now (I see an opp for some talented bilingual interns!)

But then the experience gets a bit lame.

HRChatbotexample

This wasn’t so much a conversation as it is a display of information. After I asked what a (Demo) Job Position is, I got nothing. Radio silence. So I checked in.

HRChatbot5

Aaaaand it seemed like we were back at the beginning. This reminded me of the awkward experience I had with IBM Watson’s chatbot. So I typed in an alternative job. Then we were back on track. Interestingly I was supposed to tell them about my relevant work experience. I wondered what they meant by that. Should I write paragraphs? Is this the place for a short, two-sentence summary of my qualifications, similar to what I’d write in a cover letter intro or say to a recruiter on the phone? Should it be a list of companies? Positions? Software? Skills? Am I being evaluated on the time it takes to answer questions, similar to video interviews? I had no idea. So I just added some fuckery to advance the convo and see the other questions.

HRChatbot hiring future of work

HRchatbot example

As you can see I wasn’t a good candidate. But the questions were interesting reminded me of an initial phone interview.

The flow and expectations in this exchange are a bit problematic though. As a job seeker, I may not have samples of my work or a summary of my work experience on hand during the chat interview. Since I had to put in my contact information before I saw the jobs, it’d be hard to take a break, go get that information, and return to the bot to continue the conversation. Usually a candidate takes a look at a job, builds the required documents based on the job, then returns to submit. This experience was like an application and phone interview in one.

Also the inability to engage at times makes me wonder what candidates should do when a bot fails. Any interaction with an employer should be considered evaluative. This leads to questions about best practices for candidate behavior. What should a candidate do if they’re stuck engaging with a bot? If the chatbot fails, what are the next steps for the candidates? If the chatbot misunderstands their information or can’t answer a question, does the candidates get bumped to a human? Can a candidate press 0 or some magical combo to get a live human?

If I were advising a candidate, I’d tell them to take screen shots and contact a recruiter directly with questions. It’d work in the candidate’s favor maybe. It could show that the candidate is a proactive, problem solving candidate. And it’d (hopefully) help the team improve their bot.

So for all job seekers out there: brush up on your professional written communication skills. You’re going to need them beyond writing cover letters to get past the bots.

The pain of upskilling

The benefits of the comfort zone are appealing. Steady (though not always satisfying) incomes, “secure” jobs, relaxed routines, and predictable schedules are as comforting to humans as they are to animals. In this phase, people limit their learning to things they learn on the job, not knowing that yesterday’s lessons rarely solve tomorrow’s challenges… Without skill upgrades or a willingness to learn, people are caught in a rut. They are unable to see when the next trend is about to catch up or when the current one is about to die. For the few that can see the new trend, the pain of having to upgrade their skills far supersedes the pleasure of staying in the comfort zone.- How to stay relevant in today’s rapidly-changing job market

Since reading this article I’ve thought about the above paragraph multiple times. The last part about the pain of upgrading our skills nailed it.

No doubt, professional change is painful. I’m part of a generation where the narrative has always been college degree = career success, full stop. Two degrees and five professional jobs later and I’m wondering if I’m staring at irrelevance in five years if I don’t upskill. I quit my well paying, secure job at Yale last year because I was stagnant with little hope of gaining new, relevant skills that prepared me for the future of work. While I’m starting my own company, I’m concerned I’m not keeping pace with the technical skills needed to stay relevant. Should I take a side job designing chatbots? How can I fit in learning to code in python so I can get closer to working with AI systems? I’m not thrilled by self-paced learning, so what are my options? Where do I find the time?

Telling people they need to update their skills and #alwaysbelearning is the first step. But the next step is harder. How do we teach people reskill? How do we help them identify what to change and how to change it?

That’s what I’m setting out to change in my upcoming book. If you’re curious, subscribe.