The conversational interface is coming, and it’s coming fast. Remember when in 2010, about three years after the iPhone was launched, people were telling you mobile was going to be it? And you didn’t believe them? And even though you recently tried to upgrade your corporate careers website to be mobile friendly, it still isn’t really delivering results? It’s that time again and now it’s called the conversational interface — chatbots and voice assistants. Just like with mobile, we don’t know exactly what will happen and how it will look. We do know this is the new way to communicate.
Understand how this works. Learn and adapt so you have the skills to build a great candidate experience when, like with mobile these days, 50 percent of your potential hires demand this channel. Right now you can still make mistakes. In two or three years, you can’t.
My advice: start with a chatbot right now. That’s the easiest thing. Voice assistants are going to be big too, no doubt about that. The adoption rate is huge, and the way the interface works is so natural. The way a voice assistant works is basically the same as a chatbot; you need to create a great conversation. Since chatbot technology is easier, more people use chatbots, you get more feedback, and you can improve faster. Hence: start with chatbots.
What type of jobs are suited for conversational interfaces? Well, basically all of them. I know Facebook uses a chatbot for recruiting technical talent. But the most profitable right now are:
- Jobs with a great international reach
- High-volume vacancies
- Internships and traineeships
In the examples below I’ll explain why.
Example 1: informing PhDs
The University of Twente in the Netherlands is, as far as I know, the first university in the world with a chatbot for recruiting. The bot is used to answer questions about doing a PhD in the Netherlands and specifically in Twente. Full disclosure: I was a consultant with this project.
PhDs can apply to jobs all over the world. In the Netherlands visas are needed, but for a PhD that is never a problem. However, students have many questions. How much do I get paid? How much tax will I have to pay? Will you help me find a place to live? All kind of simple questions that can easily be put into a chatbot.
The chatbot doesn’t care about time zones. That’s important when you have applicants from Japan and Korea to Mexico and the east coast of the U.S. Of course, your website doesn’t care about that too, but a great, user-friendly FAQ website is almost impossible. People search in different ways; a conversational interface can help find the best way.
Example 2: selecting interns
The other example unfortunately went offline recently. It might have something to do with the HR manager, who was the big sponsor of the project, changing jobs. But it was great. This German incubator worked a lot with interns, and the HR manager found she was always asking the same four questions at the beginning of the interview. Basically, she based the entire selection on the answers to those four questions, since with an intern, the resume can usually only say: I’m studying X at university Y. So the chatbot asked those four questions. She now made the selection who to interview based on the answers. The interviews went much deeper, and there was more time to check if someone truly fitted with the company.
So if a resume is useless, with students and starters, a chatbot can ask the questions you usually start an interview with and help you select better candidates.
Example 3: high-volume recruiting
Anchor is a British Elderly care facility using a chatbot in a really great way. This is high-volume recruiting. It needs a lot of people, and you don’t need formal training to get started. With unemployment being at low levels, it is willing to train people who are interested. For this type of work, jobsearches are verly local. So it builds that in as well. it made this video about how it works.
In this case of high-volume recruiting, a resume isn’t always very useful if you train people yourself. You need motivation, and a chatbot can ask the right questions to uncover it. Just with the interns.
But there was another aspect here. Many people don’t believe they qualify because they don’t have the right diploma. So they don’t apply. The don’t even call, since they do not want to waste the precious time of the recruiter. And even though a website says otherwise, they don’t believe that. They need to hear it from a person … or a chatbot.
Another prime audience for chatbots in my opinion is trainees. Trainees, especially for international traineeships, can get an answer to all their questions in their own time zone. And international trainees have plenty of questions about working conditions and other nation specific things like taxes and visas. A trainee, like an intern, usually has a resume with absolutely nothing on it that has any predictive value in selection. So you need to select on other things, like motivation, and that’s much easier done in a conversation than by reading a 20-times re-written letter.
A great thing about chatbots is that it’s possible to make candidate-based alerts instead of company-based alerts. Let me explain based on the data that I gathered as a consultant for the PhD chatbot (example 1). There are three kinds of PhDs in the world:
- Those who decided they want to do a PhD.
- Those who never considered it until their professor asked them
- Those who are just looking for a job and figure that a PhD is just another career option.
The chatbot focusses on the first category: the PhDs who make the conscious decision they want to do a PhD. This group sometimes starts researching their options between two years to six months before graduation. Professors expect PhDs to start within three months (usually even faster) after the application procedure is finished. A chatbot can ask the questions about when a candidate expects to graduate and, if the candidate wants to receive one, send an alert about three months before graduation asking if the student is still on track and if he or she wants to look at open vacancies. This way it’s not a organization-centric alert such as “we have a vacancy now, let’s alert everyone.” But a candidate-centric alert is about the candidate being be ready to make a move — let’s see if we can help him or her. It helps you keep in touch with candidates who liked you when they were doing early research, but later forgot about your organization.
I assume, but I haven’t researched this, that the same timeline might apply to trainees. Since many companies have multiple starting moments with traineeships and you should always try to get the best talent in, not just the talent that’s ready when your traineeship starts, this would be a valuable option as well.
Assessment by chat
Another part of conversational interfaces I’m really enthusiastic about is the possibility a company like Talent Swot offers. Its chatbot software is loaded with assessment tooling based on LIWC (Linguistic Inquiry Word Count). Simply explained, based on the words you use, how often you use them, and more, it’s possible to build a scientifically sound personality profile. So from the chat you have, the answers to questions you’ve asked, you do not only get the answers, you also get an unbiassed personality report of the candidate. Now that’s added value from using a chatbot for selection, right?
So what about those voice assistants? Well, to be honest, I have no idea how we are going to use them. They are the next interface we are going to use. Just about everybody that uses them is very satisfied. Kids understand them right away, a sign that the user interface is very natural. And they basically use the same conversational interface as chatbots. One of my developer friends recently showed me a tool where you can build a conversation and export it to either Facebook Messenger or an Alexa Skill. The interface is the same, it’s about creating a great conversation. If voice is the new interface, chatbots are the way to learn about it. That’s why my advice to everybody right now: start building conversations with your candidates trough conversational interfaces.