The Fascinating World of Smart Chatbots and Recruiting
I am fascinated by the Chatbots Revolution. And yes, we can call it a Revolution as it seems it is turning out to be even greater than the apps revolution! The numbers of chatbots hosted on Facebook Messenger alone are over 100,000 within 1 year! Chatbots technology is hailed as the disruptive new comer and it seems it will be greater than anything we have witnessed to date. I will add on the same note that not all is alive and well in the chatbot kingdom as Facebook also released that these bots have just hit a 70% failure rate…But I will be very forgiving regardless as I believe it only mirrors the natural process of a technology evolution (According to Gartner) – Innovation, inflated expectations, disillusionment, enlightenment and finally a ‘productivity plateau’, which refers to mainstream adoption. It seems we are now experiencing stage 3.
I will remind you that downloading apps was going through the same process, now that the hype and excitement is over it seems that out of the dozens if not more apps we downloaded – We actually don’t use more than 6 apps on a daily basis…(I have 8 screens on my iphone, loaded with apps I hardly ever use..)
All this compiled with the fact that messaging apps have an unbelievable adoption rate (over 2.5 Billion People hv at least 1 messaging app and many using several on a daily basis..) can only point out to what will, for sure, become a norm in the not-so-far-away future.
Pondering and wondering about this topic and how this revolution will affect recruiters and sourcers, I decided to give my friend Alon Mei-Raz, a Chief Product Strategist at HPE and a very sought after machine learning & chatbots key note speaker, a call and luckily managed to lure him for a quick lunch meeting with me so I could get his view on this topic.
Alon quickly distinguished between the “gofer bots”, “stupid bots” or “purpose based bots” – Simple decision or rule based tree bots that answer simple questions and carry out simple tasks and cannot answer anything it wasn’t designed to (reach a dead end) and the rising smart bots, based on machine learning, big data, AI and other advanced technologies that not only conducts a discussion in natural language but also “observes” the conversation and provides insights.
For example, when we tried to find examples in the recruiting world Alon mentioned the goal to be able to write or speak to our bots in natural language – Find me a “pretty senior” candidate with “vast” experience in software management and have the bot answer back to us – Here’s that candidate, but I also have this candidate, who doesn’t comply with all of the criteria but based on what you said about this company in the past, I think he/she could be interesting for this role.
Alon points out that the reason there is such a dramatic surge in smart and insightful chatbots is that it’s a good time now – AI has been around for many years but in the last couple of years, it reached an accuracy level as never before. Data quantity also grows exponentially. Numerous cloud based machine learning platforms were introduced lately offering strong capabilities, such as NLP (The natural language technology), sentiment analysis and more. And of course, the great adoption of messaging platforms (such as Facebook Messenger), that host large number of bots as a one stop shop.
Let’s talk abit more about smart bots. A smartbot is a bot that has added value – it is INSIGHTFUL. For smart chatbots to be good – you need to teach them. They need to learn. Sometimes they can arrive with knowledge from the “factory”, but you need to customize it to your own needs (the analogy that comes to mind is from the movie “The Little Shop of Horrors..“Feed me Syemour!..Feed me all night long”…).
Another feature to help it succeed is PERSONALITY. There is an ongoing discussion on the personality of chatbots. We are now also able to “emotionally charge chatbots”(Work of Hao Zhou and his team at Tsinghua University). For me that means we are getting one step closer to saying “empathy WILL PROBABLY be automated”. Users will stop using a boring bot that doesn’t emotionally engage with them. An emotionally aware chatbot could significantly change the recruiting industry! Scripting an engaging bot conversation that makes users come back for more is the task in hand. “Personality” in the bot world means defining the purpose, matching required characteristics and a suiting INTONATION. They also have to be ENGAGING and prepared for “what’s the meaning of life” and “is there a god” questions.
The bottom line is that the recruiter or sourcer’s “Minny Me bot” will have to be engaging, have a great personality, empathic intonation and insightful. Piece of cake…
“The basic added value and purpose of a chatbot is to replace human repetitive tasks” Alon said. “Bots don’t have a schedule, they are available 24/7/365 and they don’t require an employment agreement…” Where does this formula apply in the recruiting world – What are those repetitive tasks that we carry that can be replaced by smart chat bots. Without thinking too much, here’s several such tasks…
1. Searching my ATS for potential candidates
2. Screening and reviewing CVs and profiles
3. As much as I hate to say it…Searching Linkedin and the web for potential candidates
4. Engagement – Initial and screening phone calls to potential candidates
5. Rejection letters
7. Maintaining an ongoing relationship with candidates in any one of the recruiting stage
In order to understand what is already happening in our space, I stepped into the web, as well as spent time on several demos. I am not mentioning the simple bots that mainly deliver roles according to your search criteria. I wanted to see what are the more advanced options out there today.
1. Mya – By FirstJob – Someone wrote about her “The moment I met her, I was hooked. She was sensitive to my needs, responsive, inquisitive, direct, and had a sense of humor. Most importantly, she improved my chances of getting the job I want. The only problem I’ve encountered so far: I won’t be able to take her out to lunch to thank her”.
Mya boasts a whopping 75% automation of the recruiting process. improves recruiter efficiency by 38% and increases candidate engagement by over 150%. Even more impressively, candidates are 3 times more likely to hear back from a recruiter if they respond to Mya’s questions. She instantly engages with applicants, poses contextual questions based on job requirements, and provides personalized updates, feedback, and next-step suggestions. She is able to answer any question a candidate has related to the employer, including topics about company policies, culture, benefits and even the hiring process. When she can’t answer a candidate question, she queries the recruiter, gets back to the candidate, and learns how to respond the next time. Ultimately, the platform takes the data that Mya obtains through her conversations and turns it into quantifiable intelligence around the engagement level of each candidate and how closely they fit the target profile.
2. Job Pal – This white label chatbot is available on FB Messenger, Skype, Slack and Telegram. Their machine learning and NLP powered chat bots enable companies to automate candidate engagement, pre-screening and interview logistics. At this point their chatbot is not very intelligent but it is learning and growing as we speak…Everytime it encounters a question it cannot answer, it is forwarded to the recruiter who answers the question and the machine, in return, learns how to answer next time.
3. Lara Bot – Lara Bot was launched by the dating service Match.com. It communicates with more than 280,000 users per month in 10 different languages. At some point the company was looking for an intern and decided to create a new mission for Lara: ‘Lara the Recruiter’. It was finished in one weekend 🙂 Rather than using a ‘classical’ method for the internship, sending a CV and motivation letter, they told the applicants to use Lara. To their grand surprise, within a few weeks, Lara had interviewed around 400 applicants, around a hundred of them serious candidatures. Very rarely are the bots capable of “undersanding”, and thus of responding to open-ended requests. First, Lara detailed the mission of the internship, which teams would be involved and the practical details. Then, she would ask some questions, open-ended or closed (“Why is this internship right for you?” etc.).
4. RAI – From HiringSolved – Rai is an experimental AI interface. You interact with it like a human assistant. It can find candidates for your job and even help create and send messages to them.
5. XOR (pronounce EX-OR) – A chatbot recruitment assistant. It’s Available 24/7, so it immediately responds to every candidate. Guides every candidate through the application process. Asks pre-screen questions. Provides information about the hiring process. Updates candidates if the application status changes. Informs candidates, if a vacancy is closed. Delivers recruiter’s messages to candidate’s messenger. Provides onboarding and training materials. Available on Messenger, Viber, Telegram, Slack and by SMS.
This is my view on chatbots and sourcing/recruiting (besides loving their very cool new age names…)
1. Market adaption – To me, it seems that it is just a matter time, that smart chatbots will infiltrate and conquer the sourcing/recruiting industry like some kind of a (good?) epidemic. The only serious caveat I see is the hard work needed to translate the professional recruiting world into information to “feed” the chatbot. This may, perhaps, slow down adaption.
2. Augmented Self – We will become super recruiters and sourcers, suddenly able to process much more information much more quickly. Scalibility is the norm.
3. Involvement – Hybrid Chatbots will keep us aware and still involved when the person in the other side wants a better answer than the bot can deliver at that moment.
4. Availability – We will be more available to deal with stuff that we want to be on top of – such as closing the deals, strategies etc.
5. Red flags – I cannot overlook the very much needed way to be able to frequently QA all these processes. How can we know that our chatbot is not taking “too much liberty” in learning new stuff and come up with answers that we do not agree with (as a company?) This whole topic is still a major “black box” and we are already finding out problems.
5. Linkedin Chatbot – Shannon Pritchett mentioned chatbots in relation to Linkedin – “It would be a huge win for LinkedIn to experiment with this technology for recruiters and sourcers to help with tedious tasks like scheduling and reminders. A LinkedIn chatbot would surely have its critics, but it could reduce many nuances on the website such as recruiter spam”.
6. One-stop-super-smart-people-aggregators chatbots – I predict, that there will be a time when the super advanced recruiting chatbot will MERGE and REPLACE tools like AmazingHiring/Aevy/Talentbin/Entelo etc – Once a potential candidate converses with our chatbot – All his information will be presented to me on the spot – As I see it today on my supercharged (full of gazillion chrome extensions and tools) Linkedin/Facebook etc public profile.
Remember where you heard it first…
Karen Azulai is a Global Sourcing Evangelist and HR Tech Enthusiast. Karen, lectures in sourcing conferences and conducts advanced sourcing training in Israel. Feel free to contact me at firstname.lastname@example.org