To Use, or Not to Use, a Chatbot?
Although chatbots can make recruitment more efficient and productive, it's important to have a realistic understanding of their capabilities and limitations.
By Steve Hamrick
It's starting to seem like everywhere you look, people are integrating chatbots into their lives -- ordering a taco? Meet tacobot. Booking a vacation? Work with automated SnapTravel. These consumer use cases exemplify the potential of chatbot technology and how it can streamline processes -- showing promise for workers, especially HR and talent acquisition employees, to leverage the technology to increase productivity and reduce administrative burdens.
HR and recruiting teams are often slowed down with things like administrative work and answering routine questions from employees and job candidates. While these tasks are certainly part of the job and sometimes don't take much time, it can add up to a significant portion of the 40-hour work week.
Most recently, HR and recruiting teams have turned to the innovation of artificial intelligence and machine learning to streamline certain processes and ultimately get more time back in their day. Sounds great, but is it too good to be true?
What to Trust a Chatbot With
There are some cocerns over giving a chatbot responsibilities, and those concerns are valid as we continue exploring the limits of the technology. In their current state, while chatbots may not be ready to take on a recruiter's full workload, they are certainly equipped to help kick off tasks and provide basic information.
Automate Outreach For recruiters who are flooded with enormous amounts of resumes, chatbots can provide relief for repetitive tasks by, for example, automating outreach to candidates to provide next steps in the interview process. For internal HR needs, in some organizations it might make sense to have a chatbot on the front line of support to provide employees with an immediate response. HR can work to develop a search index that matches frequent employee questions to a known set of articles to help them find the answer themselves. However, it's good to have a human close by in case there are additional questions or clarifications needed.
Offer Catered Alerts and Notifications Chatbots can easily follow parameters and automate alerts based on information, which can be helpful when searching for candidates with specific requirements. For example, a recruiter could set an alert to be notified when a resume appears with 15 years of IT experience because that is required for a specific role. They would be able to change the parameters of notifications and alerts from job to job based on the role, responsibilities and department.
Data Entry and Search Retrieval Similar to the alerts and notifications, chatbot technology can manage processes that are easily repeatable without human intervention or help. These types of processes typically fall into administrative tasks that HR takes on, such as compiling new-hire documents, or entering data about new employees. Having a chatbot complete these tasks allows recruiters to spend more time recruiting candidates and working on creative, strategic projects that can't be done by a bot. Bots can also ensure that everyone uses transactional systems where these processes and data are stored, driving ROI.
Chatbot Challenge: Communicating like Humans
The main challenge preventing chatbots from taking on a human workload and managing business processes is communication. Chatbots can point candidates and employees in the right direction, but they typically don't have the right context to fully respond to requests. In this case, trusting a chatbot to supply information can leave a candidate or employee with missing information and potentially more confused than ever. There are several gaps that need to be filled before chatbots can become a trusted business tool.
Translating Human Intent Chatbots in their current state cannot understand the nuances of intent or emotion from hearing human speech. As such, they're not able to grasp the full meaning of many phrases or commands in a human capacity. This limitation comes from under-developed natural language processing technology. NLP has a few challenges when it comes to translating complex human requests into commands because it doesn't account for the context clues of intonation or emotionality in a typical conversation.
The Learning Curve AI is becoming more advanced and will ultimately help chatbots better understand human intent and predict human needs. The AI-enhanced chatbot of the future will be able to learn from prior experiences. For example, if someone books a conference room for a 30-minute interview, an AI-enhanced chatbot might be able to recall that the last time this person booked the same conference room, the interview went over by 15 minutes. The bot could learn from this occurrence (the same way a human would) and automatically extend the reservation to account for the potential of the meeting running long. Unfortunately, today's chatbots are only smart enough to do what we tell them to do, so are not nearly as helpful as a human assistant who can learn and adapt to your needs and preferences.
Generate Limited Content Today's chatbots cannot provide dynamic content because they are not securely integrated with pre-existing business applications and, therefore, cannot draw from them to provide a rich, complete answer. As a result, chatbots can only answer a narrow, basic set of questions -- think of teaching a dog how to fetch your slippers. The animal can repeat the task if your slippers are in the same spot every time you ask for them. But you might run into trouble when you ask for something more complicated, such as your flip flops that are buried at the bottom of your closet. These basic bots are similar to our furry friend -- they can reliably do simple tasks, but anything more sophisticated will require enhanced development.
Chatbot Power Struggles
In large organizations, different departments often use different chatbots. At this point not all chatbots are interoperable, so they cannot communicate with one another. This is likely due to a lack of industry standards or guidelines that would ensure chatbots can understand each other. If the HR department isn't using the same technology as the marketing department, there will be an adjustment period where everyone needs to change to a different technology -- not to mention the power struggle of determining which department has the best chatbot.
As with all technological innovations, it takes some time to work out the kinks and identify the best use cases and processes for the technology. We are still in that incubation period of figuring out how chatbots fit in the HR world and the best way to use them to help HR and talent acquisition professionals be more productive and creative in their work.
Steve Hamrick is vice president of product management at SAP Jam.