Invited to an ‘AI for Recruiting’ Meeting? Here’s What You Need to Know

While 42 percent of HR leaders believe artificial intelligence and machine learning are “among the biggest transformation challenges they will face in the next five years” (according to KPMG’s The Future of HR 2019 report), a troubling 50 percent admitted to being “not at all prepared” to respond strategically to artificial intelligence and machine learning. This means a significant number of HR professionals are unsure how to proceed when it comes to researching and planning deployment; how to set goals, priorities, and evaluate AI recruiting solutions for their organizations.

If that describes you, don’t worry: you’re not alone.

Given that AI is new and complex, but also has great promise, you need to prepare for that inevitable meeting.  

Here are suggestions to help you get ready, so you can ask intelligent questions, and get the most out of the meeting:

Familiarize yourself with AI-related terms

In advance of the meeting, be sure to acquaint yourself with the terminology of AI technologies and applications (such as chatbots, classification, and matching). For a helpful preliminary introduction, see the sidebar below and also Wikipedia’s entry on Artificial Intelligence. 

See below in this post for a glossary.

Identify the functions you hope to address and the problems you are trying to solve

AI isn’t a panacea for recruiting challenges, nor can any technology replace many of the functions provided by talent-acquisition professionals. Automation can help you manage the recruitment process, source, and screen candidates easier, and provide candidates with a better experience. But matching, engaging, selecting, and closing applicants still requires a human touch and the expertise of external and internal recruiters. Identify your pain points and make sure your organization is clear about which recruiting processes are being targeted for improvement.

In other words, make sure everyone is crystal clear on what problem you are trying to solve with AI.

Understand the complexity (in terms of human decision making) of the tasks you wish to automate

When choosing which problems to tackle with AI, be sure to assess task complexity. And, if your organization is relatively new to implementing AI, start with simple versus complex activities. Typically, activities best suited for AI are the higher volume/lower complexity tasks at the beginning of the recruiting process. Tasks later in the hiring cycle are much more complex and not well suited for AI, at least as an initial application.

For example, AI is helpful for recruitment marketing, as well as for arduous and time-consuming activities such as scheduling candidate interviews. Similarly, chatbots can efficiently capture and analyze applicants’ answers to screening questions. However, AI cannot handle the more complicated human tasks of selecting and closing (i.e. debriefing, aligning and summarizing team feedback; making and negotiating job offers).

Set clear objectives and expectations; apply best practices

Just because the technology is new, doesn’t mean you can’t apply standard best practices you already know about for adopting new tools or technologies. For example, to effectively plan for an implementation of AI for recruiting, consider adopting the SMART approach to setting objectives and expectations.

SMART is an acronym that spells out that your goals should be Specific, Measurable, Attainable, Relevant, and Timely. Thinking through exactly what you can realistically achieve, and planning how you’ll execute your initiatives, will make it easier to document and demonstrate your success. Set specific and clear goals for your first applications. For example, many companies choose to automate one highly labor intensive task with the goal of a breakeven ROI within 12 months.

Additional best practices include conducting due diligence and checking references, ensuring proper integration points with other systems and departments, and managing data effectively, including security and privacy, across your organization.

Ensure you have the right algorithms to meet your objective

Don’t let the word “algorithm” intimidate you. At its core, an algorithm is simply a step by step process to accomplish a task.

As you evaluate systems from solutions providers, or your own team, your team and vendor should have the right algorithms for the technologies needed to address the task you want to handle. For collecting and analyzing screening questions, you will need sophisticated chatbots. For categorization and matching, you’ll need powerful machine learning algorithms.

Those with experience with the applicable technologies probably will have the needed algorithm experience as well. Ask your vendor and/or team about relevant use cases and examples.

AI applications (things AI can do) include:

Learn (Ex. Netflix recommendations)

Classify (Ex. Job type structure)

Article Continues Below

Identify (Ex. Find the cat in the photo)

Extract Meaning and Context (Ex. Smart replies)

Translate (Ex. BabelFish)

Converse (Ex. Chatbot)

Predict (Ex. Demand for pricing — hotels and airlines)

Decide (Ex. Siri, Alexa)

Confirm the appropriate data exists to be effective and to test the results

The success of AI and Machine Learning systems are directly related to both the quality and quantity of your data. As “Moneyball” (the book, and later Brad Pitt film), clearly showed, analyzing data of individuals’ performance can be helpful in predicting future success of new players.

The same  goes for new hires, but in baseball, there are great sources of high-quality data. In other fields, such as recruiting, that isn’t always the case. For recruiting applications to deliver value, your company needs accurate data that is relevant and necessary to your hiring process and success. Machine learning becomes more and more powerful the more it crunches data, and it’s crucial that the data is of high quality. “Bad” data will spoil algorithms from the outset and lead to inaccurate insights and poor, and possibly biased, hiring decisions. Ask about the data quality that your company has or how any vendor being considered ensures high-quality relevant data.

Conclusion

While most recruiters aren’t expert in artificial intelligence and analyzing data to improve hiring results, the technology is critical for the future of recruiting and more and more employers are embracing AI. Educate yourself on the basics so you can participate and help determine how and where it is adopted in your organization. Becoming familiar with AI terminology and some of the implementation basics is a worthwhile investment in your future. And, it will prepare you to ask critical questions, look smart and be prepared for that dreaded but inevitable AI meeting.

Glossary

Machine Learning — method of predicting outcomes based on data and without explicit programming

Natural Language Processing — method that enables computers to understand, interpret, and manipulate human language

Speech Recognitiontechnology that enables recognition and understanding of spoken words, by digitizing the sound and matching patterns against the stored patterns

Expert Systems — a program that is designed to emulate and mimic human intelligence, skills, or behavior

Robotics — the design, construction, operation, and use of robots, as well as computer systems for their control, sensory feedback, and information processing

Machine Vision — technology that enables a computing device to inspect, evaluate, and identify still or moving images

Ken Lazarus is CEO of Scout Exchange, a platform for marketplace recruiting.

Prior to joining Scout in 2013, Lazarus served as CEO of Lilliputian Systems, where he led the development of Lilliputian’s portable power product platform, secured international aircraft regulatory approval, raised significant capital from blue-chip investors, and signed important strategic manufacturing and distribution agreements. Previously, Lazarus served as CEO of ACX, a maker of vibration and motion control equipment that ranked #79 on the Inc. 500 before being acquired by Cymer, Inc. (NASD: CYMI).  Lazarus also cofounded DataXu and Ekotrope, and has worked with numerous other startups. He holds a Ph.D. from MIT and a B.S. from Duke University. Lazarus, who is a World Economic Forum Technology Pioneer, has served as an MIT Visiting Committee Member and holds over 20 patents.

Topics