There is substantial hype around artificial intelligence (AI), yet so many things about the technology are misunderstood. People are confused about what AI is, what it can accomplish and how to use it most effectively in a business environment.
Deloitte's Bersin 2018 High Impact Talent Acquisition study found that less than half of companies participating in the study have implemented AI, and only 14 percent are seeing some effectiveness. Of the firms surveyed, 35 percent have found that AI has been ineffective. Here's how to ensure your organization gets the most out of AI.
What Is AI?
The foundation for AI has been around for a while, but what has changed and made AI possible is increased computing speed, lower cost of computing and the vast abundance of data.
There are now essentially three types of AI helping the human resource function:
Classification. This uses regression or neural networks to study data and find ways to classify it. While it sounds simple, this is the AI that powers software to find the best candidates and match them to the right job based on the information they submit online. This type of AI can discern thousands of variables and figure out what makes someone the right fit.
Natural language processing. As mentioned in my column last month, this is the AI that knows how to communicate via chat, e-mail, voice, etc. This has tremendous potential within HR. This AI can be used as screening bots or to answer service-desk questions or schedule appointments.
Image analysis. This type of AI can interpret pictures and identify elements inside the pictures. Outside of HR, this particular field has made huge progress. For instance, AI can now be used to identify cancer by simply looking at a retinal scan. HR is beginning to incorporate image analysis, as some video-interview tools are experimenting with the potential applications.
Challenges to Understanding and Implementing AI
Deloitte conducted a survey in the fall of 2017 called the State of Cognitive, and the results were intriguing. The No. 1 reason that firms had trouble getting benefits from AI was they couldn't figure out how to effectively integrate the AI into existing processes.
A leading AI researcher in the technology industry spoke at a recent conference and announced that AI is often like alchemy, and the software engineers often don't understand how the results are occurring. This researcher and a panel of other AI engineers submitted a paper asserting that users need to dramatically increase the focus on such issues as rigor and peer review to develop AI in a measured way.
In HR, there is a need to be able to explain how and why something happens. HR should be able to document why applicants were screened out, how employees sign up for benefits and why they got (or did not get) the raises they wanted. Given how AI works, HR is going to need to get a little more comfortable with the idea that AI could be making a decision but is not necessarily able to explain why.
Leading AI Practices
HR professionals should think about learning to utilize AI. There are too many potential benefits to not tackle this challenge. We have found from organization maturity measurements that those organizations that are using AI are twice as effective as those that are not. Here are some things to consider when bringing AI into your organization:
- Pick a function or process that is ripe for AI improvement. Look for tasks that are highly manual, require high interaction and utilize a large amount of data. This could be running an administrative process like benefits enrollments, screening resumes, engagement forecasting such as predicting attrition, or suggesting learning curriculums.
- Determine how you are going to build your AI. You may have in-house data science teams that are likely capable of building machine learning models, or you can use vendors who have built point solutions with AI embedded, like interview-scheduling chatbots. If you are using a vendor, be aware that there are literally thousands. At Bersin, we currently are tracking over 2,200 solution providers, and it seems like we are learning about new firms every day. Consider using a third party to help narrow down the landscape for you, and make sure your request for proposals process is robust. It's imperative to compare solutions.
- Capture the win. You should plan for how to measure success, such as increased quality, increased process efficiency or reduced cycle time. This is critical, as you likely will have to spend a considerable sum on your first AI project, and as a result you will need to use the outcomes to help garner support for future projects.
The advances being made in AI are truly amazing, especially if you enjoy math, statistics and computer engineering. While it may be true that HR doesn't need to fully understand how AI works in order to be successful, those who take the time to study, learn and implement the technology are able to reap greater benefits: 71 percent of higher-maturity companies are using findings from their robotics, cognitive or AI methods.
Franz Gilbert is vice president, solution provider programs, at Bersin™, Deloitte Consulting LLP.
As used in this document, "Deloitte" means Deloitte Consulting LLP, a subsidiary of Deloitte LLP. Please see www.deloitte.com/us/about for a detailed description of our legal structure. Certain services may not be available to attest clients under the rules and regulations of public accounting. Copyright © 2018 Deloitte Development LLC. All rights reserved.