Industry analysts say employers are increasingly exploring artificial intelligence (AI) as a means of solving their human resources challenges. Are those who haven’t yet embraced AI about to be left behind?
Not really. A host of challenges — from a general lack of consensus about what “AI” even means to systems integrations issues — have slowed industry adoption of true AI solutions. But the landscape is fast evolving, so now is the time for employers to educate themselves on what AI means and how it can benefit the HR function.
What does AI really mean?
Most people, including HR practitioners, have a hard time defining AI. The term refers to a wide range of solutions that incorporate human intelligence into machines or computer systems, usually for the purpose of performing tasks like problem-solving, decision-making and predicting the future with greater accuracy or efficiency.
Systems can be automated without the use of AI. Consider automated telephony systems that require callers to indicate selections by pressing a button on their keypad. The rules tell the system what to do if the user presses “1,” “2,” and so on. There is no room for interpretation or adaptation; callers will always be presented with the same array of options, and pressing “1” will always produce the same result.
AI goes much further than automation, imbuing systems with human capabilities such as learning, interpreting and predicting. Two of the best-known AI disciplines are machine learning and natural language processing. Machine learning is the process by which systems learn from experience. Using computer algorithms, systems parse data, draw conclusions from it, and then make determinations or predictions about something in the world. Natural language processing is the process by which systems learn to understand human language and present data back to the user in human language, so that neither information inputs nor outputs have to be formatted as data.
How are HR functions deploying AI today?
Gone are the days when HR departments managed cabinets full of employee files and documented procedures in three-ring binders. The digitization of HR data made it possible for humans to retrieve information and perform HR tasks with much greater speed and efficiency. Now, forward-thinking organizations are advancing the industry even further by using computers programs to perform those same tasks. But which of these solutions count as AI?
Consider the now-widespread use of chatbots to answer common employee questions. When an employee chats a question — say, about their health plan copay — the chatbot doesn’t understand the full text of the question asked. Rather, it picks up on the keyword ‘copay’ and offers questions it can answer based on information saved in a database. Sometimes that information even includes data fields that are unique to the individual, like their name and the benefit plans they are enrolled in.
“It sounds like you have a question about your copay,” a chatbot might say. “I can tell you what a copay means. I can tell you what the copay for your health plan is. Did one of these answer your question?” This kind of chatbot can greatly reduce the resources a company must devote to answering employee questions, but it’s an example of process automation — not true AI.
More recently, vendors have begun developing smart chatbots, or ‘smartbots,’ with the capacity for machine learning and natural language processing. These AI-enhanced smartbots can interpret and answer employee questions in more sophisticated ways. For instance, a smartbot might remember previous interactions with an employee and use that contextual information to answer follow-up questions more efficiently. Or, using natural language processing, a smartbot might be able to accurately interpret employee questions that contain typos or that don’t contain an exact match for any keywords in the bot’s database.
Process automation is a particularly hot topic in the recruitment funnel. There, deployments range from basic chatbots that are little more than after-hours answering machines to sophisticated AI assistants that automate the sourcing, screening and scheduling of candidate interviews. Employers have even turned to AI as a means of eliminating human bias in the recruitment process — for instance, using predictive analytics instead of human intuition to determine an applicant’s likelihood of succeeding in a given role.
Any HR task that is time-consuming, prone to manual errors or subject to human bias is ripe for automation, and any time that process automation is augmented with technologies like machine learning and natural language processing, it’s an example of AI at work.
While it’s true that employers have begun to employ AI in the workplace, the “AI revolution” is still in its infancy. Several challenges must be solved before AI goes fully mainstream:
- Vendors are moving quickly
AI providers have been moving at light speed to deliver AI solutions to market while keeping up with advances in AI technology. Since very few AI vendors actually specialize in the HR space, employers may find it especially challenging to identify an AI vendor that suits their HR needs. Before shelving their dreams of a more modern and efficient HR function, employers should seek the assistance of a third party that’s versed in working with both plan sponsors and AI vendors. An experienced vendor search consultant can help employers ask potential providers the right questions and poke holes in proposals that seem too good to be true.
- Systems don’t talk to one another
Chatbots and other AI-driven process improvements are frequently deployed in isolation. For instance, a company might work with its health and welfare plan administrator to implement a smartbot that helps with benefits enrollment, then engage a completely different vendor to develop an AI assistant that screens job applicants. If the company later decides it wants to provide a more comprehensive, centrally managed experience, those independent solutions will need to share data with one another via an API connection or custom integration. Some vendors are good at this kind of collaboration, and some only want to play in their own sandbox.
- Projects come with hefty price tags
It should come as no surprise that, as the “shiny objects” of HR technology, AI projects come with sizable price tags. The more advanced the system you want to put in place, the more your cash outlay will be — and the greater the potential value for your organization. The best way to determine if an AI project is worth its price is to quantify in advance how much the manual process you intend to replace is costing you on an annual basis.
AI has the potential to improve HR functions by elevating the level of service offered to employees while reducing departmental spending. But before you jump in head-first, thoroughly vet prospective AI technologies to ensure you receive the intended return on investment.
This article originally appeared in Employee Benefit News.