AI in HR

Original article was published on Artificial Intelligence on Medium

Why AI used in HR

Today, AI’s capabilities are being used to augment business operations and consumer solutions. We have identified five primary reasons for implementing AI in HR:

➢ To solve pressing business challenges

AI enables HR organizations to deliver new insights and services at scale without ballooning headcount or cost. Persistent challenges, like having the people resources to deliver on the business strategy and allocating financial resources accordingly, can be addressed through the thoughtful application of AI solutions.

➢ To attract and develop new skills

The business world is constantly being disrupted. In order to cope with this disruption, businesses need to respond faster to opportunities, and to work in an agile way to stay ahead of competitors. This means finding an effective way to compete for the skills required to innovate in this new operating environment. AI applications enable HR departments to acquire and develop employee skills in lockstep with shifting market demand.

➢ To improve the employee experience

People have started to expect something different when they come to work; they want a personalized experience, not a standard one. They want things to be tailored and offered to them in a way that works

for them from the start to the end of a process. Today, people can also look inside a business from the outside with sites like Glassdoor, which puts a huge premium on the employee experience.

➢ To provide strong decision support

The speed of change and rate at which information is being generated means that business decisions today are best made analytically. Because the amount of information that needs to be considered is vast, AI can be used to make sense of it and deliver recommendations. As a result, the information managers and employees require is there just when they need it. AI also provides the opportunity for employee voices to be heard and acted upon in real time.

➢ To use HR budgets as efficiently as possible

AI can enable HR to become more efficient with its funding. HR spend can shift to higher value and more complex problem solving, without reducing levels of service for workers who have more routine HR queries. HR savings made in this way can be reinvested in further AI deployment, increasing HR’s ability to solve business challenges, continuously develop strategic skills, create positive work experiences, and provide outstanding decision support for employees.

Its benefits

AI in HR has the potential for significant organizational benefits. This was achieved by clear articulation of expected outcomes tied to business problems, careful selection of appropriate metrics to measure the expected outcomes, and regular tracking of results to enable iterative improvements. Return on investment Tracking return on investment (ROI) in HR needs to be a business imperative. HR practitioners should have a direct line of sight from AI applications to the outcomes AI will produce and the associated ROI that occurs in the business. Establishing the expected connection between AI and its return should occur before the AI application is implemented.

Let’s consider an example of an organization struggling to close sales because of a lack of technical knowledge among its sales people. Here, AI might be proposed as a method to personalize learning for the individuals’ needs, delivered in a consumable format. Personalized learning is the outcome that the AI produces. The HR metric that the learning connects to is the skill breadth and depth of sales people in the business.

Steps for successful adoption of AI in HR

IBM has learned many lessons in its deployment of AI, and interviewees recounted many that could be helpful to practitioners. Some of the major learnings are shared here.

➢ Don’t wait until you have the perfect solution

It is better to release a minimum viable product and position it as such with your employees. When IBM released the performance management chatbot, people knew it was ready but not perfect, and were encouraged to ask questions that would challenge the bot. This allowed IBM to deploy the AI solution quickly, while at the same time enabling the chatbot to improve.

➢ Empower people with AI

Designers of AI systems have a stewardship responsibility to ensure that AI empowers workers. Designers should keep in mind that people feel most empowered when their decision autonomy is augmented rather than replaced. AI has been viewed most favorably when managers have the option to override the AI recommendations when not seen as optimal. IBM encourages managers to be comfortable overriding the AI recommendations as appropriate.

➢ Ensure transparency

For managers to feel comfortable working with an AI recommendation, it is important that there is clarity and transparency about why AI recommendations are made. This should include making clear to managers and employees what the recommendation aimed to achieve, what data were used to make the recommendation, which variables influenced the recommendation most, as well as identifying all the variables on which the recommendations were based, and the expected accuracy of the recommendation.

➢ Consider language and culture

AI requires context, reflected in data from different regions where AI will be deployed, to learn and make appropriate recommendations. An AI solution designed in one region of a multinational organization’s operations may need to be entirely retrained before deployment in another region — even if the same language is spoken in both areas. This relates to both the development and training of the model, as well as the experiences of the end users of the AI solution.

➢ Design each app with other apps in mind

As you develop your AI solutions, it is a good idea to have a holistic view of the end goal in mind. This will avoid, for example, having a proliferation of unconnected chatbots all addressing related questions but not drawing on common infrastructure and data. It is increasingly important to do this because apps often do not sit in the traditional siloed HR subdomains. For instance, the compensation example discussed earlier requires information from compensation and learning, among other HR functions.

To learn about Artificial Intelligence and its impact visit: Allicient technologies