Lion.

Silvia Cohen

We are sitting on the cusp of the AI age and are beginning to see the potential of AI-powered Chatbots. Cutting-edge AI technology could eliminate mundane tasks, enable mass personalisation of employee experiences, and allow us to combine vast amounts of data to increase the accuracy of our predictions and strategic decisions. This will change the nature of work, especially what and how work gets done in HR.

You may wonder why the hype surrounding Generative AI tools has exploded over the last few months. Effectively, tools such as Chat GPT and Bing Chat are language models that generate natural language text in response to users’ questions. These tools are trained using enormous data sets and can answer questions quickly and succinctly. In addition, it allows you to ask follow-up questions and can provide you with references.

Several HR use cases have been identified. The obvious is eliminating transactional HR tasks through an HR self-service chatbot. This self-service chatbot could integrate internal policies and practices with external leadership to answer employee questions and coach leaders. Examples of these could include performance and conduct management. The chatbot would share the relevant policy and coach the leaders on structuring the conversation, and it may even script your key messages.

There are numerous advantages of HR self-service chatbots. These advantages would include providing just-in-time support for employees and leaders, consistent and accurate information and advice, and value creation, as you would require fewer resources in this space or reallocate them to more valuable tasks. However, whilst this is helpful, it isn’t what excites me.

"Generative AI lets us look at a variety of textual data or written language, including transcribed verbal data. It learns from data patterns and relationships to make predictions by combining and analyzing different data"

The game changer would be that this technology can elevate our work and position us to drive company performance and create exceptional employee experiences. Generative AI enables us to analyse a vast array of textual data, that is, written language, which could include transcribed verbal data. It combines and analyses the different data, and learns from the patterns and relationships within the data to make predictions. These predictions are likely far more nuanced than our current approach, allowing for greater accuracy.

With this capability in mind, my first thoughts go to talent. AI and machine learning solutions already exist that enable us to match candidates to job requirements. For example, some emerging assessment companies identify 50-60 traits that predict success in a given role and then apply a 15-minute assessment that provides stronger predictors of success than the traditional batteries of psychological tests. The generative AI solutions may take it further and allow us to consider other factors, such as customer feedback and financial performance, and apply this approach to all talent moves.

Beyond talent, generative AI could help to create more personalised employee experiences. For example, we could analyse what skills, traits, and knowledge lead to success in a given role. With this insight and knowledge about the candidate, we could automatically deploy onboarding and development material to target any gap areas they may have.

Another area that could be turbocharged using generative AI is employee listening, i.e., measuring employees’ experiences. It could open up how data is collected, moving beyond static surveys, integrating performance data and generating powerful insights. For instance, a Chabot could ask employees about their experiences in the moments that matter, deep diving where it makes sense, and then integrate other data sources, such as sentiment analysis, performance data, and employee behaviour data, such as turnover. This would allow us to predict likely outcomes for performance and turnover and implement interventions. 

Similarly, you could train the generative AI on the desired culture and share data on the current culture and the company strategy. The model would help to identify the gaps and potential interventions. Of course, operationalising this would require training and testing the model before it is cut loose and released to the enterprise. 

Whilst the potential of generative AI is exciting, it has its risks. Some of the answers provided by models such as Chat GPT and Bing Chat are incorrect. It would be best to weigh the risks and benefits in determining when a decision requires human verification. The role of HR would be to critically evaluate the outputs of models.

The other risk that has received much attention is that these models may propagate systemic biases, given that they are trained using biased data. The creators of these models say that they are working to reduce biases through actions such as diversifying the training data and monitoring their responses. However, given the pervasive nature of these biases, they are challenging to eliminate. We would need to monitor and test for this.

The role of HR will evolve as this technology is sharpened. It is a time for us to be open, curious, and experiment to see how we can amplify our capability in the AI age. It will allow us to play in more strategic spaces whilst leveraging our uniquely human skill set.