Six recommendations for launching or expanding your virtual agent in this crisis

As we work together to respond to our current crisis, artificial intelligence (AI) is a force multiplier, helping us effectively and efficiently navigate this overwhelming storm.

Artificial intelligence (AI) virtual agents are not new news. Today they are being reimagined to help citizens and workers access trusted data sources and to help get the answers they need. How many infectious cases are in my area? How do I get tested? What should I do if someone in my household is sick? When should I go to the hospital?

Both public sector agencies and private sector organizations are building on these capabilities to rapidly train AI on policy. What is my organization doing to respond to this crisis? What happens if I need to take an unscheduled day off? How much sick leave can I take? Organizational policy is rapidly evolving, each week is a new frontier as we learn, together, how to navigate an unprecedented global pandemic. As organizations amend their business policy, the AI can be quickly updated to reflect these new changes, becoming the single true source of data for an organization and its entire workforce.

AI virtual agents are simple yet powerful solutions. In early deployments around the world, we are seeing a significant decline in the volume inbound inquires to call centers, which means shorter waiting times for those critical cases. When workers get answers from AI, this means more time connecting with managers and colleagues on mission-critical initiatives. By analyzing the AI interactions, leaders have deeper insight into their workforce and use the data to help prioritize their crisis response strategy.

Whether you are launching a new virtual agent or expanding the capabilities an existing one, here are six recommendations:

1. Keep your employee personas at the heart of your design. Understand the needs of your employees and prioritize those when building AI capabilities.

2. Balance security requirements with access and leverage the cloud. Take a hard look at the need to access confidential information and balance that with the need for the AI to be available anytime, anywhere, on any device.

3. Start small and iterate quickly. Consider a prototype with limited functionality, deployed to a smaller group of users. Scale quickly to a broader set of users and expand the AI’s capabilities through frequent updates.

4. Consider integrating publicly available data from trusted sources. Public health organizations, government agencies, and educational institutions are making large amounts of information available. Train your AI to use these data sources or provide links to additional sources of information.

5. Make one of your senior leaders the solution owner. Your chatbot team needs near-real-time insight into the changing organizational policies, in order to adapt quickly. Select someone who is a focal part of the leadership team as solution owner, helping minimize the real-time between policy decisions and training the virtual agent.

6. Use the AI data. Take your cues from your workforce. The questions asked are the topics of most interest or concern. Adjust your crisis response strategy accordingly.