AI and Cloud Could Make Getting Unemployment Insurance Easier in the Next Recession 

According to a 2021 survey of state-level CIOs, the pandemic has highlighted the value of technology and how it is critical to the delivery of government services and the productivity of state employees. The survey also revealed the biggest driver for state governments is the need to digitize: “better online experience for citizens.”

Perhaps nowhere is this truer than with unemployment agencies. Clearly a black swan event, many states experienced massive problems processing unemployment claims during the pandemic. The reasons for this are varied but included an inability to perform the needed work from remote locations, antiquated legacy systems unable to process an unprecedented volume of claims, overloaded phone systems, an inability to detect high numbers of fraudulent applications, and a lack of funding for staff and system upgrades.

Since the onset of the pandemic in 2020, states have been attempting to both catch-up and to prepare for the future. For some at least, this remains challenging. An op-ed in a Nevada paper noted that the state has yet to complete processing claims from nearly a year ago and stated: “If they operated in the private sector, those running Nevada’s unemployment system would have been unemployed long ago.” Though in fairness, if this was the private sector it is likely that more resources would have been provided. Nevada is hardly alone in this experience.

The need is urgent as another round of unemployment claims could come with concerns of a recession mounting.

States are attempting to undertake the needed upgrades, but talent and funding are often obstacles. This is a significant challenge as state unemployment agencies need to leap-frog from legacy systems to new models enabled by cloud and powered by disruptive technologies including artificial intelligence (AI). For example, Oregon is in the midst of replacing systems and software dating back to the 1990s.

Fortunately, modern systems are available to help unemployment and other public sector agencies by leveraging the cloud, AI, blockchain, and data management technologies. AI-driven systems are particularly crucial for government agencies to best meet their constituents’ needs, automating answers to user questions in real-time, distributing checks faster and identifying complex claims for humans to review. Advances with AI-enabled software are now able to transform the unemployment insurance experience – from making a claim to uploading weekly proof of active employment search – regardless of the device an applicant is connecting from. These next generation systems can also help the unemployed find new work. New York has recently implemented a new AI platform to create rapid re-employment, as well as enable long-term career planning.

One tactical option chosen by states during the pandemic to quickly update legacy systems was to modify existing code, move the user interface to the cloud or develop a new interface using low-code tools, add people to operate the phones, and increase support for dated applications. While this may have helped some and was perhaps the only short-term option available, the approach was tantamount to a bandage and has not worked very well. Certainly, this approach is not a viable long-term solution to a massive problem.

The only real solution is to replace existing systems with modern technology. Some states may opt to build this themselves, but a lack of available talent and time are not on their side. They are better off obtaining commercial applications running in the public cloud and designed by those with expertise in state unemployment solutions design and implementation. By running native in the cloud, these systems will be nearly future-proof and hugely scalable to manage through a future extreme event. Expanded use of AI and automation within these new unemployment systems will better enable states to manage staff constraints, improve efficiency, and stakeholder service and support.

Pennsylvania and other examples of this approach are appearing such as in Hawaii where AI, administrative data, and cloud computing are being harnessed to generate custom recommendations and job matches. Indiana has taken a similar path. Prior to helping people find new jobs, updated unemployment systems will proactively inform and update assistance applicants to ensure they always know the current state of their claim — where it is in the process and what the timeline is for decision and payments. A modern system also provides a high degree of self-service so that an applicant can not only submit their claims online. They could also see the status of their claim at any time, as well as engage with chat or other resources to obtain answers to their questions to eliminate confusion and concern. At the same time, these systems will use AI to spot fraudulent applications to both facilitate processing of legitimate claims and reduce waste.

Harnessing the latest technologies to provide modern, scalable, responsive, and future-proof unemployment systems is something that is possible right now. If society wants to avoid the unemployment systems debacle of the pandemic, it would be well-served to provide states with the resources to replace legacy hardware and software with modern AI-powered, cloud-native applications. The good news is that Congress has provided the Labor Department with $2 billion in pandemic relief funds through the American Rescue Plan Act of 2021 to help states upgrade their unemployment programs. Now the question is whether there will be a recession adding stress to current systems before those new programs are put into place.

About the Author

Bhanu Prasad Narayana – AVP, Head, US Public Sector, Infosys Public Services – has 25 years of IT experience working in many global locations and industry sectors. In his role leading Infosys Public Sector Market for North America, Narayana advises CIOs and senior leaders of the agencies in federal, state and local markets on emerging industry trends and innovative solutions to solve problems for citizens and constituents.