Five keys to deploying automation in government
Connecting state and local government leaders
Deploying AI and automation requires planning, good communication and thoughtful evaluation of policy and existing systems.
Just a few years ago, only a handful of government agencies (mostly military and intelligence organizations) were truly using artificial intelligence and automation. Since then, while not great in number, many solid proofs of concept have emerged with demonstrated results that display AI and automation in all of its forms, including robotic process automation, machine learning, natural language processing and other cognitive tools.
Examples include automating repetitive back-office functions in HR, finance and procurement to harnessing AI to gain deeper insights into data, to deploying chatbots so agencies can more accurately and rapidly respond to citizens’ inquiries.
Agency leaders are to be congratulated for this progress, but we have noticed some trends that give us pause. For example, while agencies are clearly embracing AI tools, most are applying them to individual use cases without a strategy to scale them across the agency where they would have a greater impact on productivity, costs and improved workforce morale.
Now that we’ve helped more than a dozen government organizations strategize and implement AI and automation, we’d like to share a few keys to success.
1. Establish an agencywide strategy. For a major digital transformation strategy to work, all functions must be involved -- not an individual business unit or office. Clear objectives, documented goals and accountability are required. Choose a use case with the highest potential return on investment, and then agency leaders should actively support the project and promote it to departmental and administrative officials.
2. Communicate a plan to the workforce. While some employees remain fearful of automation, we’ve found agencies aren’t planning layoffs as a result of adopting AI, but are leveraging these capabilities to reduce backlogs and enable employees to address more value-added tasks. Regardless, agencies must address employees' fear upfront and clearly communicate how workers will be affected. Involving staff in brainstorming and identifying use cases is a good idea. At some organizations, we’ve noticed improved morale when workers realize these tools can take mundane tasks off their desks.
3. Address governance and policy changes. Introducing new technology can create risks. For instance, does the new technology have the required authority to operate? Does the agency’s IT system or website permit the use of bots? Some do not. The CIO and chief information security officer should advise on these critical questions of policy.
4. Evaluate business processes and data. Agencies generate mountains of data, but unfortunately much of it is not well formatted, structured or machine readable. This can be an obstacle in adopting AI and automation, but it is also an opportunity to address data quality problems, including inefficient processes. Documenting a process prior to automation can point to the need for it to be redesigned or eliminated -- there’s no sense in automating an inefficient process.
5. Support CIO innovators. Interest in AI and automation is largely coming from program offices and functions such as finance and HR -- groups that may view the CIO as an obstacle, a tech gatekeeper. However, our experience is that most CIOs are looking to integrate technology to help the agency function better. Back-, middle- and front-office groups should look to partner with the CIOs to help them through each stage of the AI/automation journey.
Deploying AI and automation requires planning, good communication and thoughtful evaluation of policy and existing systems. And while risk is inherent in any major transformation, the potential ROI -- in improved citizen satisfaction, efficiency and cost savings -- far outweighs the risk.
NEXT STORY: Air Force, MIT launch AI accelerator