Artificial intelligence is nothing new to governments as future-minded institutions have been piloting AI technology for years. However, what is currently being changed is that AI is no longer an option. Recent Management and Budget Directives and Executive Orders require agencies to adopt AI as a core component of their service delivery strategy. However, AI alone is not enough. That must be responsible for ai.
Responsible AI means adopting and applying technology in ways that promote transparency, accountability and security. This is an approach that ensures coordination with key ethical AI principles. Federal agencies need to clarify how data is collected and use it to train AI models to enhance the outcomes of efficient and effective missions. Agency should also maintain the ability to intervene when necessary and actively prevent problems from occurring with continuous testing, human review and surveillance. These measures can ensure that the system protects confidential information against misuse or exploitation.
Additionally, the Government's Improved Service Delivery Act, the Integrated Digital Experience Act, OMB Directives M-25-21 and M-25-22, and Executive Order 14179 are combined to further elevate the public sector's concept of responsible AI. Institutions must develop and maintain minimal basic knowledge of how to use AI as part of their formal duties. They also need to leverage commercial, American-made, cost-effective AI solutions designed to accelerate federal efficiency, stunt innovation, improve self-service, and remove unnecessary bureaucratic requirements for improving self-service. It was implemented in a way that improved the civic experience and overall service delivery.
It may seem challenging when the public sector is taking less, challenging them to embrace change, raise the stick of experience that can be delivered to the public, and navigate ways to effectively use AI to support the agency's mission to provide better outcomes to the public. Fortunately, agencies such as the Bureau of Veterans Affairs, the IRS, Federal Student Aid, and the Social Security Administration have already begun implementing responsible AI and are offering lessons others should follow.
A blueprint for implementing AI responsibly to the government
This is an important step that institutions can follow to adopt AI in a fast and effective way while promoting trust with citizens.
Start with a clear AI use case in mind – what goals it will achieve, the problems it will solve, the needs of citizens that can be better met, or how to optimize employee workloads.
Your agency should take your mission and start with the last thing in mind and ask yourself. How can AI make everyday processes better for citizens and employees? Implementing AI just for innovation is unsustainable and often leads to wasted resources. Focused use cases ensure that AI investments are related to measurable outcomes, such as improving decision-making and faster, more efficient service delivery.
This was the case in VA that adopted AI in 2017. They started with goals in mind. It means wanting to better understand the veteran experience. By implementing AI, they were able to instantly analyze the empirical data points of millions of veterans, automatically identify veterans at risk of self-harm, suicide, and homelessness, and intervene with appropriate resources faster. Agents also used these features to prioritize resource allocation and improve service delivery.
Identify user experience friction and reveal the improvements that can be made using AI
Frontline employees often have first-hand knowledge of where the government system is lacking. Their insights can contribute to the specific identification that AI can have the greatest impact by removing bureaucratic bottlenecks and enhancing the end-user experience.
For example, one federal agency is evaluating ways AI streamlines workflows by removing unnecessary steps, reducing administrative burdens, and allowing staff to focus on high-priority work. These improvements not only increase efficiency, but also support employee well-being and retention.
The best balance between AI and human interactions and the high-class citizen experience
AI is a powerful tool for delivering the right information to the right people at the right time and place, manually aggregating, analyzing, triaging and delivering.
For example, a state DMV is considering using AI to detect when users are experiencing website problems and leveraging these signals to proactively trigger outbound contact center outreach via smart scheduling tools.
Another federal agency uses AI to detect citizen feedback related to fraud and uses automation to notify the right team.
Both examples demonstrate the advantages of using technology to automate specific processes and leaning towards human intelligence to handle complex cases when simple self-service is not sufficient.
Choose the right technology partner
With AI applied directly to government official workflows and processing the most sensitive information of citizens, agencies cannot afford to make this mistake. Leaders need to choose American-made AI technology from vendors who are committed to responsible AI. They need to have safeguards in place to ensure that sensitive data cannot be utilized by foreign interests or bad actors. Process data in ways that protect individually identifiable information, train models to ensure correct outcomes, intervene when necessary, and provide protection to prevent AI from delivering unintended outcomes.
Develop a process to gather ongoing feedback to understand whether AI is delivering the desired results
As agencies adopt this new phase of change and modernization, the most advanced is using citizen and employee feedback to assess the impact of AI to assess whether AI-powered processes are more effective and efficient.
For example, when SSA used AI to analyze website and contact center insights, it was able to uncover and address one of the root causes of calls. Citizens had a hard time getting back to open applications once they started the process online. Thanks to AI analytics, agents were able to add a “Return to saved applications” button, and through continuous feedback and monitoring, SSA was able to see that this fix helps achieve the desired outcome of reducing complaints and calls.
Lessons from the private sector
There is a great opportunity to learn from brands that use AI for years to improve quality of service and increase efficiency.
First, the customers who interact with the brand and citizens who use government services are the same people who share the same expectations. Second, the administration encourages the federal government to utilize the best practices and lessons learned about how the private sector approaches business.
This is what the IRS did. During the 2022 tax season, agents responded only to approximately 10-15% of calls. The following year, the agency deployed a smart callback option, ensuring that up to 95% of callers do not have to wait on hold for long periods of time. This change is consistent with what consumers want and expect. A 2024 Medalia survey asking consumers what improves during customer service interactions found that short waiting times were at the top of the list. The study also found that customers experiencing short waiting times were more likely to report positive satisfaction ratings. Findings from another Medalia survey published in 2023 revealed that most consumers (58%) feel they are waiting properly for five minutes, but if they are given the opportunity to receive a callback (66%), they would prefer this option.
The urgency to adopt AI in government is for the same reasons it is accepted by the private sector. To enable transformation, modernization, create more efficiency and become more effective, teams can increase less capacity and accelerate effective service delivery.
AI is not a silver bullet, but when implemented responsibly, it helps institutions optimize their operations, reduce inefficiencies, and strengthen public trust. However, if done inadequately, the agency risks experiencing the opposite effect. As federal requirements evolve and citizen expectations rise, successful agencies will become institutions that take a thoughtful, mission-driven approach to responsible AI.
Lee Becker is Medalia's Industry Executive Advisor and Senior Vice President of the Public Sector.
Copyright©2025 Federal News Network. Unauthorized reproduction is prohibited. This website is not intended for users within the European Economic Area.
