--- title: Building the Foundations for AI-Enabled Public Service description: Artificial intelligence can help state and local governments improve the services residents rely on, but moving from experimentation to sustained results requires more than selecting a model or launching a pilot. Agencies need clear mission objectives, strong information foundations, and the ability to understand how AI systems perform as they become part of everyday operations. image: https://www.stategovtoday.com/hubfs/karthick_thumb.jpg --- [![logo-sgt-white](https://www.stategovtoday.com/hs-fs/hubfs/logo-sgt-white.png?width=146&height=146&name=logo-sgt-white.png "logo-sgt-white")](https://stategovtoday.com) - [Episodes](https://www.stategovtoday.com/state-gov-today-blog) - [Events](https://www.stategovtoday.com/events) - [About](https://www.stategovtoday.com/about) - [The Future Public Servant](https://www.stategovtoday.com/the-future-public-servant) - [Episodes](https://www.stategovtoday.com/state-gov-today-blog) - [Events](https://www.stategovtoday.com/events) - [About](https://www.stategovtoday.com/about) - [The Future Public Servant](https://www.stategovtoday.com/the-future-public-servant) Search Search - [CONTACT US](https://www.stategovtoday.com/contact) - [CONTACT US](https://www.stategovtoday.com/contact) Search Search # Building the Foundations for AI-Enabled Public Service Share - [mailto:?&subject=Building%20the%20Foundations%20for%20AI-Enabled%20Public%20Service&body=Building%20the%20Foundations%20for%20AI-Enabled%20Public%20Service%0A(https%3A%2F%2Fwww.stategovtoday.com%2Fstate-gov-today-blog%2Fbuilding-the-foundations-for-ai-enabled-public-service)](mailto:?&subject=Building%20the%20Foundations%20for%20AI-Enabled%20Public%20Service&body=Building%20the%20Foundations%20for%20AI-Enabled%20Public%20Service%0A(https%3A%2F%2Fwww.stategovtoday.com%2Fstate-gov-today-blog%2Fbuilding-the-foundations-for-ai-enabled-public-service)) - - - *Presented by Granicus in Partnership with Carahsoft*   Artificial intelligence can help state and local governments improve the services residents rely on, but moving from experimentation to sustained results requires more than selecting a model or launching a pilot. Agencies need clear mission objectives, strong information foundations, and the ability to understand how AI systems perform as they become part of everyday operations. In episode 4 of **The Future Public Servant**, **Karthik Anbalagan, General Manager of Emerging Technologies at Granicus**, joins **George Jackson** to discuss how governments can build those foundations. His perspective connects AI adoption to the broader work of service delivery: aligning leaders around outcomes, establishing governance, maintaining human oversight, and using resident insights to guide future improvements. For Anbalagan, the opportunity begins with the purpose that already drives public servants. Government leaders focus on delivering services and experiences for the people they serve. Technology can accelerate that work when its deployment is tied to those responsibilities. #### Connect AI pilots to mission priorities [![Post Summit\_300 x 250 copy](https://www.stategovtoday.com/hs-fs/hubfs/Post%20Summit_300%20x%20250%20copy.png?width=300&height=250&name=Post%20Summit_300%20x%20250%20copy.png)](https://granicus.com/watch-gxvs/?utm_campaign=373931715-GXVS2026&utm_source=State_Gov_Today)The transition from pilot to scale starts with identifying an organization’s most critical objectives. Anbalagan notes that pilots often address easier projects or limited areas of friction. Those experiments can help agencies understand the technology, but they may have a weak connection to the broader outcomes that matter most. More consequential objectives are often harder to tackle because they involve multiple departments, stakeholders, or processes. That complexity makes it especially important to define the intended result before choosing where AI should fit. “And so the objective orientation, I would say, is job number one,” Anbalagan says. Once agencies understand the larger objective, they can break it into smaller components. Pilots then become a way to test how AI can be operationalized within a broader plan, with a clearer path toward material, measurable results. This approach places experimentation in the context of institutional capability. The value of a pilot lies partly in what it teaches an agency about delivering outcomes across its operations, beyond the initial test. #### Build a coalition that can put AI into practice Scaling AI also requires participation from the people responsible for both technology and service delivery. Anbalagan identifies technology leaders as essential to governance policy, security, and architectural decisions. They need to understand the agency’s existing technology environment and how new AI capabilities will work within it. Program owners bring a different but equally important perspective. They often manage resident-facing services or the processes where automation could improve speed, efficiency, and the overall experience. Department heads and agency executives help align these efforts with organizational priorities. Bringing those stakeholders together early creates an opportunity to agree on objectives, determine how a plan will be operationalized, and address the change management required to use AI effectively. The conversation highlights an organizational challenge alongside the technical one. A successful deployment depends on a coalition that understands what the agency wants to accomplish and how its people and processes must adapt to support that work. #### Make governance part of the service delivery foundation Anbalagan sees a pattern among organizations with strong governance foundations: they understand the business processes that matter most and have invested in their underlying data, architecture, and content. Those investments help agencies define the principles their governance policies must preserve. Information architecture is particularly relevant to resident-facing services, where the organization of content contributes to the experience an AI system can support. “And so there’s no one size fits all approach to governance,” he says. Governance rules need to reflect the services an agency provides, the residents it serves, and its mission objectives. Agencies with a clear understanding of those factors are better positioned to articulate requirements and select partners that can work within them. In this view, AI governance belongs within the agency’s overall service delivery infrastructure. Data quality, content organization, architectural decisions, and operational objectives provide the basis for deciding how AI should be used and what standards it must meet. #### Evaluate partners against public sector needs ![Screenshot 2026-10-12 at 4.14.50 PM](https://www.stategovtoday.com/hs-fs/hubfs/Screenshot%202026-10-12%20at%204.14.50%20PM.png?width=450&height=317&name=Screenshot%202026-10-12%20at%204.14.50%20PM.png)Anbalagan encourages agencies to examine how deeply a prospective technology partner understands government workflows and resident experience. He argues that public sector use cases can carry demanding expectations for security, accuracy, and visibility into system behavior. Agencies should ask what the provider’s solutions are designed to solve, how much public sector experience it brings, and what data, context, and expertise inform its approach. They should also ask partners to explain their evaluation methods and how they provide visibility into AI performance. These questions help establish whether a solution’s underlying frameworks are suited to government use cases. Anbalagan cautions that a provider with a broader or diluted focus may overlook requirements or edge cases that matter to an agency. Partner selection therefore becomes part of the foundation for implementation. The agency’s governance requirements and the provider’s technical capabilities need to support the same service objectives. #### Keep human oversight as AI workflows grow One of the interview’s central themes is observability: the ability to see what an AI system is doing and assess whether its behavior meets expectations. For a system that retrieves information and responds to resident questions, Anbalagan describes visibility into inputs, outputs, and conversations. That record helps an agency understand how the system is performing and whether it is doing the work intended. The requirement becomes more complex when AI agents perform tasks across systems and communicate with one another to complete a workflow. Agencies then need visibility into the inputs and outputs of individual agents and their interactions along the way. Anbalagan explains that AI behavior can vary, creating the possibility that a workflow will drift from its intended performance. A robust monitoring framework can help identify changes before they become problematic. He describes evaluation and observability as core areas of focus for Granicus, supporting transparency and opportunities for human oversight across resident-facing and internal workflows. As automation takes on more repetitive or complex work, this visibility gives public servants a basis for supervising it. Agencies need to understand the process behind a completed task, along with the result, to determine whether it meets their expectations. #### Use resident insights to guide improvements For Anbalagan, resident analytics and digital service delivery are closely connected. Agencies need to understand how people actually use their services before they can determine where improvements are needed. Insights can reveal where residents encounter friction, where they succeed, and where gaps remain in the overall experience. Agencies can then turn that information into actions that improve services and remove obstacles. The process creates a feedback loop: observe how services are used, make changes, and measure whether the experience is improving over time. That evidence also helps guide the next set of investments. Anbalagan notes that residents bring expectations shaped by the consumer technologies and services they use. Government leaders need data to assess whether their digital services are meeting objectives and moving in the right direction. This connection makes measurement part of ongoing service management. AI adoption should be assessed through its contribution to the experience agencies deliver and the outcomes they seek. #### Make AI costs predictable The interview also addresses a practical obstacle to sustained adoption: unpredictable costs. Anbalagan explains that AI compute expenses can vary with model choices, workloads, and effort levels. Higher spending does not always produce a meaningfully better outcome. He argues that agencies should work with partners that focus on business and operational objectives while managing the technical variables behind compute costs. Flexible architecture and model choices can give providers room to manage those expenses and offer a more predictable pricing basis. Rather than passing through a bill based entirely on token usage, he describes pricing tied to a proxy for expected volume, such as population or the number of phone calls. The appropriate basis would be determined with the customer. “We want real predictability,” Anbalagan says. For governments operating within fixed budgets, that predictability is part of making AI operationally sustainable. The interview connects responsible scaling to both dependable system performance and costs agencies can plan for. Three key takeaways 1. **Start with mission outcomes.** Connect pilots to critical organizational objectives and involve the leaders responsible for implementing change. 2. **Build foundations for accountable AI.** Data, content, architecture, governance, and visibility into system behavior support effective deployment and human oversight. 3. **Plan for sustained service improvement.** Use resident insights to guide decisions and work toward predictable costs that fit government budgets. #### About The Future Public Servant **The Future Public Servant** is a **State Gov Today** program presented by **Granicus in partnership with Carahsoft**. Through conversations with government leaders and technology experts, the series examines how artificial intelligence is changing the way government works, the services it delivers, and the capabilities public servants need for the future. Explore the individual interviews at [TheFuturePublicServant.com](https://TheFuturePublicServant.com). [![logo-sgt-1080x1080](https://www.stategovtoday.com/hs-fs/hubfs/logo-sgt-1080x1080.png?width=150&height=150&name=logo-sgt-1080x1080.png "logo-sgt-1080x1080")](https://stategovtoday.com) [Newsletter Signup](https://www.stategovtoday.com/newsletter-signup) | [Privacy Policy](https://www.stategovtoday.com/privacy-policy)   Copyright © Fed Gov Today, LLC 2026. All Rights Reserved. ```json { "@context" : "https://schema.org", "@type" : "BlogPosting", "author" : { "@type" : "Person", "name" : "State Gov Today", "url" : "https://www.stategovtoday.com/state-gov-today-blog/author/state-gov-today" }, "dateModified" : "2026-10-13T16:07:26.553Z", "datePublished" : "2026-10-13T15:11:17.000Z", "headline" : "Building the Foundations for AI-Enabled Public Service", "image" : [ "https://www.stategovtoday.com/hubfs/karthick_thumb.jpg" ], "mainEntityOfPage" : { "@id" : "https://www.stategovtoday.com/state-gov-today-blog/building-the-foundations-for-ai-enabled-public-service", "@type" : "WebPage" }, "publisher" : { "@type" : "Organization", "logo" : { "@type" : "ImageObject", "url" : "https://www.stategovtoday.com/hubfs/logo-sgt-1080x1080.png" } } } ```