---
title: "From Pilot to Priority: Investing in AI Where It Delivers the Greatest Value"
description: State leaders from Illinois, Kansas and Missouri, with Carahsoft’s AI perspective, explain how to move government AI from pilot projects to measurable public value.
image: https://www.stategovtoday.com/hubfs/SGT%20Ep%203%20Midwest_thumbnail%20still.jpg
---

<https://www.stategovtoday.com/state-gov-today-blog/from-pilot-to-priority-investing-in-ai-where-it-delivers-the-greatest-value#body>

[![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

# From Pilot to Priority: Investing in AI Where It Delivers the Greatest Value

Share

- [mailto:?&subject=From%20Pilot%20to%20Priority:%20Investing%20in%20AI%20Where%20It%20Delivers%20the%20Greatest%20Value&body=From%20Pilot%20to%20Priority:%20Investing%20in%20AI%20Where%20It%20Delivers%20the%20Greatest%20Value%0A(https%3A%2F%2Fwww.stategovtoday.com%2Fstate-gov-today-blog%2Ffrom-pilot-to-priority-investing-in-ai-where-it-delivers-the-greatest-value)](mailto:?&subject=From%20Pilot%20to%20Priority:%20Investing%20in%20AI%20Where%20It%20Delivers%20the%20Greatest%20Value&body=From%20Pilot%20to%20Priority:%20Investing%20in%20AI%20Where%20It%20Delivers%20the%20Greatest%20Value%0A(https%3A%2F%2Fwww.stategovtoday.com%2Fstate-gov-today-blog%2Ffrom-pilot-to-priority-investing-in-ai-where-it-delivers-the-greatest-value))
- <https://www.linkedin.com/shareArticle?mini=true&url=https%3A%2F%2Fwww.stategovtoday.com%2Fstate-gov-today-blog%2Ffrom-pilot-to-priority-investing-in-ai-where-it-delivers-the-greatest-value&title=From%20Pilot%20to%20Priority:%20Investing%20in%20AI%20Where%20It%20Delivers%20the%20Greatest%20Value&summary=&source=>
- <https://twitter.com/intent/tweet?text=From+Pilot+to+Priority%3A+Investing+in+AI+Where+It+Delivers+the+Greatest+Value&url=(https%3A%2F%2Fwww.stategovtoday.com%2Fstate-gov-today-blog%2Ffrom-pilot-to-priority-investing-in-ai-where-it-delivers-the-greatest-value)>
- <https://www.facebook.com/sharer/sharer.php?u=https%3A%2F%2Fwww.stategovtoday.com%2Fstate-gov-today-blog%2Ffrom-pilot-to-priority-investing-in-ai-where-it-delivers-the-greatest-value>

<iframe style="position:absolute;top:0;left:0;width:100%;height:100%;border:0;" title="YouTube video player" allow="accelerometer; autoplay; clipboard-write; encrypted-media; gyroscope; picture-in-picture; web-share" referrerpolicy="strict-origin-when-cross-origin" src="https://www.youtube.com/embed/z1amxCkbfWI?si=xfxs7PK_0wC06S0I" width="560" height="315" frameborder="0" allowfullscreen></iframe>

*Presented by Granicus in partnership with Carahsoft.*

This program examines how state governments can move artificial intelligence from experimentation into trusted operations. Hosted by George Jackson, the conversation brings together state IT leaders from Illinois, Kansas and Missouri with an industry perspective from Carahsoft. Their discussions examine the practical work behind high-value AI: setting policy, preparing data and employees, choosing use cases, protecting residents and determining when a pilot is ready to become part of everyday operations.

### Moving Illinois from AI Governance to Enterprise Operations

**Brandon Ragle, Secretary, Department of IT, State of Illinois**

![Screenshot 2026-11-02 at 1.50.06 PM](https://www.stategovtoday.com/hs-fs/hubfs/Screenshot%202026-11-02%20at%201.50.06%20PM.png?width=450&height=254&name=Screenshot%202026-11-02%20at%201.50.06%20PM.png)Illinois is entering a new phase in its AI work. After spending time establishing controls and governance, the state is building the operation needed to guide AI across government. Brandon Ragle describes this as a shift from setting the framework to staffing and running it: defining responsibilities, developing procedures and creating a dedicated capacity that can help the state move from individual experiments toward sustained use. He says Illinois is working toward a fully dedicated chief AI officer role, supported by an operation that can coordinate policy, data and technology decisions.

Ragle’s enterprise-first approach reflects the size and variety of Illinois government. The state supports roughly 36 executive branch agencies and many additional public entities, each with its own mission and information needs. With limited resources, he says the state has to ask where one investment can solve more than one problem. Reusable capabilities can help agencies avoid duplicating tools and make it possible to scale a successful approach beyond the original use case. That enterprise view depends on understanding the information behind the work. Data must be organized, classified and protected before agencies can determine whether it can be used safely and responsibly with AI.

The interview also connects AI value to the experience of employees and residents. Ragle says Illinois looks for opportunities that are relatively low cost and low risk while offering meaningful potential benefit. An employee-facing tool may help staff work more effectively; that improvement can then affect the quality or speed of service residents receive. The state therefore needs to look at who benefits, how many people are affected and whether the solution improves the work of the employees delivering public services. The business problem should come first. Technology is one possible response, considered after agencies understand the need, the people involved and whether the organization can adopt the change.

Ragle distinguishes a pilot from production. A pilot or proof of concept can show whether a technology is feasible, whether a use case exists at meaningful scale and whether legal, security and governance questions can be addressed. Moving into production requires a further operating plan: data protection, cybersecurity, compliance, training, change management and the capacity to support the system over time. He also stresses that staff need preparation and a reason to trust a new tool. Introducing technology without helping employees understand it or incorporate it into their work can undermine adoption even when the tool itself works.

To connect these responsibilities, Illinois has brought AI, data and enterprise architecture expertise together through a Strategy and Innovation unit. The interview’s central message is that responsible scale depends on this foundation: understand the mission, make governance workable, prepare the data and workforce, and then evaluate how technology can contribute to better outcomes.

**3 Key Takeaways**

- **Build the operating model:** Staffing, procedures, governance and data readiness support durable AI operations.
- **Prioritize reusable enterprise solutions:** Start with shared problems and capabilities that can help more than one agency.
- **Measure value through people and mission outcomes:** Assess whether employees can do their work more effectively and residents receive better service.

### A Whole-of-State Path from Policy to Practice

**Jeff Maxon, Chief Information Officer, State of Kansas**

![Screenshot 2026-11-02 at 1.48.18 PM](https://www.stategovtoday.com/hs-fs/hubfs/Screenshot%202026-11-02%20at%201.48.18%20PM.png?width=450&height=252&name=Screenshot%202026-11-02%20at%201.48.18%20PM.png)Kansas has been working on generative AI governance since 2023. Jeff Maxon describes the state’s next steps as a combination of clearer policy, practical use cases and broader access to expertise. Updated guidance addresses transparency when AI is used in public-facing services, as well as safeguards for sensitive information and expectations for vendors. The aim is to give agencies a consistent foundation for responsible adoption while recognizing that the right solution depends on the work each agency needs to do.

Maxon places AI within Kansas’ whole-of-state approach to technology and cybersecurity. Many communities in the state are rural, and local governments may not have the same staff or resources as a large state agency. Shared services and statewide coordination can help extend access to tools and technical knowledge. Maxon’s point is not that every locality should use the same product for every need; it is that common standards and shared capacity can make useful capabilities more available while reducing fragmentation. That same discipline can help agencies avoid buying multiple tools that perform similar functions.

The state’s early applications include employee productivity and policy support. Kansas introduced a limited chatbot to state employees so they could begin exploring what the technology can do, with guidance and training around its use. Maxon also points to policy bots designed to help caseworkers find relevant procedures, statutes and policies while working through complex cases. Casework is repetitive in some respects, but every person’s circumstances are different. The intended role of the tool is to help employees locate pertinent information more quickly so they can apply their expertise and serve people; it does not remove the need for professional judgment.

AI is also being considered as a way to understand legacy technology. Long-running systems may be poorly documented, and staff who knew how they worked may have moved on. Tools that help explain code or identify what a system does can give agencies a clearer starting point for modernization. That work can help the state make decisions about what to retain, connect or replace, rather than assuming that every older system must be discarded at once. Maxon ties modernization to standards and shared expertise: a fragmented environment can make integration harder and more expensive, while enterprise coordination can help agencies move at a more manageable pace.

Workforce preparation is part of the same effort. Maxon describes a need for broad awareness alongside training that reflects employees’ roles and the tools they will use. He also emphasizes that IT should respond to business objectives, not promote a technology simply because it is new. Value may be reflected in cost, employee workload or how easily residents complete government processes. Kansas’ approach is therefore to define the objective, choose an affordable solution that fits it, and support staff and local partners with the governance and training needed to use it well.

**3 Key Takeaways**

- **Set clear policy and transparency expectations:** Address sensitive information, public-facing AI and vendor responsibilities.
- **Extend capability across the state:** Shared services can help rural and local partners access useful technology and expertise.
- **Connect tools and training to business needs:** Use AI to support defined work, with learning tailored to employees’ roles.

### Choosing AI Investments That Deliver Public Value

**Mike Adams, AI Program Executive, Carahsoft**

![Screenshot 2026-11-02 at 1.49.35 PM](https://www.stategovtoday.com/hs-fs/hubfs/Screenshot%202026-11-02%20at%201.49.35%20PM.png?width=450&height=251&name=Screenshot%202026-11-02%20at%201.49.35%20PM.png)From the industry perspective, Mike Adams focuses on how public-sector organizations can decide which AI opportunities merit investment and how to tell whether they are working. He describes a shift in the market: agencies have moved from initial experimentation and familiar tasks such as summarizing information or reducing repetitive work toward questions about production use cases, agents and more substantial changes to how work gets done. For Adams, that shift makes disciplined selection more important. A compelling demonstration is not enough; a use case has to connect to a mission need and produce an outcome the organization can observe.

That begins with defining the problem before choosing a tool. An agency should know who is affected, what part of the work needs to improve and how it will recognize progress. Depending on the mission, the desired result may be time returned to employees, faster processing, a better interaction with a constituent or added capacity for a team. Adams cautions against treating return on investment as a single financial calculation. Public value can include dollars, but it also includes the quality and accessibility of service and whether employees can focus on work that requires their knowledge and judgment.

Pilots can help agencies collect evidence, provided they are designed to answer useful questions. Teams can compare the process before and after a test, determine whether the solution fits the workflow, and identify what would be needed to continue. Adams urges leaders to consider whether the organization has the people, infrastructure and culture to support the tool. They should also understand costs over time. AI usage can vary, and consumption-based charges—such as costs associated with tokens or other usage measures—may make a solution harder to budget if leaders do not have clear visibility into expected and actual use.

Moving from a pilot to an operational service also raises acquisition and integration questions. Agencies need to understand how a solution will work with existing systems, what support is required and whether the vendor can explain its capabilities and limitations. Adams encourages practical conversations with providers: ask how similar organizations have used the technology, what outcomes they achieved, what challenges they encountered and what the solution requires to operate securely. Cooperative purchasing options and knowledgeable implementation partners can help government teams navigate procurement and deployment, but the agency still needs to retain a clear view of its objectives, costs and responsibilities.

Adams’ advice is to treat scale as a decision supported by evidence. Define the mission outcome first, establish measures before the pilot, evaluate both benefits and operating requirements, and make sure the organization can sustain the solution. If the results do not justify the cost or the operational commitment, a pilot can still be useful by showing what needs to change. If the evidence is strong, leaders can use it to plan a responsible next step rather than moving forward on enthusiasm alone.

**3 Key Takeaways**

- **Start with a mission problem:** Select use cases because they address a clear need for an agency or the people it serves.
- **Define success before testing:** Measure service, workforce and operational outcomes alongside financial return.
- **Plan for ongoing costs and support:** Understand usage, integration, security and staffing needs before scaling.

### Putting Pilots to Work Across Missouri

**Tim Marczewski, Director of AI and Innovation, State of Missouri**

![Screenshot 2026-10-27 at 5.14.46 PM](https://www.stategovtoday.com/hs-fs/hubfs/Screenshot%202026-10-27%20at%205.14.46%20PM.png?width=450&height=254&name=Screenshot%202026-10-27%20at%205.14.46%20PM.png)Missouri is testing AI across resident services and internal operations while building the safeguards and workforce readiness needed to use it responsibly. Tim Marczewski describes public-facing assistants as one way to help people navigate government information, but also as a source of insight for agencies. Questions residents ask can show where information is hard to find or what people need from government. The state monitors these tools and adjusts them; they are not meant to be launched and left alone. Missouri is exploring ways to make service interactions easier across channels and to provide clearer updates as an application moves through a process, so residents are not left waiting without knowing what is happening.

The conversation also looks at the role of AI in more complex workflows. Missouri used an AI assistant to help review 3,000 handwritten beef and dairy grant applications. Marczewski says the work was completed in about 70 hours, compared with more than 1,000 hours of review time. The assistant helped people move through the applications, but human reviewers remained involved at each step. That example makes efficiency measurable while preserving human accountability. Marczewski says agencies should also consider whether people receive benefits on time, whether an interaction is resolved and whether residents are satisfied. Public-sector return includes more than financial savings.

Marczewski says the lessons from one project may travel farther than its code. Different programs can have distinct rules and inputs, so a model or application built for one purpose may not transfer directly to another. Agencies may still be able to reuse the workflow, the review process and the guardrails. In consequential work such as adjudication, AI should assist the person reviewing a case, not make the final decision. A shared approach can help employees apply consistent procedures while leaving responsibility with the people authorized to decide.

Security is built into the testing process. Drawing on experience in secure environments, Marczewski emphasizes creating a properly configured sandbox where teams can test a system, identify weaknesses and correct configuration problems before broader use. Agencies also need privacy protections, monitoring and clear limits on what information can be used. Missouri is considering AI as a way to support modernization too: teams can use it to examine legacy code, understand how an older program works and explore ways to connect it with newer systems without interrupting services that residents depend on.

Workforce training has evolved through trial and adjustment. Small pilots helped develop employee champions, but a large general training session did not work equally well for everyone. Missouri is moving toward more focused instruction for business and functional areas, helping employees understand the approved tools, security requirements, what information they can enter and their responsibilities when using AI. The state’s internal GenAI portal brings training and certification together with access to approved tools in a secure environment. Marczewski describes this as a way to prepare employees while building the state’s ability to operate and maintain the platform over time.

**3 Key Takeaways**

- **Use pilots to measure real outcomes:** Missouri’s grant review illustrates how a test can quantify time saved while informing a scale decision.
- **Keep people responsible for consequential decisions:** AI can assist review, with human judgment at each decision point.
- **Pair secure access with tailored training:** Provide approved tools, role-focused learning and protected test environments.

*Presented by Granicus in partnership with Carahsoft.*

[![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)

 

<https://linkedin.com/company/state-gov-today/>

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" : "2026-11-18T17:38:57.920Z",
  "datePublished" : "2026-11-18T14:36:28.000Z",
  "headline" : "From Pilot to Priority: Investing in AI Where It Delivers the Greatest Value",
  "image" : [ "https://www.stategovtoday.com/hubfs/SGT%20Ep%203%20Midwest_thumbnail%20still.jpg" ],
  "mainEntityOfPage" : {
    "@id" : "https://www.stategovtoday.com/state-gov-today-blog/from-pilot-to-priority-investing-in-ai-where-it-delivers-the-greatest-value",
    "@type" : "WebPage"
  },
  "publisher" : {
    "@type" : "Organization",
    "logo" : {
      "@type" : "ImageObject",
      "url" : "https://www.stategovtoday.com/hubfs/logo-sgt-1080x1080.png"
    }
  }
}
```