Customer Stories
State Farm needed a governed way to explore AI-enabled capabilities that could help employees reduce friction, find information faster and support better experiences.
State Farm is exploring AI-enabled capabilities through Microsoft Copilot Studio and Power Platform in service of its Human + Digital strategy.
These capabilities help employees work more efficiently while supporting responsible use, governance and human accountability.
State Farm is one of the largest insurance and financial services companies in the United States, supporting customers across auto, home, life, health, and other insurance needs. Headquartered in Bloomington, Illinois, the company employs approximately 65,000 people.
As interest in AI grew across the organization, State Farm focused on creating a clear, enterprise-ready way to responsibly explore, build and scale AI-enabled capabilities. Microsoft Copilot Studio and Microsoft Power Platform are among the tools helping support that work. Microsoft Copilot Studio and Microsoft Power Platform, alongside technologies such as Microsoft Power Apps, Microsoft Power Automate, Microsoft Dataverse, SharePoint Online, Microsoft Entra ID, and the Power Apps vibe experience.
“We are energized by Power Platform adoption at State Farm and look forward to exploring new AI capabilities that can empower our employees and deliver greater value for our customers,” says Brad House, Technology Director at State Farm. State Farm’s acceleration was supported by close collaboration with Microsoft through labs, workshops, new capability guidance, and the Power Up Program. A continuous feedback loop also helped shape Copilot Studio capabilities over time. Together, these efforts helped teams move beyond isolated experimentation toward governed, reusable agent patterns supported at scale.
This approach mirrors State Farm’s relationship-driven culture and operating model. The company treats broad access as essential to democratizing innovation. Employees are encouraged to learn in safe environments, build confidence alongside peers, and share what works—because the people closest to each business challenge are often best positioned to solve it. The strongest ideas are then refined into reliable internal solutions that ultimately improve experiences for customers.
“We are energized by Power Platform adoption at State Farm and look forward to exploring new AI capabilities that can empower our employees and deliver greater value for our customers.”
Brad House, Technology Director, State Farm
State Farm didn’t lack ideas. Employees were curious, builders were eager, and the company already had a strong Power Platform community. The bigger challenge was turning that energy into production-ready solutions that could reduce manual work, speed up answers, and deliver value to employees across HR, underwriting, and onboarding.
As Azsure Dorton, Engineering Manager for Power Platform at State Farm, explains, "Getting governance, risk, and responsible AI right is an important part of the journey." At State Farm, that work became an accelerator, rather than a constraint. Clear approval paths, close alignment with risk and compliance teams, and practical guidance on responsible AI helped reduce hesitation and gave builders a way to move useful ideas into production with confidence.
At the same time, many teams were still dealing with large manuals, repeated policy questions, onboarding tasks, and content or translation workflows that required significant manual effort. HR teams answered policy questions that employees could have self-served. Life underwriters and trainees navigated hundreds of pages of underwriting criteria. Team owners and administrative assistants managed role changes, memberships, security groups, meetings, and channels whenever employees moved between teams.
These pain points created opportunities to explore how AI-enabled digital agents and other capabilities could help employees work more efficiently. Governance helped State Farm scale these solutions safely by earning trust, aligning with enterprise standards, and enabling teams to realize business value without slowing down.
State Farm invested early in hands-on workshops focused on Copilot Studio and responsible AI. Using Microsoft-provided labs as a foundation, the team adapted the content to reflect State Farm’s own governance requirements. Dorton describes this as a “laddering” approach: "Start for yourself. Build for your team. Build for your department. Then you begin to see where an enterprise idea takes shape and can grow."
“Start for yourself. Build for your team. Build for your department. Then you begin to see where an enterprise idea takes shape and can grow.”
Azsure Dorton, Engineering Manager for Power Platform, State Farm
The goal wasn’t to rush builders into production, but to help them build confidence and develop strong habits before scaling their work.
The workshops begin with responsible AI principles, then move into hands-on building. Sessions can reach up to 150 attendees, both virtually and in person. Open office hours give new and returning builders a place to share what they’re working on, learn from others, and get help when they need it.
State Farm has 41 builds that have passed risk assessments and are now in production, supported by a broader community of makers and agent identities within the Power Platform environment. While many early agents focused on knowledge-retrieval, builders are increasingly exploring more complex scenarios, including orchestration across multiple systems.
State Farm also makes it easier for builders to reuse proven ideas. Catalog capabilities allow teams to share connectors and solutions. One innovation sprint produced a reusable onboarding and offboarding solution that combines Power Apps, Power Automate, and Copilot Studio, giving teams a faster way to build on existing work.
3,000+ AI agent identities in Power Platform
The Ask HR policy agent shows the value of self-service in practice. Built in Copilot Studio, the agent allows employees to ask questions about HR policies and benefits instead of manually contacting an HR representative. According to Dorton, the impact was immediate: "This has saved our HR department a lot of time, and it's also helped employees feel confident in finding the information they need on their own." After seeing early success, the HR team began building a roadmap of 12 additional agents.
The same approach extended to underwriting. A data scientist built an agent to help underwriters and trainees search criteria that previously lived in large, hard-to-navigate manuals. By moving those manuals into SharePoint Online, the information became easier for Copilot Studio to surface and easier for underwriters to navigate. These tools help employees find relevant information faster and reduce manual searching. Employees remain responsible for decisions that require judgment, expertise, and care. Within 30 days, the agent became the second most-used agent at State Farm, out of more than 3,000 AI agent identities within the Power Platform environment.
Other teams began applying the same approach to their own work. Translation agents assist with translating marketing materials between English and Spanish, including through an email-based workflow that helped reduce translation turnaround time from 12 months to 24 hours.
A content agent helps writers check content against internal standards, removing the need for a separate license. In information security, a builder created design templates, documentation, and a wiki series that now help other Copilot Studio builders navigate the risk process.
The biggest shift may be cultural. As Dorton puts it, "AI elevates our employees' ability to do their job." Builders who once attended workshops as learners are now becoming champions, joining office hours, sharing templates, and helping others build responsibly.
State Farm’s next chapter builds on what teams have already proven with knowledge-based agents. Building on knowledge-based use cases, teams are considering ways these capabilities could support more complex workflows and coordination across systems. As Dorton explains, the focus is shifting beyond simple retrieval agents: "Right now, we’re looking at moving past those conversational or knowledge-retrieval AI agents towards more complex, action-oriented or autonomous AI agents." To support this shift, State Farm is planning build-a-thons focused on orchestration, system-to-system scenarios, approvals, translation scenarios, and intelligent apps.
“Right now, we’re looking at moving past those conversational or knowledge-retrieval AI agents towards more complex, action-oriented or autonomous AI agents.”
Azsure Dorton, Engineering Manager for Power Platform, State Farm
The company is also considering an AI-enabled capability that could help potential employees find relevant information and better understand the employee experience, with appropriate review and controls in place. In parallel, builders are experimenting with the Power Apps Vibe experience for rapid prototyping, architecture decision support, and executive-level product canvases.
The long-term focus is clear: helping employees reduce friction and support better experiences for customers, agents, and employees. Dorton summarizes success this way: “At State Farm, success with agents will look like employees having personal agents that support their own work, and more importantly enterprise-scale agents that can move across systems to do more work for us.”
“At State Farm, success with AI agents will look like employees having personal agents that support their own work, and more importantly enterprise-scale agents that can move across systems to do more work for us.”
Azsure Dorton, Engineering Manager for Power Platform, State Farm
Follow Microsoft