Key lessons for public service leaders 

Unified access, connected citizen journeys, empowered teams and continuous improvement — all without overhauling legacy systems. A new standard for public service is no longer on the horizon. With agentic AI, it’s already here.

Yet many public sector leaders are just beginning to explore its potential, and questions remain: Where should you start? What does it take to succeed?

To dive into what agentic AI can do and how leaders can seize the opportunity, we hosted a candid conversation with two leading voices in the space: Richie Etwaru, CEO of Mobeus, and Deon James, Head of GenAI Cohort Partner Engineering at Google.

The full discussion is available on demand here. Read on for a sneak peek at the key insights and takeaways.

The problem: A citizen’s journey, not a bureaucratic maze

Imagine trying to start a small business. You need a license, a tax ID, an environmental permit and a dozen other approvals. That's not one task; it's a frustrating scavenger hunt across a dozen different agencies, each with its own website, forms and rules.

This is the reality of government for millions of citizens and businesses every day. The single greatest force driving the need for AI in government isn’t a new technology or a policy — it’s a permanent and irreversible shift in public expectation. Citizens, having experienced the seamless convenience of the private sector, now expect the same from their government.

As Richie Etwaru, CEO of Mobeus, observed:

This reframing transforms citizen experience from a "nice-to-have" into a core mission priority. It's not just about service delivery, but about public value and trust.

The opportunity: One intelligent front door

The biggest potential of agentic AI in government is the power to eliminate this friction by creating a single, intelligent "front door" for citizens. Instead of navigating a maze of agency portals, a user can simply state their need in plain language: "I want to open a coffee shop."

The AI agent then becomes an intelligent guide, a system that not only understands the request but also knows the steps, the agencies and the documents required. It insulates the user from the fragmented complexity on the back end, orchestrating a seamless journey.

This is possible because agentic AI operates on a fundamental shift in logic: it's built for fluidity, not rigidity, as told by Richie Etwaru.

This is the non-negotiable principle for deployment: go where the ambiguity is. To succeed, leaders must pick use cases that thrive on flexibility, insulating users from the messy, probabilistic nature of real-world public service.

A blueprint for deployment: From pilot to impact

With a world of opportunity ahead, how do leaders avoid the "AI POC trap," where promising projects never scale? Our discussion revealed a three-part litmus test for selecting initiatives built for real-world impact:

• Technical suitability: The problem must align with AI’s strengths, focusing on cases where flexibility is needed rather than rigid certainty. Think of an integrated business portal, not a statutory tax calculation.

• Strategic relevance: Avoid recycling old, unfunded projects hoping AI will be a magic wand. Instead, identify new problems that couldn’t be solved before but now fit AI solutions like a glove.

• Mission alignment: Even well-suited use cases need a strong mission and mandate. As Deon James, Head of GenAI Cohort Partner Engineering at Google, emphasized, successful deployments are “mission-driven,” tied to clear human or economic outcomes.

This clear sense of purpose, supported by executive sponsorship, provides the political will and cross-agency collaboration needed to turn promising ideas into meaningful impact.

The starting point: Start small, think big

Getting started with a big transformation can feel overwhelming. The temptation to tackle everything at once (or “boil the ocean”) is strong, but it’s a sure path to costly delays and frustration.

Instead, the key is to adopt a "thin slice" approach.

By zeroing in on a single, high-impact user journey, you can deliver value quickly, prove the concept, demonstrate real ROI and build momentum for future phases. Deon James captured this approach:

This isn't about thinking small forever. It’s about being strategic: using targeted, successful MVPs to learn, adapt and earn the right to take the next, more ambitious step toward your larger goal of a seamless, citizen-first government.

The final shift: This is the worst the tech is ever going to be

Ultimately, succeeding with agentic AI isn't just a technical challenge; it's a profound cultural and organizational one. It demands a new kind of leadership that can embrace a dizzying pace of innovation and plan for a future that arrives faster than ever.

This requires a complete shift in mindset, which was one of the most compelling parts of our conversation. Richie offered a provocative piece of advice that captures this new reality perfectly:

That single idea changes everything about how you plan, invest and lead. 

From the forces driving change to the new blueprint for purpose-driven AI deployment, it’s clear we are at a pivotal moment for government innovation. The insights shared here are just the beginning, offering a glimpse into the future of public service.

In the full webinar, we unpack the critical cultural shifts required to thrive, making it a must-watch for any leader steering their organization into the future.