Bridging the gap between instant AI answers and good government decisions

The next chapter of public sector digital transformation

Public sector digital transformation has been an industry topic of conversation for nearly a decade. Private industries and commercial businesses readily embrace technology and innovative tools, but in government there are many nuances to consider. For instance, government was built to move at the speed of consensus. Modern technology like generative artificial intelligence (genAI) moves at the speed of your WIFI connection. The tension between these points is growing.

Adding to this strain, the people who know how to strike a balance between fast tools and careful decisions are retiring. The leaders and new hires replacing them have only ever worked in a world where an answer is available immediately. This isn't just a technology or people problem, this is a structural gap between 2 things moving at very different speeds, and nobody has built the scaffolding to hold them together yet.

Public sector AI strategy centered on the pace of government

Government AI adoption can give public sector leaders faster answers, quicker insights, and more efficient operations. As this efficiency becomes status quo, the pressure to keep pace grows every day. Sound judgement can’t be forgone in favor of speed, especially in the public sector. Good government decisions require human judgment, context, and often careful deliberation. AI struggles to replicate this kind of thinking, not because it lacks data, but because it lacks human traits like curiosity and productive procrastination, that let ideas develop before anyone acts on them.¹

Here’s a hypothetical scenario to illustrate the point: Picture a state unemployment agency where an AI tool flags a claim as ineligible in seconds. For years, a supervisor with 2 decades on the job could glance at a case like that and sense something was off. This long-term employee had a knack for identifying a pattern the algorithm had no reason to question. That supervisor was an invaluable resource for the agency, and they retired last spring. The caseworker reviewing the flagged claim today has no reason to doubt it either. The claim gets denied. Nobody finds out it shouldn't have been until the audit.

It’s easy to recognize the risk of approving claims that shouldn’t have been, but the risk of denying a claim that should be approved is more subtle. Agencies already deal with a lot of public scrutiny. If they’re using AI for claims processing and it’s leading to errors that require more human intervention, this quickly erodes public trust in AI tools and in the agency. The costs might be more abstract, but they will add up over time.

This is why maintaining checks and balances with humans and technology is a critical function of good government. It’s also part of the necessary guardrails around public sector digital transformation. We know tools like genAI can provide us with answers, but that’s not enough. The challenge for public sector agencies is to determine whether those answers have earned or deserve trust. This isn’t as straightforward as it seems, because the judgment for when not to trust an AI output is usually the last thing anyone gets trained on. Compounding this is the reality that agency employees who knew how to make that call are leaving the workforce right at the onset of government AI adoption.

The silver tsunami is stalling public sector digital transformation

Institutional knowledge is vanishing from state and local government at an alarming scale. With nearly 38% of the local government workforce expected to retire within the next 5 years, the public sector is confronting what experts have named the “silver tsunami.”² This is not just a problem of diminishing headcount; it’s also the loss of judgment built over years of watching decisions play out slowly. Like it or not, a particular form of wisdom only comes with time. Agencies are losing that faster than they can pass on the context behind it.

The silver tsunami exacerbates an existing problem. The public sector has historically struggled to recruit and retain technical talent. In a 2023 report, the U.S. Government Accountability Office (GAO) warned of a “severe shortage of digital expertise, including in the field of AI.”³ As it relates to the silver tsunami, it has listed strategic human capital management as a high-risk area for government since 2001.⁴

Simultaneously, government AI adoption is moving faster than any technology wave before it. Cloud computing took roughly 5 to 8 years to reach government after the private sector adopted it. Mobile-first strategy took 4 to 6 years. The Chief Data Officer role took 3 to 5. AI is taking about 1 to 2 years.⁵ Governors are issuing executive orders, legislatures are creating new offices, and agency organization charts are being redrawn in real time. This wave of retirements and staff departures has real consequences for the success of public sector digital transformation.

Public sector AI strategy and a workforce shaped by instant tools

The rising group of government leaders and the youngest members of their workforce witnessed public sector digital transformation early in their careers. Most of them have never worked without instant access to information. AI is just the newest, most extreme version of this. This isn’t a character flaw or intellectual failing; it's a byproduct of the environment and culture they’ve built their careers in. Even so, the patience required to make fair government decisions used to be developed over years of watching decisions unfold slowly, and that apprenticeship is shrinking just as the tools speed up further.

This isn't unique to government. Employers expect nearly 40% of workers' core skills to change by 2030, with resilience, leadership, flexibility, and analytical thinking growing in importance even as automation increases.⁶ Research on public sector AI adoption backs this up. AI's biggest value comes from amplifying human judgment, creativity, and empathy, not replacing them.⁷ This readiness gap is already visible across industries, not just government.

One recent study found that 63% of organizations have invested in AI training over the past year, yet 74% of technology professionals still say they need to upgrade their skills to stay relevant, and 52% are pursuing that training on their own because internal programs can't keep pace.⁸ The necessary skills of a fast-moving, AI-augmented workforce are still human, and therefore they take the longest to build. Government workforce AI training must consider this, and agencies must manage the process as many of their most experienced people are leaving.

Widespread government AI adoption will happen at a slower pace

Agencies are being told to move at a speed that private sector AI has normalized. This looks like instant drafts, instant recommendations, instant answers. But public sector digital transformation won’t look the same as the transformation that occurred in commercial businesses. The obligations beneath a government decision haven't gotten any faster, nor should they. Procurement law still requires a defined process. Equity review still requires someone to ask who a decision affects. Auditability still requires that a decision can be reconstructed and explained after the fact, not just produced.

At the same time, agencies backfilling retirements are hiring quickly, often through contingent or newly permanent staff. They’re tasking these new hires with workflows where AI tools are already ingrained, without defining what those hires are allowed to do with the tools. There aren’t clear protocols for where accountability sits if something goes wrong. This poses a major risk and hindrance for government AI adoption.

Agencies can lean on managed service providers (MSPs) to help backfill their open positions; this is a clear solution. This is especially helpful when looking for IT professionals with specific and advanced technical expertise. But the right partner can also address the tension between governance, compliance, and new technology that moves lightning fast. MSPs bring external insight and the benefit of working in diverse industries and projects, which can be tremendously beneficial to agencies.

The future of government AI adoption and the public sector workforce

There’s no denying the conundrum that state and local government agencies are facing. The retiring workforce is taking critical knowledge with them. Staffing shortages and the introduction of a contingent workforce have created entirely new ways of working in the public sector. The tension has never been greater.

Agencies that govern too tightly risk frustrating leaders who now expect instant answers by default. Agencies that govern too loosely risk violating compliance and public trust. Most agencies are doing one or the other right now, because almost no one has been able to bridge this gap.

If your agency is feeling the strain between instant tools and careful decisions, CAI's workforce services team can help you determine where that structure needs to go.