Donald Gainsborough is widely recognized as a political savant, a leader whose deep understanding of policy and legislation has shaped the strategic direction of numerous public sector initiatives. As the head of Government Curated, he sits at the intersection of traditional governance and the cutting-edge technological shifts that are currently redefining how agencies function. In this discussion, we explore a fundamental transformation in the public sector: the transition from viewing artificial intelligence as a mere tool for automation to embracing it as a catalyst for high-speed decision-making. Gainsborough provides a deep dive into how shortening the “time to insight” can bridge the gap between data collection and meaningful action, ensuring that government programs move at the pace of modern challenges.
The conversation centers on the idea that the true value of AI in government is not just doing the same work with fewer people, but radically accelerating the timeline from evidence to implementation. We delve into specific examples from the healthcare sector, such as improvements in hospital inquiry responses and the streamlining of complex regulatory reviews. Gainsborough emphasizes the shift from measuring success via static reports to measuring it by the reduction in lag time. Furthermore, the dialogue touches on the necessity of robust governance, the role of human experts in maintaining analytic rigor, and the future of AI as a critical piece of decision-support infrastructure that helps leaders navigate an increasingly complex information landscape.
How can the acceleration of decision-making fundamentally reshape the way government programs operate on the ground, and what does it mean for the people waiting on those decisions?
In the traditional public sector model, the wheels of bureaucracy often turn with a heavy, deliberate friction that can feel like wading through deep water for those on the outside. We have historically measured the process from the moment data is ingested to the point where an actual action-ready insight is produced in months or even years, which is an eternity for a citizen waiting for a service or a hospital administrator needing urgent guidance. By integrating artificial intelligence, we are seeing a shift where these grueling cycles are being compressed into days, hours, or in some cases, mere minutes. This isn’t just about efficiency for its own sake; it is about making the government feel responsive and alive to the lived realities of the people it serves. When a leader can gain the insights they need to act almost as soon as the data becomes available, the entire nature of public service changes from a reactive, lagging posture to one that is proactive and dynamic. It creates a sense of relief for stakeholders who are used to the silence of long administrative delays, replacing that frustration with the clarity of timely, evidence-based results.
When you speak about “time to insight” as a new metric for success, how does this differ from the way we have historically evaluated the effectiveness of public sector innovation?
For decades, the public sector has invested heavily in the sheer act of collecting data, often resulting in vast digital warehouses that are difficult to navigate and even harder to translate into policy. We used to evaluate success by the production of traditional outputs—the thickness of a report, the complexity of a dashboard, or the number of analytic products generated in a fiscal year. While these remain essential, they often contribute to the “noise” rather than the “signal,” leaving executives overwhelmed by the volume of information without a clear path forward. AI shifts the goalposts by focusing on the lag between evidence and action as the primary indicator of innovation. If an agency can identify an emerging risk or a program deficiency in near real time rather than waiting for a biennial review, the value of that data increases exponentially. We are moving away from valuing the “product” of analysis and starting to value the “velocity” of the insight, which allows agencies to move from identification to implementation with a speed that matches the urgency of the challenges they face.
Could you walk us through a specific instance where this technology has already moved the needle, particularly in a high-stakes environment like healthcare administration?
One of the most compelling examples of this shift in action is occurring at the Centers for Medicare & Medicaid Services, where the pressure to provide accurate and timely guidance to hospitals is a daily reality. Specifically, when hospitals seek guidance on complex quality measures—such as those related to the Severe Sepsis and Septic Shock protocols—the turnaround time for inquiries is critical because these measures directly influence how care is delivered and evaluated. By utilizing an AI-powered tool to generate draft responses and integrate them directly into the workflow, the agency has seen a remarkable 35% reduction in inquiry turnaround time. Perhaps even more telling is the 18% increase in the share of questions that are fully resolved within the same month they are received. These aren’t just cold statistics; they represent a tangible reduction in administrative burden for hospital staff and a 2% decrease in reopened inquiries, which suggests that the guidance provided is not just faster, but clearer and more complete the first time around.
The process of rulemaking and regulatory review is often described as a mountain of paperwork. How is AI changing the experience for the experts who have to climb that mountain every year?
The annual rulemaking cycle, particularly for something as dense as the Medicare Inpatient Prospective Payment System, is a Herculean task that involves reviewing thousands of pages of technical regulatory text and manually comparing changes across different years. Historically, this required experts to spend hundreds of hours on the “drudgery” of tracking line-by-line modifications, which is an exhausting and mentally taxing process that leaves little room for high-level synthesis. Generative AI tools are now stepping in to perform these comparisons in minutes, acting as a powerful mechanical assistant that handles the tedious labor of identification. This frees the human experts to step back from the granular data and focus their intellectual energy on interpreting the actual policy implications of those changes. Instead of being buried under a mountain of paper, these professionals are now empowered to spend their time weighing trade-offs and making better-informed decisions that ultimately shape how hospitals are reimbursed for inpatient care.
Beyond data processing, how does AI help in more human-centric environments, such as when researchers are trying to find common themes across diverse national discussions?
There is a common misconception that AI is only useful for structured numbers, but its ability to synthesize human conversation is proving to be a game-changer for research and convening work. Recently, in a large-scale initiative involving 18 national symposia focused on the future of health services research, the challenge was to capture and analyze the rich, diverse discussions of hundreds of participants. Traditionally, this would involve weeks of manual transcription and painstaking synthesis by researchers trying to find the “red thread” across various sessions. By using AI-assisted transcription and synthesis tools, researchers were able to identify cross-cutting themes within a matter of days. This acceleration doesn’t just save time; it preserves the momentum of the conversation, allowing the insights from the symposia to be funneled back into the policy development process while the ideas are still fresh and the stakeholders are still engaged.
In the context of oversight and compliance, particularly for something as complex as Medicaid Section 1115 demonstrations, how does AI manage to handle the scale without losing the necessary rigor?
Oversight of Medicaid demonstrations is a massive undertaking because it involves reviewing lengthy, complex monitoring and evaluation documents that must be checked against federal guidance and historical precedents. The scale of these programs can be overwhelming, and maintaining consistency across hundreds of reports is a significant challenge for human analysts alone. We have seen the development of AI-assisted workflows that help teams review these complex reports more consistently and rapidly than ever before. These tools act as a first-line filter, highlighting key areas of concern and ensuring that every document is held to the same rigorous standard. This allows the experts to focus their oversight efforts on the most critical findings and provides them with the mental bandwidth to offer more nuanced advice to policymakers. It’s a perfect example of AI functioning as “decision-support infrastructure,” allowing the government to manage growth and complexity without sacrificing the quality of its supervision.
With the ability to make decisions so much faster, there is a natural fear that we might simply be making “bad decisions more quickly.” How can agencies ensure they are maintaining governance and transparency?
This is the most critical question we face because faster insights are a liability if they are not paired with a strengthening of governance and analytic rigor. If we accelerate the production of insights without a transparent framework, we risk creating a new set of problems where speed leads to a lack of accountability. Responsible adoption of AI requires that these tools be embedded within well-designed, human-centric workflows where the expert remains the final arbiter of truth. We must be incredibly disciplined about ensuring that models are transparent regarding their limitations and that agencies establish clear governance structures to guarantee that AI-assisted insights are reliable and reproducible. The goal is to use AI to cut through the noise and highlight uncertainties, not to replace the critical thinking of a human policymaker. We have to preserve the rigor that public policy demands by ensuring that every AI-generated suggestion is scrutinized and aligned with our broader policy goals before it is ever acted upon.
As we look toward the future, how do you see the relationship between data collection and data implementation evolving within the public sector?
For the last several decades, the public sector’s primary mission was the massive investment in collecting and digitizing data, which was a necessary first step, but it often felt like we were building a library that no one had the time to read. The next frontier is moving past the collection phase and ensuring that this data actually translates into action fast enough to matter in the real world. We are entering an era where the value of an agency will be judged by how quickly they can close the gap between identifying a problem and deploying a solution. This might mean identifying an emerging public health risk in near real time or synthesizing evidence across multiple programs in a few days rather than waiting a full fiscal cycle. The future of public-sector innovation lies in the ability to operate at the speed of modern challenges, transforming our vast data repositories into a living, breathing engine for better governance.
What is your forecast for the role of AI in government over the next five years?
My forecast is that within the next five years, the concept of “time to insight” will become the universal benchmark for success across all federal and state agencies, effectively replacing the old-school metrics of volume and output. We will see AI move from being a “special project” or a standalone tool to being an invisible, foundational layer of the decision-making infrastructure, much like the internet is today. I expect that we will see a significant shift in the workforce, where public servants are no longer defined by their ability to process information, but by their ability to interpret AI-synthesized insights and apply human judgment to the most complex trade-offs of society. If we continue to prioritize transparency and human oversight, we will reach a point where the lag between a citizen’s need and the government’s response is measured in moments, finally fulfilling the promise of a truly agile and evidence-based public sector. It is an exciting, albeit demanding, path forward that will require us to rethink not just our technology, but our entire philosophy of administrative action.
