Government agencies at all levels are searching for the answer to one question: What can AI do for me today? An underused AI application is to generate leading indicators of issues on the horizon, empowering agency leaders to get ahead of challenges before they become full-blown crises.
And all it requires is a mindset shift.
THE LIMITS OF TRADITIONAL DASHBOARDS
In very broad strokes, government agencies are working with quantitative and qualitative data. Agencies most commonly view quantitative data (hard numbers and program metrics) in dashboards. In other words, it is traditional data.
This data is critical for assessing the efficacy of an initiative after-the-fact, and AI applied here can accelerate the production of data visualizations and analyses. But such aggregated quantitative dashboards nonetheless lag behind real-time, ground truth. This is, in part, by design. And it has real utility in government operations.
But it doesn’t help agency leaders see around corners. In other words, it doesn’t surface leading indicators that can help government leaders detect gaps and prevent crises before they happen.
QUALITATIVE DATA IS DIFFERENT
Qualitative data manifests in things like questions asked and resources needed by public servants. To identify what public servants on the front lines are thinking, feeling, and asking about, an agency typically needs an army of consultants to manually collect this information and collate it into something useful. This can take weeks, if not months.
AI can dramatically expedite and scale qualitative data analysis to capture the “why” behind the “what.”
Data from a case management system, for example, is quantitative data that pipes into a dashboard. Qualitative data is the transcript of the weekly task force meeting where questions are asked and answered before the case management system is updated.
What happens in the community meeting or conversations between peers translating policy into practice is also valuable qualitative, real-time data.
I’d guess most agency leaders don’t think about taskforce meetings as critical, real-time data that is useful input for AI systems. And that’s the mindset shift we need.
THE INSIGHTS BURIED IN QUALITATIVE DATA
Almost every state and municipal coordinator in agencies across the country addressing homelessness will have a view of key metrics, like number of people experiencing homelessness in their jurisdiction and the number of available beds in shelters.
If the number of people experiencing homelessness goes up or available beds go down, the dashboard reflects that data after the fact.
But imagine that same coordinator knew that across their state, frontline public servants were asking their peers about eviction rules. After a number of government workers received calls from constituents about tenants’ rights, they seem concerned.
AI that ingests that data in real time—the public servants’ questions and what they hear from constituents—can surface the signal early. If frontline workers are looking up eviction regulations and people are calling about tenants’ rights, there could be an uptick of people experiencing homelessness in the near future. Agency leaders can then prepare housing support programs before the surge occurs.
AI SURFACES LEADING INDICATORS
Functionally, collective intelligence from field activity becomes a leading indicator of potential public policy challenges. It’s a fundamental shift, from “we see a problem based on reactive data, what caused it?” to “we see early indications of a problem, let’s get ahead of it.”
This model of collective intelligence as qualitative data for proactive problem solving works across sectors. Public health providers in clinics may ask about a specific drug interaction or symptom cluster. AI could surface the spike in these provider questions before case counts show up in reportable disease data. This enables health officials to pre-position resources, issue guidance, and alert other providers in the area.
An emergency management field coordinator might ask about shelter capacity, generator fuel, and potential road closures in a specific region. AI can flag the concentration of questions as an early signal that in an emergency, resources might be strapped. Emergency managers can shift resources before an emergency, before formal requests come up the chain of command.
A SHIFT LEFT
This operating model shifts problem-solving in government agencies to the left, to occur earlier, potentially even before problems arise. This is instead of reactively meeting challenges once they’re crises. When AI captures qualitative signals at scale, government operations reorient towards anticipatory governance. Operations close the gap between ground truth and leadership awareness. And ultimately, these agencies will serve their communities most effectively.
Madeleine Smith is cofounder and CEO of Civic Roundtable.
