Soldiers clear a roadblock set by demonstrators in the Cibitoke neighborhood of Bujumbura, Burundi, on May 22. (Jerome Delay/AP)

Makeshift roadblocks dot the capital of the tiny central African state of Burundi. The country of 10 million people is once again in the grips of crisis that has the potential to take the lives on a scale that last engulfed the country just over 40 years ago. Then, rival Hutu and Tutsi ethnic militias unleashed a swell of genocidal violence in a civil war that took as many as 300,000 lives.

This time, however, ethnicity isn’t driving violence. It’s political. President Pierre Nkurunziza’s late April announcement that he would seek a third term in office in the upcoming June 26 presidential vote — violating not only the constitution but the very peace accords that brought an end to the country’s last violent episode and still keep the fragile peace — has sent the country into convulsions.

Since late April, pro-democracy protesters have jammed the streets demanding that the Nkurunziza reverse his decision—prompting, in turn, the president and his supporters to ratchet up and dig in, firing on protesters, jailing opposition leaders, and banning independent media outfits. Youth militia supportive of Nkurunziza’s ruling CNDD-FDD have intimidated and harassed civilians in Bujumbura and further upcountry. Fearing new violence, more than 100,000 Burundians have fled to neighboring Rwanda, Tanzania and Congo in the past two months.

Events since late April seem to confirm these fears, as Nkurunziza’s allies in the Burundian police, military and intelligence services have continued, violently, to strengthen their grip on power. The hunt to bring plotters and supporters of a short-lived May coup attempt to “justice” has begun, and the country is rife with fears that Burundian civilians could suffer mass atrocities as a consequence of the paranoia of the Nkurunziza regime.

The scenario feels ominously familiar. In another context, we might be lamenting our collective inability to anticipate what seems eminently foreseeable. But in this circumstance, the U.S. government actually did anticipate this scenario and the potential for group-targeted violence — and have quietly been planning for it in a little known effort to avoid it entirely. The current Washington response to the crisis in Burundi tentatively confirms what the bipartisan Genocide Prevention Task Force suggested in 2008, and what atrocity prevention scholars have stated for years: The earlier we can recognize signs of future mass atrocity risk, the better prepared we will be for preventive action.

In a little heralded trip to Burundi’s capital, Bujumbura, in April 2014, prophetically only the day after neighboring Rwanda marked the 20th anniversary of its genocide, U.S. Ambassador to the United Nations Samantha Power — a principal architect of the Obama administration’s atrocity prevention efforts — planted the seeds of a diplomatic effort by urging “the government and the opposition to reject violence and to engage in the political process in a civil and non-confrontational manner.”

In an address this March at the Council on Foreign Relations, the State Department’s undersecretary for civilian security, Sarah Sewall, shed greater light on how those diplomatic messages has have been backed up by action, outlining how more than 18 months ago, “the Department’s atrocities watchers grew very concerned about escalating tensions in Burundi,” and initiated a review process by the President’s Atrocities Prevention Board, the interagency body charged not with crisis response, but — for the first time within civilian agencies — early warning and preventive action to avoid atrocities.

That months long review included a team of conflict experts from across the U.S. government traveling to Burundi to conduct a thorough “analysis of the potential risks of violence … leading to a diplomatic engagement and programmatic strategy” to understand the likely drivers of new conflict and strengthen Burundi’s — and our own — infrastructure for preventing it.

Through a meager $7 million program, that has meant engaging leaders at the village, local and national levels to facilitate dialogue, build resiliency and implement activities that could detect and deescalate tensions leading to violence. It’s also meant having a conflict adviser remain on the ground in the U.S. Embassy to closely monitor local airwaves, political statements and religious gatherings for the early warning indicators of impending conflict.

But while promoting dialogue and monitoring tensions are necessary, they are by no means sufficient. Were more flexible funds available other programmatic efforts could have been employed to more proactively defuse tensions and promote resiliency, for example: programs to actively counter hate speech and vile political rhetoric using new Internet and SMS technologies, ensure non-state media’s voice in society via external broadcasting from neighboring states and strengthening pro-democratic forces’ ability to organize with secure Internet platforms — all measures used to varying degrees of success in what might be deemed higher priority cases like Kenya or Iran. Even more could still be done.

In other areas it has been. In the days leading up to the Nkurunziza’s decision last month, new public warnings were issued, with, importantly, a preview of possible punitive measures to come with threats from U.S. officials to sanction those responsible for violence and deny further security assistance before even one civilian had became the target of government attacks. Already, visa bans are in place against known instigators within the government, and U.S. support to new Burundian forces to join a peacekeeping operation in Somalia — a substantial source of prestige and military funding — have been put on hold indefinitely.

Programs and diplomacy can look like one-offs if they are not themselves the products of a well-grounded policy process. Here too there are signs of important policy reforms that if properly resourced could well survive the Obama administration and represent a game-changer in how the United States thinks about its role and ability to prevent atrocities. Last month’s release of the State Department’s second Quadrennial Diplomacy and Development Review (QDDR) included for the first time an entire section devoted to emerging efforts to “Strengthen our ability to prevent and respond to internal conflict, atrocities, and fragility … through an increase in use of early warning analysis to drive early action.”

But as frightened civilians continue to run for Burundi’s exits and reports emerge of mounting targeted killings of presumed coup plotters, with last month’s assassination of Burundi’s main opposition leader Zedi Feruzi a foreboding example, it would be premature — and painfully out of sync with outcomes on the ground — to trumpet too triumphant a tone.

Even if Burundi walks back from the brink by avoiding violent reprisals and smoothing its way to a political transition, how can we demonstrate that this nascent network of early warning and preventive action materially contributed to a less bloody outcome? Establishing causal links in development and diplomacy is hard enough; proving a counter-factual is even harder. But though the empirical evidence in support of preventive action is at best debated and at worst absent, we can demonstrate some of its dividends: societies become more resilient against future risks of mass violence, the international community is better prepared to confront new risks of mass atrocities, and we build a more responsive network of global institutions.

Perhaps a more important question when attempting to measure the value of prevention is to instead examine the costs of doing nothing. Sadly on this score, we know all too well the impact of not heeding the warning signs of atrocities. Recent anniversaries in Cambodia, Rwanda and Srebrenica stand as annual reminders of the consequences of our collective failure to anticipate and to act. If the slow emergence of an atrocity prevention doctrine suggests that we have learned from these failures, ongoing threats to civilians in Syria and South Sudan reiterate how far we have left to go.

By carving out even minimal financial and human resources from more immediate crisis response, where the political and financial costs are the highest, we create opportunities for upstream prevention efforts, where the tools available are more varied and their costs greatly reduced.

Today, that proposed resource shift represents a wager that we have learned from past mistakes and can act in more effective ways to avoid the worst outcomes. That has always been the promise of Never Again and it’s a gamble we have been far too unwilling and unable to make until now.

Cameron Hudson is the director of the Simon-Skjodt Center for the Prevention of Genocide at the United States Holocaust Memorial Museum.