The news media is abuzz with reports of the ongoing public confrontation between President-elect Trump and members of the intelligence community, particularly outgoing CIA Director Brennan. Alarms are being sounded over the danger posed by an incoming president in confrontation with his own intelligence agencies, and who publicly questions the value of the intelligence they provide. Dire predictions are being made about the implications of this widening schism and the price we may pay down the road.

Many of the questions posed are without doubt legitimate. A president and his intelligence advisors must work hand-in-glove, and it is imperative that a president have confidence in the sensitive intelligence he is provided. Lost in all the back and forth of Twitter exchanges and the president’s daily brief, however, is one inescapable and critical truth: The quality of the intelligence we are producing, particularly our human intelligence, is not what it must be, and it has not been so for a very long time.

This is not a reflection on the men and women of our intelligence agencies, particularly those in my former organization, the CIA. They work long hours under often dangerous conditions for mediocre pay and little or no recognition. They perform miracles everyday. Nonetheless, due to systemic and bureaucratic factors outside their control, they all too often come up short and we operate without the vital inside information we so desperately need.

Out there on the horizon, dark clouds are gathering. The Chinese, who have seized the entire South China Sea and waged economic war on us for decades, are girding to resist any efforts on our part to redress the strategic balance. Russia, emboldened and surging into every vacuum we leave in the Middle East and Eastern Europe, will not quietly climb back in its box and allow us to reassert our primacy. Iran, flush with cash and sensing a chance to dominate the entire Middle East, has no intention of squandering this opportunity or backing down. Across the planet, terrorist groups with apocalyptic world views are acting with the certain knowledge that these are the final days before the Mahdi comes and final victory is achieved.

The way ahead as we attempt to recover from the unprecedented foreign policy debacle brought on by the Obama administration is perilous. Only with the aid of high-quality human intelligence will we navigate it successfully.

To do so, the Trump administration must move quickly and decisively to address the factors that are strangling our espionage capabilities. This is no time for half measures, nor is it time for us to continue the failed policies of the past, throwing people, money and process at a business that is fundamentally about none of those.

Eliminate the office of the Director of National Intelligence. This was a bad idea when first conceived. It has been made infinitely worse by the ballooning size of what is now a gargantuan bureaucratic edifice sitting above all the other intelligence agencies, adding nothing but paperwork and micromanagement. Transfer the responsibility for overseeing the intelligence community back to the CIA. It is called the Central Intelligence Agency for a reason.

Roll back the “reforms” Director Brennan instituted at the CIA. I have no doubt Director Brennan is a dedicated public servant and acted in good faith. He is also a career, D.C.-based bureaucrat with no idea whatsoever about how operations are run in the field or how the business of espionage really works. All of his efforts to draw new lines on wiring diagrams and shuffle desks at headquarters have contributed nothing. The CIA must be a field-centric organization in which headquarters answers the mail, shovels support forward, and otherwise gets out of the way.

Stop trying to recreate the CIA’s capabilities elsewhere in the intelligence community. Since 9/11, everybody and their brother has been attempting to play in the business of human intelligence. We now have thousands of additional “case officers” clogging the system. Most of them contribute very, very little, but they suck up a lot of money and resources.

We don’t need 10 different CIAs. We need one that gets the job done. The DOD has plenty of work to do on its own. Get it out of the business of providing strategic human intelligence.

Face the fact that 9/11 was fundamentally a failure of collection. Prior to the attacks in New York and Washington, we had already seen al-Qaeda attack two of our embassies and come perilously close to sinking a U.S. warship. Despite that, we did not have a single source inside that organization to warn us about an impending attack, years in the making, involving 20 separate operatives.

We did not have the sources we needed, because the structure of our human intelligence apparatus, overwhelmingly dependent on officers under official cover and reporting from friendly liaison services, was not capable of crawling into the belly of the beast. A machine built for the Cold War was not able to put spies next to Osama bin Laden or inside his worldwide jihadi organization.

Since 9/11 we have changed some things. We have embraced the necessity for lethal action, which we had effectively walked away from in the years between Vietnam and 2001. We have surged officers into remote areas in the Middle East and South Asia. We have given the CIA more money and more people.

We need to recruit a much more diverse set of personnel. We will still need a certain number of native-born Americans straight out of college or graduate school. Increasingly, though, we need a much more varied pool of recruits. We need former special operations personnel. We need international bankers. We need people whose families came over on the Mayflower. We need people who were born in Beirut, speak native Arabic, and moved here in their twenties.

We need to train these individuals for the world into which we are sending them. Once upon a time the center of gravity for the CIA was West Berlin, and espionage was a gentleman’s game in which the worst that happened was you got sent home short of tour. Today, the center of gravity for the CIA is somewhere in the shattered, failed states of the world, wherein a failed op means death by suicide vest or hideous, prolonged torture at the hands of madmen. CIA case officers need to be selected and trained from the day they are brought on board to operate and excel in this kind of environment.

We need to cultivate and promote leaders. The CIA, which began in 1947 as the natural outgrowth of the Office of Strategic Services (OSS), has over the years calcified into an organization increasingly focused on risk avoidance and internal politics. One of the many results of that transformation has been the rise of “yes men” and bureaucrats to the top ranks of what should be a dynamic, agile, and daring organization. The reforms that must be made require that we put leaders and risk-takers back in the top ranks and send those people looking for a safe, secure federal job elsewhere.

We need to accept risk. No operation to put a source inside a vicious terrorist organization planning attacks on nuclear plants in Western Europe can be conducted without accepting the risk of failure—failure in this case likely means the death of the operators involved. This is no excuse for sloppiness or poor planning. This is a recognition that with the potential for great gains always goes great risk.

If you cannot accept this, you should not be playing at the game of spying. If you want perfect security, shut down, stay inside, and accept that you will achieve nothing. An operation without risk is not worth running.

We need to demand results. The CIA exists for one reason: to provide the critical human intelligence that no one else can. We cannot accept anything less.