The following is adapted from a speech delivered on February 15, 2017, at a Hillsdale College National Leadership Seminar in Phoenix, Arizona.

The performance of our country’s intelligence service is the latest example of an issue exploding into the headlines and becoming a shouting match, while failing to clarify anything about the issue itself. This explosion was ignited last fall by allegations that the Russians hacked into Hillary Clinton’s campaign to help Donald Trump win the election. The blast radius expanded after the election, when rumors surfaced that the Russians had deployed their nasty tactic of kompromat to undermine President Trump’s credibility by spreading rumors about his private behavior while in Moscow years ago. All this, on top of failures that had already wreaked havoc at the CIA and our other intelligence agencies—the 9/11 attacks themselves, the mess over weapons of mass destruction in Iraq, the weird 2007 National Intelligence Estimate whose key judgment was that Iran had abandoned its nuclear bomb program, Edward Snowden’s NSA espionage activities—has kept the issue of our intelligence service in the headlines.

But before addressing the question of why these failures have occurred, we need to define clearly the role and purpose of our country’s intelligence service, with a focus on how intelligence really works when it’s working properly.

Just utter the word “intelligence” and most people conjure up images of spies, secret satellites peering down on foreign cities and terrorist camps, and rooms full of young technocrats reading private emails and listening to private conversations. These images are accurate, but they reflect the tools and techniques of our intelligence service, rather than its purpose. To understand its purpose, think of a jumbo jet flying at night through turbulent skies—thunder clouds, lightning, other airplanes streaking in all directions and at all altitudes. To navigate through this, the pilot and his crew rely on their radar—the instrument that paints a picture of their environment, enabling them to see what’s going on around them and what lies ahead so they can chart a safe course. Radar doesn’t tell the captain and his crew what to do, but it gives them the accurate information they’ll need to make good decisions.

Our intelligence service is our nation’s radar. Its purpose is to provide the president and his national security team with an accurate picture of what’s going on in the world and what’s likely to happen in the days, months, and years ahead. The assumption is that if the president and his team have this information, they can chart a safe course for our country. And if they can see the distant future soon enough and clearly enough—and if they don’t like what they see—they can take steps to change the future before it happens.

Good intelligence is a combination of information and insight. Information is the raw material, while insight is the finished product. Sometimes this insight takes the form of a top secret report that alerts the president and his team to something that’s about to happen, such as a terrorist attack or the military invasion of one country by another. At other times it is a National Intelligence Estimate, whose purpose is to provide an overall assessment of a major issue—such as North Korea’s nuclear bomb program or the rapid growth of Africa’s middle class—along with a prediction of its future course.

The key to producing good intelligence lies in getting this combination of information and insight right. Intelligence work is like science. You don’t collect information randomly and then stare at it in hopes that something important will pop up. You start with a thesis—in other words, you decide what you want to know. Then you send your collectors out to get it. This is why the key to producing good intelligence lies in asking the right question, rather than in just poring over what’s been randomly collected in hopes that somewhere in the pile of reports and intercepts on your desk you’ll spot something important.

Let me give you an example of how this worked during the Reagan administration. From the end of World War II until 1981, every president’s objective had been not to lose the Cold War. If things were no worse when a president left office than when he took office, he’d done a good job. But President Reagan didn’t want to tread water—he wanted to win the Cold War. In other words, he switched from defense to offense. So Reagan’s great director of Central Intelligence, William Casey, asked the CIA’s Soviet Division two obvious questions: Where is the Soviet Union weak? and Where is it most vulnerable? The answer he received was: We don’t know. No one’s ever asked this before. Our spies had been so focused on Soviet strengths—infantry divisions, nuclear missiles, tanks, submarines, and so forth—that we had no intelligence on Soviet weaknesses, such as its imploding economy. Under Casey’s leadership, we refocused our collection efforts and, not surprisingly, found all sorts of Soviet vulnerabilities that hadn’t been grasped because no one had bothered looking for them. President Reagan used these weaknesses and vulnerabilities to put more and more pressure on the Kremlin. Eight years later the Berlin Wall came down, and two years after that the Soviet Union ceased to exist.

In the intelligence business, just as in scientific research, a thesis sometimes turns out to be wrong. The collectors can’t find what you want, because it isn’t there. When this happens—and it happens to even the best scientists and intelligence officials—you must abandon your flawed thesis and re-think the issue. If you refuse to do this, you’re like a scientist who continues to insist that the Earth is flat—or a president who continues to insist that ISIS is like a “junior varsity” team.

***

When the collectors have done their work—when they’ve told the analysts what they want to know—the intelligence process shifts from gathering information to creating insight. It’s the difference between shopping for food in the supermarket and actually cooking dinner.

Insight is the product of knowledge, experience, and, above all, good judgment. You cannot say something insightful, or even something intelligent, on a subject or issue about which you don’t know anything. So the most senior intelligence analysts must be among the world’s most knowledgeable individuals in their fields of expertise—the Mideast, Russia, China, nuclear weapons, economic development, etc. And they must have that one elusive and unquantifiable skill that so often brings success in every venture: the ability to spot a pattern with the fewest possible facts—the ability to look at what’s known and combine this with their own knowledge, experience, and good judgment, to come up with a new idea or insight. This is the skill we see in great scientists like Albert Einstein, in great entrepreneurs like Steve Jobs, and in great intelligence chiefs like Bill Casey.

Back in January, when U.S. intelligence chiefs released an unclassified version of the briefing they gave to President-Elect Trump about Russian efforts to influence the November election, Americans learned a phrase that’s unique to the world of intelligence: key judgment. It was a key judgment that Russia had hacked into John Podesta’s email server, and a key judgment that Vladimir Putin preferred Donald Trump to Hillary Clinton. Since these key judgments understandably erupted into a nasty political brawl, let’s take a moment to understand what a key judgment really is. Simply put, it’s the conclusion reached by our most senior intelligence officials, based not only on the evidence they were able to collect, but also on the insights it enabled them to reach based on their knowledge and experience.

A key judgment isn’t the same as a jury verdict. A jury verdict is based solely on the evidence presented to it. In a murder trial, unless the prosecutors can prove beyond a reasonable doubt that the defendant is guilty, you must vote for acquittal. But in a National Intelligence Estimate, you reach a key judgment by starting with the evidence, then combining it with your own knowledge and experience to reach a conclusion.

Precisely because key judgments go beyond evidence, an intelligence service must be trusted by policymakers to be effective. Policymakers may not always like what they’re told—as when the obvious implication of a National Intelligence Estimate is that a favorite policy is heading for catastrophe—but if they trust the intelligence service, they will know that what they’ve been told is likely true. And this trust needs to be earned.

This is how it was during the Reagan administration, because ev­eryone from the President on down knew perfectly well that the intelligence official who not only had read the final version of an Estimate and signed off on it—but also played a major role in writing it—was the CIA director himself. Like every other member of the cabinet, Bill Casey was a busy man. But to Casey, being in charge of our intelligence service meant more than merely being its top administrator and dealing with budgets and bureaucracies. It meant that he himself was our country’s top intelligence analyst. When the final draft of an Estimate landed on his desk—more precisely, when I walked into his office and handed it to him—Casey would take that draft, pick up a pen and a yellow legal pad, and go through it word by word.

Sometimes he made a change that clarified a sentence. Other times he asked a question that forced us to go back and re-think what we’d written. When that happened, we either changed the draft or asked to meet with Casey to try and persuade him that the original version was better. He would listen and then make his decision. All of us who worked closely with Bill Casey—he insisted that everyone, including the CIA’s most junior analysts, call him Bill—were astounded by the amount of time he devoted to getting the final draft of an Estimate, or the final version of the President’s Daily Brief, just right. He did this by sitting quietly in his office, reading, writing, and—something that so few officials in Washington, D.C. set aside the time to do—thinking.

***

So why has our intelligence service suffered so many failures during the last decade or so, losing the trust of so many? Because it’s been run by career bureaucrats and administrators who rose to the top by managing intelligence rather than actually doing it. That’s like putting an airline executive with an MBA and a law degree into the cockpit of a jumbo jet. And like bureaucrats and administrators everywhere, our recent intelligence chiefs focused on structure rather than on people. Of course all organizations, including intelligence services, need the proper structure. But especially in an intelligence service, good structure is worthless without the right people—in this case world-class analysts who are deeply knowledgeable about the Mideast, China, Russia, terrorism, and all the rest. Make a list of our country’s leading experts on these subjects. How many of them have held top-level jobs in our intelligence service during the last dozen or so years? How often have the leaders of our intelligence service reached out to these people to seek their advice? The correct answers are: none and rarely.

We are still in the early days of the Trump administration, but to borrow an overused Washington cliché, we should be cautiously optimistic about the future of our intelligence service. Neither Director of National Intelligence Dan Coats nor Director of Central Intelligence Mike Pompeo are professional bureaucrats. They’ve built their careers on substance rather than on management. Each of them has proven he can talk about the key issues that confront us with an impressive level of personal knowledge and insight. Each is capable of actually doing intelligence rather than merely overseeing it.

This will require restoring the correct balance between collection and analysis. Obviously, collecting information is crucially important work. Collecting information through technology—satellites, intercepts, and so forth—is intense to the point of exhaustion. Collecting information through espionage is dangerous and sometimes fatal. All of us owe these collectors a huge debt of gratitude. What they need now is guidance from the top—a clear sense of what to look for, rather than just being told to sweep in whatever information they can in hopes it will prove useful.

Turning this raw material into first-rate intelligence will require the active participation of our country’s best geo-strategic experts in think tanks, universities, corporations, and increasingly the blogosphere. Directors Coats and Pompeo should recruit the ones they can, and be in close touch with the others. This doesn’t mean agreeing with everything these experts say and write. It means listening to them and blending their information and insights with what’s been gathered covertly, in order to reach the clearest, most accurate conclusions about what’s happening now and what’s likely to happen in the future.

Finally, Coats and Pompeo will need to do the one thing their recent predecessors didn’t do, either because they didn’t recognize the need to do it or didn’t have the ability. They will need to set aside time—quite a bit of time—to sit quietly in their offices and think. Their objective must be to paint an accurate picture of what’s going on in the world and of what’s likely to happen in the future. If they can do this, President Trump and his national security team will have what they need to see America safely through today’s global turbulence: radar.