Jim Bourg / Reuters

Smoking guns are the stuff of spy movies. In real-life intelligence-gathering, they are exceptionally rare. That’s why the business of intelligence typically requires collecting and analyzing fragments of information—putting together secret nuggets with unclassified information—to try to make sense of complex reality. If nothing else, the whistle-blower who filed a complaint against President Donald Trump clearly followed his or her training.

I’ve spent 20 years reading intelligence reports and researching the U.S. intelligence community. And I’m not automatically inclined to believe the worst allegations about any administration; everyone has agendas and incentives to reveal information, some more noble than others. Trump and his allies have dismissed the complaint as hearsay and accused the whistle-blower of acting on political motives. But a close reading of the whistle-blower’s lengthy complaint, which accuses Trump of “using the power of his office to solicit interference from a foreign country in the 2020 U.S. election,” yields a lot of concrete leads for investigators to follow.

Here are three things I learned:

The whistle-blower sounds like an experienced intelligence analyst, and the complaint appears credible. Because intelligence officials are trained to be oh-so-careful with what they say and what they leave out, reading their reports is something of an art. You have to ask: Is the author clear about where the information came from, and careful about protecting confidential sources and collection methods? Does the information come from a single source or multiple sources, and what context is provided to assess their credibility? Does the overall tone suggest the author is carrying a vendetta, cherry-picking facts, discarding opposing evidence, or going further than the evidence warrants? Were other, more benign explanations considered for a given fact pattern?

On all of these counts, the whistle-blower complaint comes across as highly credible. The writer expressly acknowledges the limits of his or her knowledge and also measures contemporaneous accounts from public sources against the private information reportedly conveyed. The full complaint also carefully points out which information was reported in the press, including by the Ukrainian government; which information came from private sources; and what information was potentially classified. The whistle-blower is also careful to delineate what information is known and what is still not known—such as why President Trump allegedly directed that nearly $400 million in military assistance to Ukraine be suspended in July.

Notably, the complaint is not based on direct knowledge. Secondhand information is, of course, less reliable than firsthand information—as Trump’s defenders were eager to point out. But intelligence is frequently secondhand. If intelligence reports included information only from those who were literally in the room when the nuclear weapon was made, there wouldn’t be many intelligence reports.

Importantly, the whistle-blower makes clear up front, on page 1 of the complaint, “I was not a direct witness to the events described.” There is no hiding the ball. The whistle-blower then says, “I found my colleagues’ accounts of these events to be credible because, in almost all cases, multiple officials recounted fact patterns that were consistent with one another. In addition, a variety of information consistent with these private accounts has been reported publicly.”

The most important word here is multiple—which, according to the writer, means “more than half a dozen U.S. officials.” And the word appears in relation to each of the most serious claims in the complaint. Multiple officials allegedly told the whistle-blower about: the July 25 call between Trump and Ukrainian President Volodymyr Zelensky; the disturbing transfer of the verbatim transcript to a classified system in an alleged effort to “lock down” all records of the phone call; contacts between Trump’s personal attorney, Rudolph Giuliani, and Ukrainian officials that reportedly circumvented the national-security decision-making processes; the communication to Ukrainian leaders that a meeting or phone call between Trump and his Ukrainian counterpart depended on Zelensky’s willingness to “play ball” on the issues Giuliani had publicly raised involving investigations that would help Trump in his 2020 reelection bid; and the instruction to suspend military assistance to Ukraine, which they said had come directly from the president.

We’re not talking about relaying rumors by the lone office gossip or low-level bureaucrats on the fringes of the decision-making apparatus. The multiple sources in the complaint are all said to be U.S. officials, many with “direct knowledge” of the events described, including “White House officials.”

Importantly, Michael Atkinson, the inspector general for the intelligence community, conducted a preliminary review of the whistle-blower’s complaint and, in his letter to acting Director of National Intelligence Joseph Maguire, noted multiple corroborating sources as well. In fact, though Atkinson said he had reason to believe the whistle-blower may have “political bias” and favor a rival political candidate, the complaint nonetheless “appears credible.” Why? Because of “the other information” Atkinson uncovered during the preliminary review.

A word-for-word transcript of the July 25 call between President Trump and Ukrainian President Zelensky exists. “In the days following the phone call,” the whistle-blower says, “I learned from multiple U.S. officials that senior White House officials had intervened to ‘lock down’ all records of the phone call, especially the official word-for-word transcript of the call that was produced—as is customary—by the White House Situation Room.”

The intelligence community’s inspector general wants that word-for-word transcript. On August 26, the same day that Atkinson sent the whistle-blower complaint to Maguire, he sent a “document hold notice” to the White House counsel requesting that all relevant records regarding the July 25 phone call, as well as “alleged related efforts to solicit, obtain, or receive assistance from foreign nationals in Ukraine, directly or indirectly, in connection with a Federal election,” be preserved, and that his office be given access to them as required by law.

In the coming days, it will be important to learn whether this actual verbatim transcript of the July 25 call still exists. If it does, will Congress or Atkinson’s office get it? And why would the White House release a summary rather than a redacted version of the verbatim transcript unless the verbatim transcript provides information more damaging to the administration?

Classified systems may have been misused. The whistle-blower claims that White House officials were “directed” by White House lawyers to remove the July 25 verbatim transcript from the computer system where transcripts of calls are typically stored. Instead, it was “loaded into a separate electronic system that is otherwise used to store and handle classified information of an especially sensitive nature.” If true, this is a much bigger deal than it may seem. Using classification to protect information that’s embarrassing or politically sensitive is against the law. Looking ahead, the congressional intelligence committees are likely to examine whether the alleged misuse of classified systems implicates individuals in criminal violations, and whether the alleged misuse involves additional documents.

In the face of these revelations, the temptation to exaggerate and hyperventilate—to make sweeping claims that go beyond what the evidence supports—is strong and growing. Old-school analysis is more crucial than ever. The congressional intelligence committees need to focus on the fundamentals, asking: What exactly do we know? How credible is it? What don’t we know? And how can we fill in the gaps?

We want to hear what you think about this article. Submit a letter to the editor or write to letters@theatlantic.com.