Part of Foresite's The Mission Endures series, marking Veterans Day.
In the field, every operator knew the commander's intent: what we were there to do and why. Plans fall apart the minute you step off, so the people closest to the problem have to know the goal well enough to adapt without waiting for permission.
I run the SOC the same way. Our analysts don't just work a queue. They know what we're protecting and what a bad day looks like for each customer. When something strange hits at 3 AM, the person on shift shouldn't need to wake me up to know what matters. They should already know, and they're trusted to act on it.
That doesn't happen by accident. Running a security operations department means answering for outcomes twice: to customers, for what happens in their environments, and to my own people, for whether they have what they need to do the job. At any real scale I can't sit on every bridge call or look at every alert. I shouldn't be the only person who can make the call, either. So the work of leading a SOC is building the conditions where good calls get made without me in the room.
A plan gives people steps for the situation you expected. Adversaries rarely cooperate. Intent is what an analyst falls back on when the playbook runs out: what are we protecting, what does failure look like, and how much risk are we willing to take to stop it.
In a SOC, intent can't live only in the head of whoever runs it. It has to be written down, customer by customer: what matters most in that environment, what gets contained right away, what gets watched, and where the line sits before you take a business system offline to stop something moving fast. Leave that line vague and good analysts hesitate, because they're guessing what you would want.
Writing it down once isn't enough. Intent goes stale. Threats change, customer environments grow, and the document from last spring stops matching the network in front of you. The useful review is simple: look back at recent incidents and find the moments where someone had to guess. Each guess is a gap in the intent, and closing it is leadership work, not analyst work.
Then you repeat it. You say the same things in team meetings until people are tired of hearing them, because that's when they start using them.
Recon taught me that a handful of people who trust each other completely will outperform a much bigger group that doesn't. You know each other's strengths, you cover each other's gaps, and nobody hides a mistake, because the whole team pays for it.
That's why our department is built as focused teams instead of one big org. Each team owns its piece end to end, and they lean on each other. When Threat Hunting turns a finding into a new detection, the SOC is the team that lives with it, and they say so quickly if it's noisy. That loop only works if both sides know each other and assume good intent.
Trust is also what lets someone act at 3 AM, and it runs both directions. I have to earn my analysts' trust as much as they earn mine. The fastest way to earn it is to back a sound call made under pressure. If an analyst sees something spreading and isolates a server in the middle of the night, they need to know I'll stand behind that decision with the customer in the morning, even if hindsight shows a softer option.
Second-guess that call in public, or punish someone for acting on the best information they had, and the next time they will wait for permission. Waiting is expensive during an active incident.
My job is less about giving orders and more about making sure every team has what it needs and knows the others have its back.
Hours of preparation, studying the ground, checking and rechecking, for a moment that may last seconds. You learn that the outcome is decided long before the shot, and that panic is the most expensive thing you can bring to a bad situation.
Security works the same way. The incident you handle well is the one you prepared for months ago: the runbook, the drill, the detection someone tuned on a quiet Tuesday. That boring work is what the team falls back on when the real thing arrives.
When an incident does hit, the tone on the bridge call comes from the top. If I'm calm, the team is calm, and calm people make better decisions. I treat that as part of the job, not a personality trait. If the leader on the call is loud or demanding answers nobody has yet, engineers stop thinking about the adversary and start managing the room.
So I keep the call to a few questions. What do we know is true? What are we doing about what we don't know? When is the next update? Set a cadence and people can put their heads down and work between updates instead of reporting every five minutes.
It's also why we've leaned so hard into automation and AI on our SOC floor. Let the machines handle the volume, so our people have the time and focus for the calls that actually need a human.
Every mission ended with a debrief. What happened, what was supposed to happen, why those were different, and what we change next time. Rank didn't matter in that room and neither did ego. The point was to get better, not to find someone to blame.
We do the same after every significant incident: a blameless review, an honest timeline, and real follow-through on the fixes. If people are afraid to admit what went wrong, you never find out what actually happened, and you see the same failure again.
Blameless doesn't mean nobody is accountable. It changes the question. If an analyst missed an alert, the useful question isn't why that person failed. It's what made the alert easy to miss: the missing context, the noisy rule, the screen that buried it. Fix that and you've fixed it for everyone on every shift.
The same honesty applies to our tools. In July, about 12% of our triage agent's first verdicts came back "unknown," and each one went to a named analyst. I'd rather see that number than a tool that pretends to be sure.
The real test of a debrief is the follow-through. Every fix that comes out of a review gets an owner and a date, and I want to see the open ones until they close. You know the culture is working when you stop having the same incident twice.
That brings it back to the analyst at 3 AM. They act without waking me up because the intent is written down, the team has their back, the preparation is done, and the last review fixed what tripped someone up before. That's what commander's intent looks like in a SOC.
Have you served, or are you serving now? If you’re a Foresite customer or partner, we’d be honored to send you a Mission Endures challenge coin.
Available while supplies last.