After 30 years in this industry, one lesson has become impossible to ignore: the problem is not always the warning sign. Often, it is the gap between seeing a concern and doing something meaningful with it before the situation becomes harder to interrupt.

In targeted violence prevention, we talk about “missed warning signs” as if no one saw anything coming. Sometimes that is true. In other cases, someone noticed a behavioral change, heard a concerning comment, or became worried about an escalating grievance. Someone felt that a situation was moving in the wrong direction, even if they could not fully explain why.

In those moments, the concern existed. The real question is what happened next.

That is where prevention opportunities can be lost. Concern is minimized because there was no direct threat. Behavior is explained away as “just venting,” “just a personnel issue,” or “not serious enough yet.” Leaders wait for more proof before initiating a reasonable review. Information remains scattered across conversations, departments, files, inboxes, and memories. Each person may hold only part of the account, and no one is responsible for examining how those accounts fit together.

That is the cost of waiting. But the issue is not simply that an organization paused. There are times when patience, restraint, and caution are necessary. Not every concerning behavior indicates a threat. Not every person in distress is dangerous. Not every unusual comment, emotional reaction, or workplace conflict should become a security matter. A mature prevention system must avoid panic, overreach, and unfair labeling.

The difference between restraint and passive waiting

The deeper problem is passive waiting. It happens when an organization notices concern but does not assess it, hears something troubling but does not document it, or delays a decision without establishing a plan. Information stays with whoever first received it. Nobody is assigned to clarify the situation, coordinate the review, or determine what would require a different response.

Passive waiting can feel responsible in the moment. Leaders do not want to overreact or stigmatize someone who may be struggling. They do not want to damage a reputation, create unnecessary conflict, or escalate an allegation they cannot yet substantiate. Those concerns are valid. They should shape a fair assessment process, rather than prevent one from beginning.

Measured restraint looks different. A team may decide that immediate escalation is not warranted while still documenting the available information, identifying unanswered questions, assigning follow-up, and setting conditions for reassessment. The decision to wait has a reason, an owner, and a review point. It is a considered response that can change as the circumstances change.

That distinction matters because prevention does not require certainty before every reasonable step. It requires purposeful, proportionate movement. The next step may be quiet and supportive: preserving an account, clarifying the context of a grievance, checking relevant information through authorized channels, or connecting someone with appropriate support. Where the circumstances warrant it, qualified reviewers may also need to examine access, protective factors, potential targets, and intervention options within their responsibilities and applicable requirements.

None of that requires panic. It requires process. An immediate safety concern calls for the organization’s emergency response, rather than waiting for a routine team meeting or a completed case summary.

Seeing and saying need a dependable response

A warning sign is not a conviction, a diagnosis, or proof that violence will occur. It is information that may warrant structured attention in context. Threat assessment should help the team understand the situation, test explanations, coordinate an appropriate response, and revisit its decisions. It cannot predict individual behavior with certainty or guarantee that every harmful outcome will be prevented.

Organizations that take this work seriously build a capability around it. People know how to report a concern. The people receiving reports know how to preserve and route them. Authorized team members can examine relevant information together, distinguish observation from interpretation, and decide who is responsible for what happens next.

After three decades in security and violence prevention work, I am convinced that teaching people to “see something, say something” is only part of the task. Seeing and saying matter, but the real test is whether the organization can do something meaningful, timely, and proportionate with what was seen and said.

That is also where assistive intelligence can support the work. Threatalytics AI can help organize information supplied for review, bring decision-relevant gaps into view, and support documentation of the team’s reasoning. It cannot make an unmonitored reporting channel effective or take responsibility for the response. People still have to receive the concern, examine the information, and carry out the agreed actions.

Warning signs do not prevent violence on their own. Responses create opportunities to reduce harm. The cost of waiting is not that an organization took time to be fair or careful. It is that concern remained unexamined, coordination never began, and nobody established a plan for what should happen next.

A pause can be responsible. A pause without assessment, ownership, or a way to recognize change leaves the organization depending on hope when it needs a process.

Does this sound familiar?

Have you heard, “Let’s wait and see”?

Have you seen a concern put on hold without a clear owner, follow-up date, or plan for recognizing change? The difficult part may be knowing how to move forward without overreacting.

We welcome a practical conversation about your review process and whether Threatalytics AI could help your team organize concerns, clarify unanswered questions, and document its reasoning.

Talk Through Your Challenge
Explore the reporting gap

About this essay

Adapted from Jason Destein’s April 28 essay. This version was published on October 29, 2026.