Industry Thought

Preparedness That Keeps Up With Reality

Practice is the flywheel of preparedness. Preppr makes it fast, and keeps it grounded in what’s true about your place.

What kept bothering me is that preparedness is supposed to be continuous: a living thing that keeps up with a changing place. Hazards shift, capabilities change, and people move on. A plan written for last year’s reality ages on the shelf the day it’s approved.

What keeps preparedness alive is practice, meaning exercises and simulations. Practice is how a team finds out whether the plan still matches the place, and it’s the flywheel that turns findings into better plans. But that flywheel has always turned slowly. Exercises and simulations were slow to design and slow to run, and by the time you’d designed and run one, the reality had moved.

Speed was the first problem

So the first problem was speed: design and run exercises and simulations fast enough to keep up with the place they’re meant to test.

We solved it. Preppr designs and runs exercises and simulations in a fraction of the time they used to take. Producing the content isn’t the hard part anymore.

That shift changes what a program can do. When an exercise no longer takes months of staff time, you can practice more often, practice more of the plan, and come back to the same scenario after you’ve fixed what it found. Speed is what makes practice continuous.

The harder problem is context

Speed on its own isn’t enough. AI is only as useful as what it knows about your place. An exercise built from the internet’s average disaster tests a generic jurisdiction, not yours. It can be fluent and polished and still be wrong about your hospitals, your shelters, your agreements and your people.

So the deeper problem is context: capturing what’s true about your place, keeping it accurate, finding the right piece at the right moment, and maintaining it as things change. When AI designs or runs an exercise, it should be working from your reality, not a generic one.

That’s what we’ve built Preppr around: a living context layer. It goes well beyond passages pulled from documents. Preppr observes what happens in the room, finds patterns, forms rules and confirms them with you, relates everything it knows, surfaces gaps and tracks the improvements that close them. And it checks its own work. It comes down to five jobs.

Capture

Exercises start from your own plans, risk data, after-action reports and past findings, not the internet’s average disaster. Preppr reads more than the text: the diagrams, org charts and tables in your plans are read too.

Documents are only part of the record. The work your team does in Preppr goes into it as well: exercise designs and the conversations that shaped them, the findings from every exercise you run, and the after-action record of every simulation. Preppr’s Observer can also join your meetings and capture the decisions, action items and moments people flag. A meeting tied to an exercise joins the record by itself, and other meetings wait for a person to review them.

Accuracy

Preppr pulls values from your documents with the page and sentence they came from, so anyone can check where a number or a commitment was found. Each value stays a candidate until a person checks it. When two plans give different values for the same thing, Preppr flags the disagreement and shows both sources, and you can aim an exercise right at that seam to see which plan people follow. When you upload what looks like a new version of a plan, Preppr suggests the match and a person confirms it. The old version then drops out of search, and any answer that cited it is marked as made against an earlier version. Every Ask Preppr answer is labeled by how well your sources support it: grounded, needs review, or largely unverifiable. You can see at a glance what rests on your own sources and what still needs a person to look at it.

Connection

Preppr doesn’t just store what it reads. It forms rules from the “shall” and “must” in your own plans. And because your team’s work flows into the same record, Preppr can see how your team actually works, not just what the plan says. When the same issue comes up on separate occasions, such as a finding from an exercise, a decision made while designing another and an action item from a planning meeting, Preppr draws out the pattern and sets it against your rules: does practice match the plan, or drift from it?

The bar for a pattern is deliberately high. Points from a single conversation count once, however often they come up, and a simulation played only by AI crews can support a pattern but never create one on its own. Rules and patterns stay suggestions until a person accepts them. Preppr also learns your team’s preferences from how people correct and edit its work, and it lists every one it has picked up.

It relates plans, requirements, facts and findings on one graph, each traced to its source, so anything Preppr knows can be followed back to where it came from.

Retrieval

Knowing a lot isn’t the same as using the right thing at the right time. When Preppr designs an exercise, it searches the sources attached to that exercise and the collections you choose. Ask Preppr answers from the documents you select and your knowledge base, and it cites the passages it used. In a simulation, you decide how well each seat can find things in your plans, from perfectly to not at all, so you can test what happens when people can’t find the procedure.

Maintenance

Context goes stale the same way plans do. Findings from an exercise can become improvement work, each with an owner and a due date, and Preppr reminds owners as dates come due. When a new document looks like it closes a gap, Preppr suggests it to the owner, but it never closes a gap by itself. You can design the next exercise straight from the open actions, so the fix gets tested.

And Preppr reviews itself. Every night it goes over what it has learned. It makes small changes on its own, like new topics and connections, but only after testing that they don’t make it worse at finding your sources, and every one can be undone. Bigger changes, like revising a rule or settling a clash between two rules, wait for a person to decide.

Together, that’s a living context layer: a record of your place that grows with every plan, meeting, exercise and fix, and keeps up as the place changes.

The judgment stays with your team

Grounding goes beyond your own documents. Scenarios are checked against FEMA and Hazus data for your jurisdiction, so the hazard in an exercise behaves the way it would in your place.

And AI findings stay evidence until a person confirms them. Preppr can surface a gap, cite where it came from and suggest what to do about it, but your team decides what’s true and what changes. The point isn’t to hand preparedness to a machine. It’s to give the people responsible for it a faster, better-grounded way to practice.

Keeping up

Preparedness that keeps up with reality needs both halves: practice fast enough to run often, and a living context layer that keeps the practice about your place. Speed was the first problem. Context is the harder one, and it’s the problem we built Preppr to solve.

See how Preppr works or talk to us.