You rolled it out. Did it land?
Everyone was briefed. Completion is at 94%. But nobody knows whether your people are actually making the calls the new process intended — or where the gap is by team, site, or shift. A Dry Run finds out.
After the rollout
measured
unmeasured
The problem this addresses
You introduced the new model — memos, town halls, a training session, sign-offs collected. Now someone has to say whether it worked. This is the evidence they have:
What the rollout gives you
- Completion rate — 94%
- A post-rollout survey
- Whatever managers report upward
What it can’t tell you
Whether a team lead facing two escalations at 11am makes the call the new process intended.
Attendance isn’t adoption.
What we do
A measurement, not a survey
We build a simulation of your new process and put your people through it — a working model of the decisions, not a quiz on the memo. Then we compare the calls they make against the ones the design intended.
Where the escalation rule actually gets followed
By site and shift — the pattern a completion rate can’t show.
You get two things out of it
You find out what actually landed. Not who attended. Which decisions are being made the way the design intended, and where the divergence is concentrated.
You find out where the design is at fault. Some gaps aren’t comprehension. A rule that reads clearly and collapses under two simultaneous escalations is a design defect, not a training failure. The report separates them.
The second one is the part nobody else offers. Completion rates tell you people attended. Running the decisions tells you which ones your design gets wrong.
Non-negotiable
Group-level, not individual
Leadership sees the pattern — which sites, which shifts, which decisions. Individuals see their own results and nobody else’s.
This isn’t diplomacy. If people believe they’re being personally graded, they play to look correct instead of deciding naturally, and the data stops being worth having.
Leadership sees
The whole pattern
Which sites, which shifts, which decisions.
Each person sees
Their own result
And nobody else’s.
The engagement
Fixed scope. Fixed price. Finite.
Every Dry Run runs the same shape.
Discovery
We sit with the people who designed the process and those who’ll run it, and pin down the decisions that carry risk.
Build
We model those decisions and calibrate against your own numbers — volumes, targets, staffing.
Run
Two cohorts go through it. Browser and mobile — no installs, no facilitator required on site.
Report & revise
We show you where understanding broke and where the design broke. One revision to the simulation.
Total: 5–6 weeks. Then we hand over the results and go.
What you receive
- A working simulation of your process, running on our infrastructure
- A model diagram your ops lead can read and correct — plain enough to argue with
- Group-level results: where divergence concentrates — by team, site, and shift
- A findings report separating “people didn’t understand X” from “X doesn’t work as designed”
- A record of every assumption we calibrated and who confirmed it
What you don’t receive
No system to maintain. No licence to renew. No integration into your stack. If you want the simulation kept alive for future cohorts, that is a separate conversation you can have later, or not at all.
Where rollouts leak
Rollouts don’t fail on the steps. They fail on the decisions.
Nobody forgets which button to press. What goes wrong is the moment the new process asks someone to choose — and two experienced people would choose differently.
A new client account with SLAs nobody has run against.
A staffing model that trades cost against coverage.
An escalation rule that reads clearly, until two things escalate at once.
Every one of those is a place where your design meets someone’s judgment, and where a briefing can’t tell you what will happen. That’s where a rollout quietly diverges — one site interpreting it one way, another site the other, and nobody finding out for a quarter.
What you’re actually buying
Not the software
That part gets easier every month. Three things instead.
The reduction
Finding the one decision that carries the risk and cutting everything else away. The hard part — and what decides whether a simulation teaches anything. Model the whole process and you get a faithful replica that teaches nothing, because everything is in it.
An outside read
Someone who didn’t design the process, checking whether it holds up. Structurally hard to do in-house.
A date
Five to six weeks, fixed scope, fixed price. Internal projects with no client waiting tend to ship at 70% and stall.
Pricing
Priced per engagement, not per seat, not per year
Dry Run
As scoped above.
Additional cohorts
Beyond the first two.
Ongoing use for future intakes
Optional — only if you want it.
No per-participant licensing. Run it with fifteen people or eighty; the price is the same.
Bracketed figures are placeholders — final pricing is quoted per engagement.
Starting
The smallest useful first step
In order of preference:
- 1
Send us the process document — the design, the new rules, the deck. We’ll tell you within a week whether there’s a real simulation in it, and if there isn’t, we’ll tell you that too.
- 2
Ninety minutes with the people who’ll run it. We’ll run one of our existing simulations with them. If they don’t come out of it arguing about a decision they got wrong, there’s nothing here for you.
- 3
Your last three incident or post-mortem reviews. If the same failure appears twice, that’s the simulation — and you’ve already documented the model for us.
Live work
Simulations you can play right now
Strategy · AI Safety
Race to AGI
Run a frontier AI lab racing three rivals to build AGI. The twist: getting there first isn’t the win — getting there safely is. A browser strategy sim about long-term thinking under pressure.
Management Sim · Cloud Architecture
CloudArchitect
Run a live cloud environment and juggle cost, latency, and availability under pressure. A terminal-styled management sim that teaches real infrastructure trade-offs — with an assessment mode for hiring and upskilling.