How it works
How Simi reads your work — and coaches it.
A live case of real work, coaching with a citation under every line — this page is the method, in public.
Real work, not quizzes.
You write the decision in your own words, not from a list of options — and every observation in your feedback cites the moment it came from.
And not a scored conversation either. A Simi run is the whole job — people, documents, a deliverable, a decision — read for judgment, not delivery.
Two layers of guidance.
Your debrief reads your work on two levels — the structure published, the tone a coach’s. Then a manager debrief on video walks through your decision and a stronger candidate’s.
Completeness — the floor
The moves a strong response includes — clear, fair, the same for everyone, each linked to your own work.
Wisdom — the ceiling
Was the call wise — what it cost, what a stronger professional did, and why. Coaching, not a verdict.
Evidence-gated.
Feedback draws only on what your session actually surfaced. Your run is read against your run — never a perfect transcript you never saw.
You asked for the channel split before recommending — and cited it.
cited · your 12:41 messageAnything the case never showed you
never held against youSubstance over fluency.
English as a second language is never penalized — structure and synthesis count, grammar doesn’t. Simi is built to read judgment, not prose.
Plain sentences that make the right call — reads strong.
A polished paragraph that dodges the decision — reads weak.
Teaching at the moment of failure.
Wherever your run fell short, you see the move that was available and the moment you could have made it — what stronger looked like.
“The 40% figure changed the recommendation. Here’s the moment it was on the table — and the question that would have surfaced it.”
The craft
What a simulation is made of.
The claims ladder, in public.
Platforms tend to claim everything on day one; we publish a ladder and climb it in order.
Preparation & development
Rehearse a role, build a skill, walk in prepared — this is what Simi claims today.
Onboarding formats
New hires living the role before week one — the next rung, in development now.
Formal verification
Only when our validation program earns it — we publish the evidence when we make the claim. Hold us to it.
Data & AI transparency.
Your run is yours — the same promises we sign with institutions.
Consent, up front
You see what a run collects before you start. Nothing rides along silently.
Minimization
We keep what the simulation needs to run and coach you — no more.
No training on your runs
Your transcripts, artifacts, and results are never used to train AI models.
Retention, under your control
Ask, and your runs are deleted. Institutional cohorts set their own windows.
What the AI does
Plays your colleagues, drafts the working documents, prepares your feedback against a published structure.
What the AI doesn’t decide
Simi doesn’t gate jobs, grades, or credentials today — and won’t claim to verify anything until the validation evidence is published.
A question this page doesn’t answer: elie@simild.com — a person reads it.
Read about the method. Or feel it.
No subscriptions. No multiple choice. The screening room is free; a real simulation is $16.