Skip to content
AnantCausal
Book a walkthrough

Bring a decision, not a use case.

Forty minutes on a decision you actually have to make. You will see the model built, the identification verdict, where every probability came from, and the record that comes out the other end — including the parts where the answer is no.

Walkthrough

What you will see

Six screens from one decision, in the order they appear. The first two are the ones a slide cannot show you.

  1. 01

    The identification verdict

    Whether your question can be answered from the structure you have — with the reason, and the specific thing that would have to change if it cannot. A refusal is a result here, not an error, and you will see one in most sessions because most decisions have at least one version that is not answerable.

  2. 02

    Where each number came from

    Every probability carries its source: elicited from a document, asserted by a named expert, fitted from your rows, or mixed. You will see an elicited guess and a fitted count side by side and be able to tell them apart at a glance, which is the point of storing the tag per row.

  3. 03

    How far the answer moves

    The sensitivity envelope: how much the effect changes when each table is shifted by a fixed amount. It names its own basis, so you can see that it is a sensitivity band rather than a confidence interval — a distinction that is routinely blurred and decides whether the number means anything.

  4. 04

    The attacks that failed

    Three checks run against the result and reported individually, including inconclusive, which is never counted as a pass. If a check fails, that is the most useful thing in the session.

  5. 05

    The decision record

    What was approved, against which stored analysis, what was assigned, who was actually reached, and the measured outcome reconciled against the design-time claim. This is the artifact a reviewer asks for eighteen months later.

  6. 06

    The handoff

    The graph as DOT, the adjustment set, and the role of every variable — including which ones you must not adjust for. If you already run an estimation stack, this is what it has been missing.

What this is not: a slideware walkthrough. If your decision has a structural problem, the session will surface it and we will say so — including when the answer is that this is the wrong tool for the question. That has happened, and it is cheaper to find out in forty minutes than in a procurement cycle.

Fit

Is your decision a fit?

Good candidates

Start with a decision that has consequences.

The strongest sessions involve a decision someone is already committed to, where a small number of discrete variables drive the outcome and the relationships between them are knowable from documents, experts or data. Capacity is a real constraint: a defensible model of twenty-four variables, not a replica of your organization.

What to bring

A description is enough to start.

Tell us the decision and who has to defend it, the options you would choose between, the variables you believe matter and how you know they matter, and what you already hold as data. Please do not send patient records, credentials or production data — the session does not need them and this page says so for a reason.

Ask for one

The more specific the decision, the more useful the session. If it is not a fit we will tell you rather than run the meeting.

Email a demo request

Or copy this address into your email service: sales@ananthq.com

This opens your email app. Nothing is submitted by this website; review and send the email to reach us. Please do not include patient data, credentials, financial account information or other sensitive production data.