1. 01

   Discovery

   1–2 weeks (2–4 if data assembly is needed)

   We pin the system to one or more named, recurring decisions, inventory your data sources, and agree the scoring or forecasting approach in plain terms before anything is built.

   You provide

   A decision owner for weekly sessions, access to the data landscape, and honesty about what data doesn't exist.

   You receive

   A written system definition: decisions supported, dimensions, sources, gaps, and a phased plan with week bands.

2. 02

   Research & Modeling

   2–5 weeks

   We build and validate the model — scoring dimensions, weights, forecast structure — with your team in the loop. Formulas are readable, assumptions are documented as we go.

   You provide

   Domain review of draft scores and assumptions; source credentials.

   You receive

   A working model with a methodology document, run against your real data, before platform work begins.

3. 03

   Platform Development

   3–8 weeks

   The model becomes a product: ingestion on a schedule, dashboards shaped around the decisions, alerting where it earns its place, and the methodology published inside the tool.

   You provide

   Feedback on working software in weekly reviews.

   You receive

   The platform, running on live data, with documentation and access controls.

4. 04

   Launch & Improvement

   1–2 weeks, then ongoing as agreed

   We launch with your team, watch how the system is actually used, tune weights and views against real decisions, and hand over ownership.

   You provide

   Real decisions run through the system.

   You receive

   Tuning, training, a runbook, and a defined improvement cadence.

The rules behind every phase are published: read the seven analytical standards.