# Start Here (Non-Technical)

If you are not an engineer, start here.

## What this project is

Institutional Reasoning Engine (IRE) is a blueprint for building AI-assisted investigation systems that are:
- Evidence-grounded
- Human-reviewed
- Tamper-evident
- Auditable

## What this project is not

- Not legal advice
- Not a ready-made software product
- Not an autonomous decision-maker

## 10-minute orientation

1. Open the public site landing page (`docs/index.html`).
2. Read the overview in `README.md`.
3. Open `IRE_Builder_Guide.html` (or `docs/ire-builder-guide.html`).
4. Review deployment profiles (1 to 6) and choose your starting profile.
5. Open the matching checklist under `checklists/profiles/`.

## How to choose your starting profile

- Start with **Profile 1** if you need internal investigations with strong controls.
- Start with **Profile 2** if interviews/transcripts are central.
- Start with **Profile 3** if document review is your primary workload.
- Move to **Profile 4–6** only when legal/regulatory stakes and scale justify it.

## What to prepare before any build

- A one-page investigation methodology
- Your target jurisdictions and regulatory context
- Data handling boundaries (PII/PHI/financial secrecy)
- Human review roles and approval authority
- Audit retention policy

## Recommended first milestone

Implement and validate the 8 mandatory base components before adding advanced clusters.

## Where to ask questions

- Open a GitHub Issue using templates in `.github/ISSUE_TEMPLATE/`
- For broader discussion setup, see `docs/community.md`
