> ## Documentation Index
> Fetch the complete documentation index at: https://help.terriqon.com/llms.txt
> Use this file to discover all available pages before exploring further.

# How Terriqon Works: From Field Data to Approved Report

> Understand the end-to-end Terriqon workflow: form design, offline data collection, AI report generation, PII protection, and the human approval process.

Terriqon is built around a single promise: your field team collects raw data, and a verified, funder-ready report comes out the other end — with a named human approving every step. This page walks you through the complete end-to-end workflow, from the moment you design a form to the moment a report is available for download.

## The three-stage workflow

Every Terriqon project moves through three stages. Each stage is designed so that the next one can only begin when the previous one is complete and verified.

<Steps>
  <Step title="Build and deploy">
    An Organization Admin or Manager uses the drag-and-drop form builder to design a data collection form. Fields are configured with types, validation rules, and — where relevant — PII tags that mark data for automatic stripping later in the pipeline. Once the form is published to a project, it becomes available on every assigned Field Officer's PWA instantly, ready to use online or offline.
  </Step>

  <Step title="Collect in the field">
    Field Officers open the Terriqon PWA on their phone or tablet and fill in the form at the point of work — whether that's a remote farm plot, a rural clinic, a construction site, or an emergency distribution point. If there's no internet connection, submissions are stored securely on the device. The moment the device detects connectivity, all pending submissions sync automatically to the Terriqon platform. No manual intervention is required.
  </Step>

  <Step title="Generate and approve">
    Once enough submissions are available and have passed data quality checks, a manager triggers AI report generation. Terriqon's pipeline validates the data, strips PII, aggregates the metrics, and drafts a structured report. The manager reviews the draft, makes any edits needed, and approves it. Only after approval is the report available for download or sharing.
  </Step>
</Steps>

## The AI reporting pipeline

Between raw field submission and finished report, your data passes through four sequential stages — each designed to protect your beneficiaries and ensure report accuracy.

**1. Data validation and quality gating**
Every incoming submission is checked against the form's validation rules — required fields, acceptable value ranges, and logical consistency checks. Submissions that fall below the quality threshold are flagged and routed to a human reviewer before they can contribute to a report. This prevents low-quality data from silently distorting your results.

**2. Automated PII stripping**
Before any data reaches the AI model, Terriqon automatically removes all personally identifiable information from the dataset. GPS coordinates, personal names, phone numbers, national ID numbers, and any other fields you tagged as PII during form design are scrubbed from the aggregated dataset. The AI only ever sees sanitized, aggregate-level data — never individual beneficiary records.

**3. AI report generation**
With a clean, validated dataset, Terriqon's AI generates a structured report in your selected format — Project, Ops, Donor, or Portfolio. Each report type is structured to the conventions your funders and stakeholders expect, with narrative sections, key metrics, and supporting data tables generated from your actual submissions.

**4. Human review and approval**
The AI-generated draft is presented to a named manager for review in the Terriqon report editor. The manager can read every section, make manual edits, and either approve the report or reject it. Approval makes the report available for download. Rejection requires the manager to provide a written reason, which is logged permanently in the audit trail.

## Data safety by design

<Info>
  No AI-generated report can be downloaded or shared until a named manager has explicitly approved it. If a manager rejects a draft, they must provide a written reason — this reason is recorded permanently alongside the rejection in the report's full audit trail. Every approval, rejection, edit, and download is logged and cannot be altered or deleted. This audit trail is available to Organization Admins at any time.
</Info>

PII stripping is automatic and non-negotiable: it runs on every report generation, regardless of plan or report type. You cannot generate a report that bypasses PII stripping. This design means your beneficiaries' personal data is never exposed to the AI layer, and your organization stays on the right side of data protection obligations.

## Offline-first architecture

The Terriqon mobile experience is a Progressive Web App (PWA) — installed directly from a browser, with no app store required. The PWA is built to function fully in zero-signal environments. When a Field Officer is offline, submissions are stored locally on the device and sync to the platform automatically the moment connectivity returns. Field Officers never need to manually trigger a sync or worry about losing work.

<CardGroup cols={2}>
  <Card title="AI Reports" icon="sparkles" href="/concepts/ai-reports">
    Dive deeper into the four report types — Project, Ops, Donor, and Portfolio — and learn when to use each one.
  </Card>

  <Card title="Offline Mode" icon="wifi-slash" href="/concepts/offline-mode">
    Learn how the Terriqon PWA stores, encrypts, and syncs submissions in zero-connectivity environments.
  </Card>
</CardGroup>
