Strategy context index

Raw source preview

Raw, unprocessed file text shown below.

---
title: "Strategy Context Index (AI-facing orientation guide)"
status: active
tags: [context, internal, ai-agent, orientation, index]
last_updated: '2026-04-18'
---

# Strategy Context Index

*This document is written for AI agents and coding assistants entering this repository. It provides the minimum context needed to understand what EmpathyIQ is, how the repository is structured, and what the most important documents are.*

---

## What is EmpathyIQ?

EmpathyIQ is an enterprise market research platform for client-side insight teams at mid-to-large companies. It embeds quality assurance and knowledge management directly into the research lifecycle.

**The problem it solves:** Enterprise insight teams produce research that gets siloed, challenged, and bypassed. Other departments run their own low-quality research (polls, ChatGPT queries) because the insight team is too slow or too hard to access. Past research is buried in shared drives. Data quality is checked manually and expensively after fieldwork ends.

**The solution in one sentence:** EmpathyIQ makes every study defensible from design to boardroom, and turns research into a structured knowledge base the whole organisation depends on.

**The positioning line:** *Become the team the business trusts for customer understanding.*

---

## The three-layer system

| Layer | What it does |
|---|---|
| **Research Execution** | Run quantitative surveys, qualitative discussion groups, AI interviews. Supports multi-language. |
| **Knowledge** | Structured, searchable repository of all research - studies run on the platform plus externally uploaded documents, PowerPoints, and reports. |
| **Confidence** | AI-powered quality assurance throughout: before launch (questionnaire review, briefing), during fieldwork (Trust Centre, 38+ respondent quality metrics), after fieldwork (claim corroboration, defensible reporting). |

---

## The confidence framework (core positioning)

EmpathyIQ's differentiation rests on "confidence before / during / after research":

- **Before:** Catch design flaws (ambiguous questions, leading language, structural risks) before launch. Launch readiness checklist. Brief interrogation against the knowledge base.
- **During:** Real-time respondent quality monitoring via the Trust Centre. Dimensions: Real (is this a legitimate person?), Unique (are they genuinely distinct?), Engaged (are they paying attention?). 38+ metrics per respondent. The strongest current wedge is live review while the study is still running, with flagged respondents reviewed and actioned during fieldwork rather than only cleaned up later. This is not the same as full Qualtrics-style post-field response-quality cleanup, which remains a product gap.
- **After:** Claim checking against underlying data. Cross-study corroboration from the knowledge base. Defensible, traceable reports with audit trail.

---

## The portfolio hierarchy

Every item in the product portfolio fits this hierarchy:

```text
Portfolio (EmpathyIQ)
  -> Product (e.g., Quantitative Research Platform)
    -> Core Workflow (e.g., Survey Design / Survey Deployment / Analysis & Insight)
      -> Job Area (e.g., Topline Reporting)
        -> Feature Family (e.g., Reusable Reporting Outputs)
          -> Feature (e.g., Convert narrative outline into charts)
            -> Document (detailed spec or PRFAQ in body)
```

**The canonical source of truth for portfolio structure is [portfolio/portfolio.csv](../../portfolio/portfolio.csv).** The CSV has columns: Name, Type (Product/Core Workflow/Job Area/Feature Family/Feature), Outcome/Promise Narrative, Market Readiness, Now/Next/Later, Notes, Root Parent, Parent item.

---

## Products in the portfolio

1. **Platform (Shared Services)** - identity/auth, authorisation, billing, observability
2. **Quant Surveys** (Quantitative Research Platform) - survey design, deployment (Trust Centre), analysis & insight
3. **Discussion Groups** (Qualitative Discussion Group Platform) - async qual forums
4. **AI Interviews** - outsourced AI interview integration (Tellet), branded inside `<un>peel`
5. **Panel** - custom panel management, respondent nurturing, reward tracking
6. **Research Architect** - research planning and context gathering from brief outward; feeds into Research Guard
7. **Insight Navigator** (AI Knowledge Repository) - research retrieval via natural language; project-level and platform-level querying

---

## Key decisions (locked)

These decisions have been made and are not currently under review:

- **ICP:** Enterprise client-side researchers (not agency-side first; not SMB first)
- **Business model:** Enterprise SaaS with possible "software-under-service" delivery modes for pilot customers
- **Suite posture:** EmpathyIQ is a Research OS / suite, not a point tool
- **Architecture:** Confidence before/during/after maps to Survey Design / Survey Deployment / Analysis & Insight
- **Portfolio canonical system:** Notion Architecture Nodes DB -> [portfolio/portfolio.csv](../../portfolio/portfolio.csv) in this repo

See [decisions/strategy-decisions.md](../../pm/decisions/strategy-decisions.md) for full rationale.

---

## What to read first (by use case)

| Use case | Document |
|---|---|
| Understand what EmpathyIQ is | [strategy/in-depth-what-is-empathyiq.md](../../Strategy/in-depth-what-is-empathyiq.md) |
| Understand the positioning and differentiation | [strategy/confidence-positioning-brief-internal.md](../../Strategy/confidence-positioning-brief-internal.md) |
| Understand the feature portfolio and build what is possible | [portfolio/README.md](../../portfolio/README.md) -> [portfolio/portfolio.csv](../../portfolio/portfolio.csv) |
| Understand the product architecture taxonomy | [context/product-architecture.md](../product-architecture.md) |
| Understand AI PoC setups (questionnaire reviewer, persona builder, etc.) | [context/poc/](../poc/) |
| Understand the MEAT IQ ingestion pipeline (AI document processing) | [context/meatiq-documentation.md](../meatiq-documentation.md) |
| Understand what customers actually said in interviews | [research/opportunities/README.md](../../research/opportunities/README.md) |
| Understand the competitive landscape | [competitive/feature-set-competitive-review.md](../../competitive/feature-set-competitive-review.md) |
| Understand why key decisions were made | [decisions/strategy-decisions.md](../../pm/decisions/strategy-decisions.md) |

---

## Current state of the portfolio (as of Q1 2026)

- **Quant Surveys** (Quantitative Research Platform): ~95% ready. Survey design, deployment, basic analysis.
- **Discussion Groups** (Qualitative Discussion Group Platform): Owned; separate flow; not deeply integrated yet.
- **AI Interviews:** Outsourced integration (Tellet); branded inside `<un>peel`.
- **Panel:** Live with one customer (Musgrave/ShopTalk). Functional but early-stage.
- **Research Architect:** Page built; product in early definition. Research planning and context gathering from brief outward.
- **Insight Navigator** (Knowledge Repository): Being built; not deeply integrated with other products yet.
- **Trust Centre:** Core differentiator; 38+ quality metrics; built and functional.
- **Questionnaire Reviewer:** PoC running on GPT. See [context/poc/questionnaire-reviewer/setup.md](../poc/questionnaire-reviewer/setup.md).
- **Persona Builder:** PoC running. See [context/poc/persona-builder/setup.md](../poc/persona-builder/setup.md).

Pilot customers as of Q1 2026: Musgrave, UKOmnibus Group (active). Ryanair (target).

---

## What "confidence" means in this codebase

When the word "confidence" appears in strategy documents, it refers to **the ability to defend every decision in the research process** - not a statistical confidence interval. It means:

- The design was reviewed and issues were documented before launch
- The data was quality-checked in real time with a full audit trail
- The conclusions are linked to evidence and can be corroborated against past research

This is different from quality (a feature) or speed (a benefit). Confidence in EmpathyIQ means an **audit trail** exists for every step.