AI Dossier Generator
AI Dossier Generator
The AI Dossier Generator is an LLM-powered step in the scoring pipeline that turns raw review and feature data into three actionable text artefacts for every product analysed. These artefacts are designed to be copied directly into the SaaS platform builder to accelerate product definition.
How It Works
After a product is scored, the AI Dossier Generator is invoked automatically as part of the pipeline. It reads the product's reviews, ratings, feature signals, and scoring data, then calls the LLM to produce the three artefacts described below. Results are persisted and immediately available on the product dossier page.
The Three Artefacts
1. Auto-drafted Mission Statement
A concise, opinionated mission statement written from the perspective of a competing SaaS product. It is framed around the market gap the product could fill, based on the weaknesses found in competitor reviews.
Use it to: Seed the mission statement field when creating a new product in the platform builder.
2. Suggested Feature List
A prioritised list of 10 features, ordered by market gap and demand signals extracted from review data. Features at the top of the list represent the highest-impact opportunities relative to what existing competitors are failing to deliver.
Use it to: Copy the feature list directly into the platform builder's feature definition step.
3. Competitor Weakness Analysis
A structured analysis of recurring negative review themes across the competitor product's review corpus. Themes are grouped and described so you can clearly see where the incumbent is losing customers.
Use it to: Inform positioning, prioritise features that resolve known pain points, and craft differentiated messaging.
Viewing the Dossier
- Run a directory crawl and wait for the scoring pipeline to complete.
- Click any product in the ranked results dashboard to open its full dossier.
- The three artefacts appear in dedicated sections at the bottom of the dossier page.
- Use the copy-to-clipboard button on each artefact to copy the content instantly.
Data Storage
All three artefacts are stored as JSONB in the product_dossiers table. The schema is versioned within the JSONB payload, which means new artefact types can be added in future releases without requiring a database migration.
| Column | Type | Description |
|---|---|---|
dossier_artefacts | JSONB | Contains mission_statement, feature_list, and weakness_analysis keys |
Pipeline Integration
The AI generation step runs after the composite opportunity score is calculated. It does not affect scoring — it only enriches the dossier with human-readable output. If the LLM call fails for a given product, the pipeline continues and the artefacts field is left null; a retry can be triggered from the dossier page.