(toolNum, toolName) => ``
Tool 12 · Historical Intelligence

Analogous Project
Retrieval Engine

Surface genuinely similar past SAP engagements from a knowledge base of 15 real-world projects. Uniqueness-weighted scoring ensures no two retrieved results are the same.

Retrieval Model
TF-IDF + Jaccard Similarity
Deduplication
MMR (Max Marginal Relevance)
15
Reference Projects
8
Industries
MMR
Diversity Engine
How This Tool Works

Upload your RFP or project document, or manually set project parameters. The engine scores each reference project using TF-IDF term overlap across module, industry, and complexity dimensions, then applies Maximal Marginal Relevance to guarantee diverse, unique results — no repeated project profiles.

Step 1
Similarity Scoring
Module, industry, region, complexity weighted
Step 2
MMR Re-ranking
Penalises redundant results for diversity
Step 3
Result Presentation
KPIs, risks, lessons, outcome shown
Step 4
Export & Compare
Side-by-side comparison across projects
Live Demo

Find Your Analogous Projects

Set your project parameters or upload an RFP. The engine retrieves the most relevant and diverse reference projects.

Upload Document (Optional)
Upload RFP / Project Brief
PDF · DOCX · TXT — auto-extracts modules, industry, scope
Project Parameters

Set parameters or upload a document to retrieve analogous projects

Architecture

Retrieval Algorithm Explained

How the engine guarantees unique, relevant, and diverse results every time

TF-IDF Similarity
Term frequency-inverse document frequency weights module, industry and region keywords to score each reference project against your query profile. Rare term matches (e.g. FSCD, IS-Oil) score higher than common ones (FI, MM).
Jaccard Index
Set intersection over union of modules requested vs. modules used in each reference project. Ensures projects with more matching SAP modules rank higher — even when the industry differs.
Maximal Marginal Relevance
After initial ranking, MMR re-orders results to balance relevance with diversity. Each additional result is penalised if it is too similar to an already-selected project — guaranteeing unique perspectives in every retrieval.
Weighted Feature Vector
Each project and query is represented as a feature vector: Industry (35%), Modules (30%), Region (15%), Complexity (12%), Scale/Budget (8%). Euclidean distance in this space produces the base similarity score.
Document Extraction
When an RFP is uploaded, the engine scans text for SAP module keywords, industry terminology, and scope indicators. Extracted signals auto-populate the query profile before scoring begins.
Knowledge Base
15 curated reference projects spanning 8 industries, 3 global regions, and 30+ SAP modules — each with verified KPIs, risk profiles, lessons learned, and delivery outcomes from real engagements.
() => ``