BioMérieux x MADE
Sylvain
Project Context
Knowledge saturation Issue
The right doc at the right moment
Companionship needed
Access to knowledge is critical as GxP environments often overflow with data
Field agents work on more that 15 different ranges
Great potential in empowering Service Teams with AI
As-Is & Objective
As-IS
To-be
Multiples platforms to query knowledge
Knowledge centralized around AI
Unreliable search function
Hybrid semantic & lexical search
Knowledge Access
Relevant info buried in hundreds pages docs
AI search quotes & sources relevant section
Troubleshooting documentation is inconsistent and heavily reliant on individual initiative
Standardized content stemming from a global & structured initiative
Data Structure
Data organized in knowledge base, tagged with metadata
Unstructured, unidentified troubleshooting data
First runs results
Correctness
Real life question set tests to measure effects of data evolution Nearly 90% correctness reached over 3 months
Run 5
AI-friendly content formatting – round 2
Run 4
88%
AI-friendly content formatting – round 1
April
Run 3
Run 2
April
Custom indexing using metadata
Splitting data into separate databases
75%
March
68%
Run 1
Testing on As-Is data
March
52%
40%
February
Data rework
Next Steps Tool Fine Tuning
Safeguard & Compliance
RAG Optimization
Implement guardrails for questions & answers
Dynamic chunking, overlapping chunks, …
History & similar queries
Prompt engineering
Leverage query history to improve consistency and reduce compute usage
Customization based on user role or input
Feedback & Metrics
Integrate user feedback and track adoption metrics for continuous improvement
Project Gains
Enhanced Efficiency
Faster issue resolution through streamlined resource access
Improve Autonomy
Service Consistency
Standardized access to documentation boosts team confidence and autonomy
Better repeatability with less reliance on personal expertise & local initiatives
Customer Satisfaction
Fast Track Onboarding
Direct result of empowering service teams
New hires lose time searching for the right information
MADE - Summer event -RISE.pptx
Pierric Gubian
Created on September 9, 2025
Start designing with a free template
Discover more than 1500 professional designs like these:
View
Teaching Challenge: Transform Your Classroom
View
Frayer Model
View
Math Calculations
View
Interactive QR Code Generator
View
Piñata Challenge
View
Interactive Scoreboard
View
Interactive Bingo
Explore all templates
Transcript
BioMérieux x MADE
Sylvain
Project Context
Knowledge saturation Issue
The right doc at the right moment
Companionship needed
Access to knowledge is critical as GxP environments often overflow with data
Field agents work on more that 15 different ranges
Great potential in empowering Service Teams with AI
As-Is & Objective
As-IS
To-be
Multiples platforms to query knowledge
Knowledge centralized around AI
Unreliable search function
Hybrid semantic & lexical search
Knowledge Access
Relevant info buried in hundreds pages docs
AI search quotes & sources relevant section
Troubleshooting documentation is inconsistent and heavily reliant on individual initiative
Standardized content stemming from a global & structured initiative
Data Structure
Data organized in knowledge base, tagged with metadata
Unstructured, unidentified troubleshooting data
First runs results
Correctness
Real life question set tests to measure effects of data evolution Nearly 90% correctness reached over 3 months
Run 5
AI-friendly content formatting – round 2
Run 4
88%
AI-friendly content formatting – round 1
April
Run 3
Run 2
April
Custom indexing using metadata
Splitting data into separate databases
75%
March
68%
Run 1
Testing on As-Is data
March
52%
40%
February
Data rework
Next Steps Tool Fine Tuning
Safeguard & Compliance
RAG Optimization
Implement guardrails for questions & answers
Dynamic chunking, overlapping chunks, …
History & similar queries
Prompt engineering
Leverage query history to improve consistency and reduce compute usage
Customization based on user role or input
Feedback & Metrics
Integrate user feedback and track adoption metrics for continuous improvement
Project Gains
Enhanced Efficiency
Faster issue resolution through streamlined resource access
Improve Autonomy
Service Consistency
Standardized access to documentation boosts team confidence and autonomy
Better repeatability with less reliance on personal expertise & local initiatives
Customer Satisfaction
Fast Track Onboarding
Direct result of empowering service teams
New hires lose time searching for the right information