Decision DataViz
A decision-making DataViz product to steer ~150 projects, from discovery to rollout.
Impact
View case95%average feature adoption (target 80%)+55NPS75%reporting preparation time saved
Product Manager · AI & Data Builder · Paris area
From need to adoption.
5 years of experience, nearly 4 in Product, at the intersection of Product, Data and AI.
95%
feature adoption at AÉSIO Mutuelle
Monthly average per feature, across ~250 users
−50%
invoicing costs at Biogaran
Observed from rollout, on the MVP pharmacies
~€340k
revenue signed at Arpon Technologies
Across 3 new accounts opened in Brussels
Selected work
Tools people actually use.
Noé × BlaBlaCar
Taking new drivers’ successful-trip rate from 30% to 40%: from data to a coded prototype.
At a glance
View case40%successful-trip target in 12 months (30% today)~6points of successful trips identified in the data, out of the 10 targeted3changes after testing
Prisme
An autonomous agent that reads the news, picks, writes, then sends the email at 8:30am.
At a glance
View case14sources read every morning5updates at most per email8:30amthe email lands
Digital invoicing
Pharmacies get their invoices in their customer portal: an MVP to validate adoption before rollout.
Impact
View case90%MVP adoption−50%invoicing costs for MVP pharmacies20+pharmacies interviewed
CHANEL.com
Auditing CHANEL.com’s product organisation and defining its target Product Operating Model.
At a glance
View case20interviews conductedDeliverablesDiagnosisTarget modelRoadmap
Mon Service Formalités
Website for a formalities service, designed and shipped in 16 days (~14 hours of work) with a team of Claude Code agents (dev, QA, content).
Impact
View case16 daysfrom first version to launch~14 hof work in total, images includedLivemonserviceformalites.com
What I bring to a product team.
Product.
Vision, discovery, prioritisation (MoSCoW, RICE), prototyping, user testing, KPIs, roadmap, delivery.
AI & build.
Testing an idea in a few days with a coded prototype, before committing the tech team. A supervised agent team (agentic coding, agentic engineering, loop engineering): Claude Code, skills, MCP, Claude Design, GitHub, Vercel.
Data & delivery.
Reliable numbers when it’s time to decide: SQL, Snowflake, Qlik Sense, DataViz. Scrum delivery: user stories, Jira, Confluence, Figma, Notion.
StackClaude CodeClaude DesignGitHubVercelSQLSnowflakeQlik SenseJiraConfluenceFigmaNotion
Background
Five years across product, data and business.
AÉSIO MutuelleData Product Manager
Decision DataViz · 150 projects · ~250 users · 95% adoption2023 —Noveane (Scalian)Product & Agile consultant
CHANEL.com · Product & Agile TransformationProduct organisation audit · Product Operating Model
CNAV · IT project leadDocument management projects taken over in testing, through to production release
Arpon TechnologiesB2B Business Developer (V.I.E.)
Brussels · 3 new accounts · ~€340k signed2021 – 2022Biogaran (Servier)Digital Product Manager
B2B e-invoicing · −50% costs · 90% adoption2021Education
- NoéProduct Manager · one-month intensive programme, certifiedDiscovery, user research, data, prioritisation, prototype and roadmap, applied to a real case sponsored by BlaBlaCar and presented to its Senior PM2026
- INSEECMSc in Project Management & Business Engineering2021
CertificationsClick a certification to see the credential.
What I build with Claude.
My GitHub$ routine prisme→ 5 updates, one fixed template
Oct. 2026Prisme.An agent that reads AI news every morning, picks and writes for PMs, then sends the email at 8:30am.$ claude "Create an agent…"→ .claude/agents/veilleur.md
Sept. 2026Guide: building Claude agents.A step-by-step for non-developers: a skill, an agent, then three ways to launch it. Shared with PMs from Noé’s cohort 35 (in French).$ /feature→ framing · code · QA · go / no-go
Summer 2026A product team of agents.Four specialised agents (PM, dev, QA, UX) and a validation chain: nothing reaches code without my go, and every failure becomes a rule.$ claude /analyse-entretien→ user journey + sourced quotes
2026PM skills library.Claude skills for discovery, specs and decisions: interview analysis, interview guide, user stories, benchmark. Used on the Noé × BlaBlaCar case.