After exiting Comptasanté (110 employees) in 2023, I came back to build BoosterMail. Roast my launch 🙏
Hey Indie Hackers 👋
Quick intro: I'm Yvan, French founder. From 2014 to 2023 I built Comptasanté (B2B SaaS in healthcare admin) to 110 employees and several thousand customers, before selling to IK Partners.
After the exit I took a long pause. Tested dozens of tools, talked to hundreds of entrepreneurs, explored AI/SaaS deep. I thought I was done with operations.
I was wrong.
A few months ago I started BoosterMail — and I'd love your honest feedback before I push harder on launch.
═══════════════════════════════════════
THE PROBLEM
═══════════════════════════════════════
Email eats 30% of the average pro's workday (Adobe/McKinsey 2019). The IA assistants out there are great at generic emails — but generic is exactly what nobody wants to send.
When I read a real email from a real client, I don't want a template. I want a reply that sounds like me, calibrated to that specific person, with the context of our last 12 exchanges. In 4 seconds. Without leaving my inbox.
═══════════════════════════════════════
THE PRODUCT
═══════════════════════════════════════
BoosterMail is an Outlook add-in (not a separate app, not a Chrome extension — lives inside Outlook itself):
🧠 Learns your writing style from your last 800 sent/received emails
👤 Builds a relational profile per contact (formal vs casual, length, tone)
⚡ Drafts replies in ~4s using Claude (with speculative pre-generation while you read)
📎 Reads PDF/Word/Excel attachments and integrates content into replies
📁 Auto-files emails AND attachments into the right Outlook/Windows folders
⏰ Detects deadlines and tracks unanswered commitments
Stack: Python/Flask backend on OVH (France), Office.js add-in, Anthropic Claude API,
Microsoft Graph, JavaScript vanilla frontend. ~50K LoC at this point.
═══════════════════════════════════════
WHERE I AM RIGHT NOW
═══════════════════════════════════════
✅ Live product: https://www.boostermail.ai
✅ Multi-tenant SaaS deployed, GDPR-compliant (data hosted in France)
✅ ~5 beta users (mix of lawyers, exec consultants, sales directors)
✅ I use it myself daily on 100+ emails/day — eating my own dog food
🔧 Stuck on: marketing distribution (huge blind spot — I'm a product/ops person, not a marketer)
🔧 Pre-monetization (custom pricing only for now, public pricing once I have ~30 beta users)
═══════════════════════════════════════
WHAT I'M LOOKING FOR
═══════════════════════════════════════
1. Honest product feedback (the website is fresh — what's confusing? what's missing?)
2. Beta users in: legal, consulting, sales, executive roles, anyone doing 50+ emails/day
3. Distribution advice — what worked for YOUR B2B SaaS in early days?
4. Comments on positioning: am I clear enough that this is "your style, in Outlook" vs another generic AI tool?
═══════════════════════════════════════
ONE QUESTION FOR YOU
═══════════════════════════════════════
For those of you who built B2B SaaS: when did you stop "feature building" and start "distribution building"? I have a strong gut feeling I'm late on that switch.
Thanks for reading. Roast me 🙏
— Yvan
A few details I omitted from the original post 👇
📋 CURRENT STAGE
We're running a controlled private alpha with a limited cohort of target users (lawyers, senior consultants, sales leadership) that I onboard personally. The progressive rollout is deliberate: before scaling distribution, I want exemplary retention on the first users — a far better predictor of B2B success than signup vanity metrics.
🎯 CORE CONVICTION
The AI email assistant market is saturated on the surface, but nearly untapped on the layer that actually matters for a professional: fidelity to personal writing style and relational depth with each correspondent. A generic reply isn't a reply — it's a draft you have to rewrite. Our bet: the value isn't in the AI itself, but in everything that surrounds it (style learning across 800 emails, per-contact relational profile, thread context, native Outlook integration).
🛠️ TECHNICAL POSITIONING
- Multi-tenant architecture designed for scalability from day one (strict per-user data isolation)
- Anthropic Claude as the primary engine (superior quality/consistency over GPT-4 for tone preservation)
- Sovereign hosting in France (OVH Gravelines) — GDPR compliance isn't a feature, it's a prerequisite for our target personae
- Native Office.js add-in, not a parallel extension — the experience stays 100% integrated in Outlook
🎯 WHAT I'M LOOKING FOR AT THIS STAGE
I'm open to exchanges on three specific topics with those of you who have direct experience:
1. Early-stage B2B distribution: before scaling marketing, what "content / community / targeted outbound" mix worked for you to reach your first 50 paying customers?
2. Microsoft AppSource: any feedback on the validation process and its actual impact on acquisition?
3. B2B SaaS pricing for individual productivity tools: data points on price sensitivity between €19 and €49/month depending on persona (freelancer vs executive vs team)?
Thanks in advance for your insights.