Building [in public] an AI EA that actually executes (not another chatbot)
The Problem I'm Solving
I spent 4 years as a CTO at a digital health startup (raised 2 rounds, 90% market share in Southern Europe).
My biggest frustration? I spent more time managing work than doing work:
- 2+ hours/day triaging email
- 12 hours/week in meetings
- Constant context switching
- Following up with people manually
- Coordinating calendars like playing Tetris
I stepped down a few weeks ago to build the solution full-time.
What I'm Building
Nori, an AI executive assistant that autonomously handles admin work.
It's not another chatbot you have to prompt. Nori:
- Connects to Gmail, Calendar, Slack, Notion (via OAuth)
- Learns your communication style and priorities
- Executes on its own: drafts emails, resolves conflicts, preps meeting briefs, follows up
- Desktop + mobile + watch app (always present), although starting with desktop.
The difference: Most AI tools wait for you to ask. Nori acts proactively.
Current Status (Month 1)
Product:
- ✅ MVP built in 3 weeks (Electron + Next.js)
- ✅ 5 design partners using daily (CEOs, VPs, Heads)
- ✅ Public beta launched mid-November
Traction:
- 140 waitlist signups in first 48h (no paid ads)
- 30 user interviews completed
- All expressed high interest and needing this
Revenue: $0 (not charging yet)
Tech Stack
For the devs curious:
Frontend:
- Electron + React (desktop-first)
- Next.js App Router
- TypeScript + TailwindCSS
- shadcn/ui components
Backend:
- Firebase (auth, storage)
- ChromaDB (vector memory)
- OpenRouter (LLM gateway - Claude, GPT, DeepSeek)
- RAG for context retrieval
Integrations:
Starting only with Gmail/Calendar at the moment
Why this stack?
- Desktop-first = better UX for always-on assistant
- OpenRouter = model flexibility (not locked to one provider)
- ChromaDB = infinite memory without token limits
The Big Question
I'm a solo technical founder. My challenge:
How do I get my first 100 paying users?
My plan:
- Launch on Product Hunt (Done)
- Twitter outreach to founders tweeting about email pain
- YC/On Deck/Indie Hackers community building
- Cold email to founders (focused, not spammy)
But I'm worried about:
- Charging too early vs. too late
- Balancing feature requests vs. focus
- Building credibility as a 1-person team
What worked for you when launching your first product?
Especially curious about:
- Pricing strategy for early users
- How you handled design partner → paid user transition
- Community channels that actually converted
Following Along
I'll post weekly updates here with:
- Metrics (signups, activation, revenue)
- What I shipped
- What failed
- Key learnings
Building in public because:
- Accountability keeps me shipping
- This community gives honest feedback
- Transparency attracts early adopters
If you've ever felt buried in admin work, this one's for you.
Follow along: https://getnori.ai
Mati
