Every freelancer and agency I know sends quotes as PDF attachments. Then they wait. No read receipts. No signal. The follow-up call happens either too early (annoying) or too late (they already signed with someone else). This is the blind spot I'm trying to fix with Proposio — a quote tracking tool that turns a PDF into a trackable web page with real-time open notifications.
But here's the thing: I'm not building it yet. I'm validating whether the problem is painful enough that people will pay EUR 9/month to solve it. This post is about that process.
The Problem: Quotes Disappear Into Email Inboxes
A typical B2B quoting workflow in the DACH region looks like this:
- You write a quote in lexoffice, sevDesk, Word, or some ERP system
- You export it as PDF
- You attach it to an email and hit send
- You wait
That's it. The quote leaves your system and enters a black hole. You don't know if the prospect opened it. You don't know if they forwarded it to a decision maker. You don't know if it's sitting unread in a spam folder.
The default follow-up strategy is "wait a week and call." Sometimes that works. Often it doesn't — the prospect reviewed your quote on Tuesday, had questions, forgot about it by Friday when you called. Or worse: they compared it with a competitor who followed up within hours of the first viewing.
This isn't a niche annoyance. Any service provider sending 10+ quotes per month loses deals to bad timing. Not bad pricing, not bad work — bad timing.
Why Not Use Existing Proposal Tools?
PandaDoc, Better Proposals, Proposify — they all exist. They all track opens. So why build another one?
Because they solve a different problem. Those tools want you to build your proposal inside their editor. They replace your quoting workflow. For a 5-person marketing agency in Munich that already has lexoffice dialled in and generates 30 quotes a month from it, rebuilding every quote in PandaDoc is a non-starter.
The gap I see: a tool that works with your existing PDF. Upload the quote you already created in whatever system you use, get a trackable link, know when it's opened. No workflow change. No learning curve. No proposal editor.
That's the Proposio thesis: don't replace the quoting tool, add a tracking layer on top of it.
Validation Before Code
I scored Proposio at 58/100 in an internal validation framework before deciding what to do with it. That's a "conditional go" — interesting enough to investigate, not strong enough to start coding.
The weakest dimension was competitive moat. "Simpler and faster" is a positioning statement, not a moat. PandaDoc could add a "just upload your PDF" feature tomorrow. So the product needs to earn its right to exist through validation, not assumption.
What Phase 0 Looks Like
The validation phase has four parts, running in parallel over about four weeks:
Landing page + waitlist. A single page at proposio.app with the value proposition, planned pricing, and a waitlist signup. The page is live now. It shows the concept clearly: upload PDF, get AI-extracted data, share a trackable link, get real-time notifications. The planned pricing — EUR 0 free tier with unlimited quotes (view count only) and EUR 9/month Pro with full analytics and notifications — is visible. Pricing is part of the test.
Typebot validation form. Instead of a plain email signup, I use a conversational form that asks structured questions: how many quotes per month, how they currently track opens (most don't), whether they've lost deals to bad follow-up timing, and what they'd pay for visibility. This gives me signal beyond "someone clicked a button."
Problem interviews. 10–15 conversations with agency owners and freelancers who send 10+ quotes per month. Not pitch calls — problem interviews. The goal is to hear them describe their follow-up process and where it breaks down. If 7 out of 15 confirm active pain and willingness to pay EUR 10+/month, that's a go signal.
AI extraction side-test. Collect 20–30 real German PDF quotes from different sources (lexoffice exports, sevDesk, custom Word templates) and test whether a vision model can reliably extract line items, pricing, and customer data. If extraction accuracy is above 60% with fewer than three corrections needed, the AI angle holds up. If not, manual entry still works — it's just less compelling.
Go/No-Go Thresholds
I wrote these down before starting, so I can't move the goalposts:
- Go: 7+ of 15 interviewees confirm pain and willingness to pay, 50+ Typebot completions with 40%+ saying they'd pay EUR 10+/month, 30+ waitlist signups
- Pivot: Interest exists but pricing resistance or different pain points emerge
- No-go: Fewer than 5 of 15 show interest, fewer than 20 Typebot completions despite 500+ visitors, dominant response is "I just call them"
The no-go scenario is a real possibility. Many service providers have developed a habit of following up by gut feel, and "good enough" beats "optimised" in plenty of markets. If that's the answer, I'll know within four weeks instead of twelve.
What The Product Would Do (If Validated)
Assuming Phase 0 hits the go thresholds, the MVP scope is deliberately narrow:
Upload a PDF quote. Drag and drop, stored on EU-hosted object storage (Scaleway, Paris).
Manual data entry first. A structured editor for customer details, line items, milestones, notes. AI extraction comes later — Phase 3 in the roadmap, not the MVP. The reason: if users won't use the product with manual entry, AI extraction won't save it. Manual entry is the honest test of core value.
Publish as a trackable link. A signed URL that shows the quote as a clean, mobile-friendly web page. The recipient can view the structured proposal, download the original PDF, and ask questions through a built-in form.
Real-time "quote viewed" notification. This is the aha moment. You send a quote, you go about your day, and two hours later you get an email: "Müller Digital GmbH just opened your quote." That's the signal to follow up while you're top of mind.
Activity timeline. Every view, every download, every question — logged with timestamps. Free tier gets view count only. Pro gets the full timeline plus notifications.
Dashboard with "needs attention" view. Quotes expiring soon, never-viewed quotes, unanswered questions — surfaced in one place.
That's it for the MVP. No AI extraction. No team features. No ERP integrations. No custom branding. All of those are scoped for later phases, gated behind actual user demand.
The Technical Angle: GDPR-Native, Not GDPR-Retrofitted
Proposio targets the DACH region. GDPR compliance isn't a checkbox — it's a design constraint.
The entire stack is EU-hosted:
- Database: Neon PostgreSQL in Frankfurt
- File storage: Scaleway Object Storage in Paris
- AI processing: Scaleway Generative APIs in Paris — open-weight models, data not used for training
- Hosting: Vercel FRA region
No data leaves the EU at any point. IP addresses are hashed before storage. The AI extraction pipeline (when it ships) runs on EU-hosted Mistral models through Scaleway, not through OpenAI or Anthropic APIs where data routing is less transparent.
For DACH agencies sending quotes with client financial data, this matters. "Where does my data go?" is a real question their compliance teams ask. The answer being "Frankfurt and Paris, nowhere else" is a competitive advantage over US-hosted alternatives.
Why Validation-First Matters for Solo Builders
I run CodeAttack as a solo operation. Every week I spend building a product that nobody wants is a week I'm not spending on client work or on products that do have traction.
The temptation with Proposio was strong. The problem felt obvious. The technical architecture was clear. I could have started coding immediately and had an MVP in 8 weeks.
But "feels obvious" is not the same as "validated." I've seen enough internal tools and side projects die because the builder was convinced the problem was universal, when in reality it was their personal pain point dressed up as a market.
Four weeks of validation costs almost nothing. A Typebot instance I already self-host. A landing page that took a day. Some LinkedIn posts. Some conversations. If the answer is no, I saved myself two months. If the answer is yes, I start building with conviction instead of hope.
What The Live Site Shows Today
The Proposio landing page is live with the full value proposition: AI-powered quote extraction, real-time tracking, follow-up recommendations. It shows two interactive demos — one for how the client sees the proposal, one for the sender's analytics view. Pricing is transparent: free tier with unlimited quotes and basic stats, Pro at EUR 9/month with AI extraction, real-time notifications, and lexoffice/sevDesk integration.
The primary CTA leads to a validation chat (the Typebot form). There's also a simpler "notify me at launch" email capture for people who want less friction.
The page positions Proposio as "from PDF to closed deal" with three pillars: AI reads your quote, real-time engagement signals, data-driven close recommendations. It emphasises EU hosting and GDPR compliance — all data in Germany and France, AI models that don't train on your data.
This is the validation artifact. It's designed to test whether the positioning resonates, whether the pricing feels right, and whether people care enough to leave their email.
What Comes Next
The validation window runs through early May 2026. I'll publish the results — including the numbers — regardless of the outcome. If Proposio gets the go signal, I'll write about the build. If it doesn't, I'll write about what I learned.
Either way, validation-first beats building-first. Every time.