← All projects

Web app 2025

BLUNATECH

The marketing agency as a web app

BLUNATECH replaces the classic marketing agency with software: a planned marketing year, AI-written campaigns and editorial content, contacts and customer retention, all in one app built for small and medium businesses in Germany.

As CTO I own product and engineering, from the first line of code to production.

Status
Live
Role
CTO, product and engineering
Focus
Marketing automation for SMBs in Germany

Case study by Fabian Bitzer 4 min read

BLUNATECH is a marketing platform for small and medium businesses in Germany. It plans the marketing year, suggests fitting measures, writes campaigns and newsletters with AI, and brings contacts and customer retention together in one place. As CTO of BLUNATECH GmbH, I am responsible for product and engineering, from architecture to operations.

The problem: marketing without a marketing department

Most small businesses know they should market themselves regularly. In practice it fails for very concrete reasons. There is no time to write a newsletter every week. There are no ideas for what to write about. And when something does get written, it sounds generic. Tools like Mailchimp feel too complex to many owners, while an agency is expensive and creates dependency.

BLUNATECH takes a different approach: marketing without prior knowledge, without an agency and without leaving things to chance. The software takes over what an agency would otherwise do, namely planning, ideas and copy. Goals, data and decisions stay inside the business.

How BLUNATECH works

The path through the platform follows the logic of a good agency: understand first, then plan, then execute and keep going.

The marketing year: each planned impulse shows its category and sending date, and hovering one opens its detail.

Assessment: understand before planning

It starts with a short first assessment. Building on it, the platform produces a detailed report with 17 chapters that derives growth suggestions and priorities from the company’s answers. That makes it clear where marketing should begin before the first measure is planned.

The first assessment, chapter one, shown with example figures for a fictional café: what is typically possible for a company of this size and industry, as estimates for revenue potential, return, time saved, visibility and repeat contacts.

Onboarding: one profile that feeds every AI feature

Eight steps build a company profile: industry, identity, target audience, writing style, content strategy, tone and language, branding and practical settings. Every AI feature on the platform draws on this profile. That is exactly what prevents the generic copy people know from general-purpose AI.

Onboarding, step 7 of 8: branding. From existing copy the platform can derive the writing style.

Annual planning: the marketing year at a glance

At the core is a marketing calendar with year, quarter and month views. A library of more than 500 marketing measures provides the material. A custom scoring logic suggests fitting measures and weights them by industry fit, timing, topic priority and effort. With a communication focus, the business picks the one to three categories that matter most right now and steers the suggestions accordingly.

The measure library as a loop: a channel filter and a search narrow the eight categories down to what matches. The full library holds more than 500 measures.

Campaigns, editorial content and newsletters with AI

Planned measures turn into campaigns whose content the AI drafts from the company profile. In the editorial area, editorial prompts can be turned straight into newsletters. The newsletter builder offers a drag-and-drop editor with 15 block types, and an AI fast lane generates three variants of a newsletter at the click of a button.

The newsletter editor: a block is dragged from the block list into the stack. There are 15 block types in total.

Contacts, contact booster and customer retention

Contacts can be imported, maintained and exported. The contact booster creates a landing page with a QR code, for example for a trade fair, where interested people sign up with double opt-in. A dedicated module covers customer retention, meaning regular contact with existing customers.

The contact booster with double opt-in: a visitor signs up on the landing page, confirms by email, and only then lands in the contact list.

Demand analysis and market monitoring

What are people in the industry interested in right now? A demand analysis based on Google Trends data and a market monitoring feature help to pick topics at the right moment. Achievements and levels make regular use visible.

Who BLUNATECH is built for

BLUNATECH is aimed at businesses and organisations without their own marketing department. The platform names 21 sectors, among them medical practices, skilled trades, retail, hospitality, tourism, service providers and consultancies, as well as associations, foundations, schools, chambers, NGOs, academies, architecture and engineering firms, start-ups and the creative industries.

My role as CTO

Since June 2025 I have been CTO and technical lead of BLUNATECH GmbH. I am responsible for building the platform and its AI solutions, design AI workflows, web apps and API integrations, and translate business requirements into technical concepts and roadmaps. That includes feasibility assessments, project planning and the implementation itself, from the first data model to deployment.

The architecture behind the platform

BLUNATECH is a web app on a lean stack that is easy to reason about:

  • Frontend and server: Next.js with the App Router and Tailwind CSS, with Zustand and TanStack Query on the client. Hosted on Vercel in the Frankfurt region.
  • Data and authentication: Supabase with Postgres, authentication and row level security. Every business sees only its own data.
  • AI: Text generation runs through OpenRouter with a fast Gemini model. Which measure fits when is decided by the platform’s own scoring logic, not by the language model.
  • Background jobs: Longer generations and recurring tasks run with Trigger.dev outside the request itself, so the interface stays fast.
  • Email, payments, operations: Email through Resend, payments through Stripe, and errors and usage monitored with Sentry and PostHog in the EU.

The public website at blunatech.com is a separate project built with Astro and React.

What I learned

Good AI copy starts with good context. A language model that knows nothing about the business writes generic text. Only the detailed profile from onboarding makes the results usable. That is why onboarding is not a form to get through but the foundation for everything that follows.

Suggestions instead of an empty calendar. An empty marketing calendar is as overwhelming as an empty text field. So the platform starts with concrete, scored suggestions that can be accepted, moved or dropped.

Fewer options, clearer paths. Small businesses do not need two hundred settings. They need a clear next step. Every new feature has to prove that it makes that step easier.

Try BLUNATECH

BLUNATECH comes with a free seven-day trial and can be cancelled monthly. All details are on blunatech.com (German). I am happy to answer questions about the technology or the platform directly at kontakt@bitzer-fabian.de.

Next project

BerichtsheftKI

AI-written training reports for apprentices. iOS and Android.