Skip to content
dilthey.media

// SERVICES

AI, automation and software, built by one pair of hands

I build and run the systems I recommend. No PowerPoint. Every service follows the same pattern: your problem, what I build, the result you can expect, and the proof behind it.

// 01 · CUSTOMER SERVICE

Customer-service automation

The problem

Your support answers the same questions every day. Requests pile up, wait times grow, and your team never gets to the real work.

What I build

I build chatbots and RAG knowledge bases that answer from your own content: on your site, in your customer portal, or in WhatsApp. Grounded answers only, no hallucinations.

Typical outcome

Most routine requests get answered automatically, around the clock. Your team only handles the real cases.

Proof

For DialogAi I turned WhatsApp voice messages into text automatically and summarized them with AI. Communication without manual listening.

// 02 · PROCESS & DOCUMENTS

Process & document automation

The problem

Invoices, forms, data entry: manual back-office work costs hours, ties up staff, and creeps in errors.

What I build

I automate whole workflows: from document recognition (handwriting included) through validation to handover into your existing systems like SAP.

Typical outcome

Recurring processes run automatically and traceably. Your team wins back time for work that needs humans.

Proof

For a German law firm I automated 93% of invoice creation. For a public-sector organization I built software that automatically recognizes handwritten applications and transfers the data directly into the SAP system.

// 03 · GENERATIVE AI

Generative AI: advice & delivery

The problem

You know generative AI could help, but not which use case is worth it, or how to turn the idea into a finished system.

What I build

Together we find the use cases with the best effort-to-return ratio and I ship them: from opportunity assessment to a finished system that runs in daily use.

Typical outcome

You invest in the AI projects that pay off, not experiments that end up in a drawer.

Proof

For DoubleYou.ai I built an AI chat that captures the persona of YouTube creators. Generative AI as a real product, not a demo.

// 04 · SALES

Sales automation

The problem

Leads sit unworked, follow-ups slip, and your sales team spends too much time on data entry instead of selling.

What I build

I automate lead qualification, follow-ups and CRM enrichment, connected to the tools you already use.

Typical outcome

Your sales team talks to the right leads at the right time. Less manual work, faster response times.

Proof

Built on the same AI pipelines as the automation projects above, wired into your existing systems.

// 05 · MARKETING

Marketing automation

The problem

Product copy, channels, variants: creating and maintaining content at scale is barely possible by hand.

What I build

I build LLM pipelines that create content and product data automatically and optimize it for your channels: shop, Amazon, eBay.

Typical outcome

More high-quality content in less time, consistent across every channel.

Proof

For the Markus Molitor winery I optimized 6,000+ product descriptions via an LLM pipeline, result: +5.3% online-shop revenue. For Wenko, 6,000+ descriptions for Amazon and eBay.

// 06 · HR

HR automation

The problem

Screening applications, capturing forms, processing onboarding documents: HR processes are document-heavy and manual.

What I build

I automate the processing of application and personnel documents, including recognition of handwritten and scanned forms.

Typical outcome

Your HR team spends less time on paper and more time with people.

Proof

The same software that captures handwritten applications in the public sector transfers directly to HR documents.

// 07 · SOFTWARE

Full-stack software development

The problem

Sometimes you don’t need an off-the-shelf tool but software that fits your workflow exactly: ready to use, not a prototype.

What I build

I develop full-stack software and AI products completely: frontend, backend, AI integration and the operations behind it.

Typical outcome

You get ready-to-use software that fits your processes and that you own.

Proof

DialogAi (WhatsApp voice-to-text) and the SAP integration for a public-sector organization are fully built systems in use, not concepts.

// ORIENTATION

Not sure which you need?

That’s normal. You know your problem, not the AI taxonomy. That’s exactly what the intro call is for: we look at your case together and I tell you honestly what’s worth it and what isn’t.

// TECH

Technically real, not just slides

I deliver finished systems that run in daily use and integrate with your existing landscape: from LLM pipelines to RAG to SAP integration. Built GDPR-aware, documented and handed over. No jargon, no opaque technology.

What I build with

  • Anthropic
  • OpenAI
  • Python
  • TypeScript / Next.js
  • n8n
  • SAP integration
  • AWS / Azure

// OPERATIONS & SUPPORT

Systems I build, I also run and maintain on request as a monthly service, no lock-in.

// GET STARTED

Which of your problems do we solve first?

30 minutes, free, no obligation. We’ll frame your case and I’ll tell you what’s possible.

Reply within 24 hours on workdays.

Book a free intro call