Email classification, multi-agent systems and everything in between. I build automations that replace manual work with reliable, AI-driven workflows. Built on 100+ prototypes.
30 Min. · No Obligation · Incl. Automation Assessment
These are the three most common bottlenecks I see in initial consultations.
Staff spend hours copying data between systems, sorting emails by hand, and processing documents one by one.
AI workflows that handle data entry, classification, and extraction: reducing hours to minutes.
Manual processes lead to typos, forgotten follow-ups and inconsistent data.
Structured validation at every step: catch errors before they reach your database.
More orders mean more manual work. Your team hits capacity limits instead of focusing on higher-value tasks.
Automations that handle 10x the volume without adding headcount.
Project or retainer. You choose what fits.
One-time automation project
Ongoing support (monthly)
All prices excl. VAT. API costs (OpenAI, Claude) and tool licenses not included.
The specific tools and AI models I use: chosen per project, not per vendor deal.
Make, n8n or custom code: I choose the right tool for your workflow.
OpenAI, Anthropic, open-source models: for document understanding, text classification, and structured data extraction.
CRM, ERP, email, accounting: I connect your existing systems smoothly.
Classify incoming emails by intent, draft context-aware replies, and trigger follow-up reminders automatically.
OCR for receipts and invoices, PDF parsing, and structured extraction: tested at scale in BuchhaltGenie (71% accuracy improvement).
Sentry error tracking, uptime checks, and Slack/Discord alerts: so failures get caught before users notice.
Standard automation follows rigid if-then rules. AI automation reads context, classifies intent, and handles edge cases. Concrete result: 71% OCR accuracy improvement in BuchhaltGenie.
| Aspect | Standard Automation | AI Automation |
|---|---|---|
| Decisions | Rigid rules | AI-based, context-aware |
| Document processing | Template-based | OCR + AI (71% better) |
| Email processing | Keyword filter | Semantic analysis |
| Scaling | More rules = more effort | Learns and improves |
Decisions
Document processing
Email processing
Scaling
Four phases with fixed deliverables at each stage.
I map your current workflows, measure time spent per task, and rank automation candidates by effort-to-savings ratio.
1-2 daysTool selection (Make, n8n, or custom code), data flow diagrams, and API mapping. Deliverable: a technical spec you can review before implementation starts.
2-3 daysBuilding the automation, connecting APIs, integrating AI models, and running test cases against your real data.
2-4 weeksProduction deployment, Sentry monitoring setup, runbook handover, and 2 weeks of follow-up support included.
1-2 daysSounds interesting? Let's discuss your project.
Discuss Your ProjectReal projects with measurable results.
RAG system with 13,794 chunks for AI-powered content search.
OCR-based receipt capture with 71% improvement through AI training.
Automated content pipeline with 3 AI models and Quality Gate.
Typical examples: email classification, invoice processing (OCR), appointment scheduling, data import/export, report generation, follow-up reminders. I identify your best candidates in the initial consultation.
Depending on requirements: Make (formerly Integromat), n8n, custom code (Node.js/Python), OpenAI API, Claude API. I choose the right tool. No vendor lock-in.
Simple workflows: 2-3 weeks. Complex multi-agent systems: 4-8 weeks. I estimate effort for your specific project in the initial consultation.
API costs (OpenAI, Claude) are typically €10-100/month depending on volume. Tool licenses (Make, n8n Cloud) from €10/month. I optimize for cost-effective usage.
Yes, I hand everything over documented. For Make/n8n workflows, you don't need programming skills. For custom code, I recommend my retainer package.
götzendorfer.at