Hannes Rapp
AI automation: n8n, chatbots, email & content pipelines
New to Contra
Hannes is building their profile!
Internal knowledge chatbot at Forterro, built from scratch in Python: Rasa framework, FAISS embeddings, semantic search with source citations, bilingual German and English, feedback loop with bad-case review and knowledge monitoring. Proof of concept 2024–25 that was well received internally and recognised by the CEO and the Head of AI.
1
22
Enterprise AI at Forterro (ERP group, 2,500+ staff): I built a two-agent Microsoft Copilot Studio system for Sales — a Cloud Sales Assistant plus a Sales Coach sub-agent — and tested it against a 2,600+ page Cloud knowledge base on source citations, hallucination, FAQ approval flow and sales usability. Basis of my B.Eng. thesis (2026): a comparative analysis of AI assistance systems in the enterprise, productivity and cost-benefit potential. Finding: no single approach leads on every dimension — enablement decides.
1
11
Content autopilot for a podcast: long-form episodes are cut into clips by the OpusClip API, Claude picks the next episode and schedules one post per day on YouTube, TikTok and Instagram. Since 10 Sep 2026: 42 posts published or scheduled, a weekly review loop instead of per-clip approval, manual editing time per clip: 0 minutes.
1
10
Inbox and lead autopilot for my own company: LinkedIn messages and reactions are read and sorted, replies are drafted with free slots from my calendar, every mail account lands in one watched inbox, and lead lists are built from public job postings and scored by budget signals. Nothing goes out before I say yes. First run: 2 reactions on a post, 2 meetings booked.
1
9