Long-Term AI Partner (Architect + Builder)

We are a 14-person professional services firm based in Germany. We’ve been delivering the same core consulting service profitably for 12+ years. Our frameworks, client patterns, and delivery know-how live mostly in people’s heads, slide decks, Notion, Google Drive, and email - not in a system our tools can reliably use. Over the last year we started building our first digital product: a client-facing tool that packages part of what we used to deliver manually in workshops. We have a working beta (built with Lovable + Supabase + Claude API). A small group of clients is using it. The service business is not going away - the product is meant to extend what we already sell, not replace the firm overnight. Where we are stuck - Everyone has ChatGPT / Claude access, but usage is inconsistent. Quality varies a lot on client-facing outputs. - Our “AI product” still draws from static prompts and a handful of reference files, not from our real accumulated expertise. - Client onboarding into the tool is still copy-paste. Each new account feels bespoke. - We don’t have a shared knowledge layer that research, delivery, and the product all pull from. - We need someone who can architect the long-term foundation and build the first practical pieces with us - not hand us a strategy deck and disappear. What we want (2026–2027) A long-term technical partner (not a one-off automation gig) to help us: - Audit how knowledge and workflows actually move today (Drive, Notion, Slack, CRM, call notes). - Design a scalable AI + data architecture: what lives in the database, what agents do, what stays deterministic, how non-technical staff interact safely. - Build a first production-ready pipeline (one high-value workflow end-to-end) that proves the approach. - Set up foundations so we can add more agents / products from the same core without rebuilding each time. - Help our team become progressively more autonomous on the surface (prompts, configs, Lovable UI) while you maintain the structural layer (backend, security, integrations, AI orchestration). Tech we use or are open to - Supabase (Postgres, auth, RLS) - already in use - Lovable for product UI - business team can iterate there - Claude API - primary LLM today - n8n or similar for workflow glue - open to your recommendation - Integrations: Google Workspace, Slack, HubSpot (CRM), Fireflies (call notes) We prefer practical, maintainable systems over enterprise complexity. This is NOT - A “build us a chatbot widget” project - Prompt-engineering-only consulting - A greenfield startup with no operating business behind it - A role where you only advise and our team implements everything - We need someone who thinks architecturally and ships code. Engagement - Start with a paid discovery + prototype phase (3–4 weeks), then ongoing retainer if it’s a fit - 10–20 hrs/week to start, likely growing - Direct access to the founder and head of delivery - no agency layers Please include in your proposal - One production LLM system you shipped end-to-end - what it does, what data it uses, one measurable outcome. - How you’d approach the first 30 days with a firm like ours - What you’d prioritize first and why - High-level architecture in plain language - One example of enabling a non-technical team to keep building safely while you own the core - Availability, timezone, and earliest start date Bonus if you have - Experience productizing consulting / professional services expertise into software - Multi-agent or structured knowledge systems - Supabase + Lovable (or similar low-code frontend) in production