Unified AI Workspace
Tenders, proposals, and content produced manually with quality varying by individual. We built a structured AI workspace connecting a 15-file knowledge base to 18 workflow skills across four business domains.
The problem
Tenders, proposals, and marketing content were all produced manually. Quality varied depending on who was available. Institutional knowledge lived in people’s heads - when someone was unavailable, the work either stalled or was done at lower quality by someone else.
What we built
An AI workspace where a structured knowledge base connects to a set of workflow skills.
- 15 authoritative knowledge base files covering company profile, services, pricing, brand voice, and proof points
- 18 workflow skills across four domains: tender evaluation, proposal generation, marketing content, lead outreach
- Branded document generation (.docx, .pptx, .xlsx)
- Team access via Claude Projects for members without a technical setup
Stack: Claude Code · Claude Projects · Python · n8n · python-docx · openpyxl · python-pptx
What changed
Tender review time dropped from 3-5 hours to 20-40 minutes. Proposal preparation, content production, and lead outreach all run from the same knowledge base. The team has a consistent, repeatable process that doesn’t depend on any one person’s availability. When the knowledge base is updated, every output type improves simultaneously.