Conceptual Template: Project Reference Silicon Planet
Structure of a Project Reference Card
1. Header Section
- Client: [Company name or anonymized, e.g. „Large Logistics Provider, Germany“]
- Industry: [e.g. Facility Management / Logistics / Insurance]
- Project Title: Short, concise name
- Subtitle / Tagline: One sentence capturing the core benefit
- Project Period: [Quarter/Year – Quarter/Year]
- Project Phase: [e.g. Phase 1 of 3]
2. Situation („The Problem“)
3–5 bullet points, concrete and quantified
- What was the pain point?
- What was the scale of the problem?
- Why had it not been solved before?
3. Solution („What We Did“)
Brief, non-technical description of the intervention
- What approach did Silicon Planet take?
- What was the methodological core?
- If applicable: in which steps / phases?
4. Results („What It Delivered“)
Numbers, numbers, numbers — the heart of the reference
- Primary KPI (quantified, before/after)
- Secondary KPIs
- Timeframe of impact
5. Technology Stack
Short, scannable list — demonstrates competence without overwhelming
- Technologies / platforms used
- Integrations
6. Silicon Planet Scope
What exactly did SP deliver? (Demarcation from client’s own contribution)
- e.g. Concept design, architecture, implementation, change management support
7. Optional: Quote / Testimonial
Short verbatim quote from the client’s point of contact
Application: CWS — Phase 1, AI Classification of Customer Care Tickets
Client:CWS (or: Large B2B Service Provider, Germany — depending on approval) Industry: Facility Management / B2B Services Project Title: AI-Powered Ticket Classification in Customer Care Tagline: „From 15,000 to 2,500 tickets in the backlog — without adding headcount.“ Project Period: [insert] Phase: Phase 1 of [X]
Solution
Introduction of an AI-based automatic classification logic for all incoming Customer Care tickets — fully integrated into the existing Salesforce environment. No platform migration, no process disruption for staff.
Situation
- 15,000 tickets/month in the Customer Care backlog
- Manual classification and routing of all incoming requests
- Growing backlog despite full staffing levels
- Increasing quality loss due to prioritization errors and long response times
Results
| Metric | Before | After |
|---|---|---|
| Ticket backlog / month | 15,000 | 2,500 |
| Backlog reduction | — | –83% |
| Additional headcount required | — | 0 FTE |
| Equivalent FTEs saved | — | 20 FTE |
Technology Stack
- Google Cloud AI (classification model)
- Salesforce (integration / ticketing system)
- Custom Programming (business logic, routing, monitoring)
Silicon Planet Scope
- AI strategy & use case definition
- Technical architecture & system integration
- Implementation & go-live support
Scaling the Template
For the website, I recommend three levels of presentation:
- Tile view (overview): Client, industry, tagline, and top KPI only — for the first-pass scan
- Detail modal / subpage: Full reference card as above
- PDF download: Same structure as a branded one-pager for sales conversations
