Purpose: Run AI-guided demos for tobacco/vape store prospects without needing their real stack or data. You generate fake historical data, load it into a live demo app, and walk the prospect through a simulated "before vs after" experience.
1. How We Generate Data
URL: zeidan.ai/sales-demo
- Pick Tobacco / Vape Stores (only vertical for now).
- Pick a scenario: Inventory Management, Customer Contacts, or Promo Performance & Reorder.
- Enter business name (for labeling), history window (days), and row count.
- Click Generate → the page builds synthetic historical rows (dates, SKUs, units, risk flags, etc.).
- Click Download JSON → save the file. This is what you load into the live demo app.
Why this works: We don't need the prospect's POS, spreadsheets, or CRM. We create believable fake data so the demo is self-contained and repeatable.
2. How We Load Data Into the App
URL: zeidan.ai/sales-demo/live
- Enter Business Name (e.g., prospect's store name).
- Enter Prospect Website URL (optional but recommended — the AI uses it for context).
- Select Scenario — must match the scenario used when you generated the JSON.
- Either Upload the JSON file from step 1, or Paste the JSON into the text area.
- Click Run 5-Minute Demo.
Important: The scenario you select must match the dataset. If you generated "Inventory Management" data but select "Customer Contacts," the app will block the run and show an error.
3. How the App Works
After you click Run 5-Minute Demo, the app:
- Sends the dataset + website URL to our AI backend.
- Generates a step-by-step demo script (talk track, AI actions, impact, evidence).
- Renders three panels that autoplay:
- Simulated Owner Alerts — What the owner would see in their inbox (alerts, prompts). These are data-specific (e.g., "31 high-risk inventory rows detected. Highest pressure SKU: POD-MENTHOL.").
- Simulated KPI Movement — Baseline metrics from the dataset, plus projected improvement (e.g., "Risk reduced by 7.0%", "Loss reduced by $2,200").
- AI Action Feed — Your narration script. Each step has: Talk Track (what you say), AI Action (what the assistant does), Impact (business outcome), Evidence (data backing).
- Replay — You can click Replay Timeline to run the autoplay again.
4. 60-Second Sales Script
"Here's your baseline from 6 months of synthetic store data. These are the alerts your team would receive. This is what the assistant does automatically. This is the business impact. We start with one workflow, prove value, then expand."
5. What Still Needs Tweaking
- Owner alerts and KPIs are driven by formulas and heuristics — not accounting-grade. They're for demo storytelling, not financial reporting.
- The live demo is tobacco/vape only. Construction and other verticals are not yet wired up.
- Website URL fetch can fail (CORS, timeouts) — the demo still runs, but AI has less context.
- No presentation mode yet — you're showing the same screen you use to configure. A full-screen "client view" could be added later.
6. Why This Should Make Demos Easier
- No stack discovery — You don't need to figure out their POS, CRM, or spreadsheets.
- No real data — Synthetic data keeps the demo safe and repeatable.
- One workflow — Focus on one scenario (inventory, contacts, or promo) per demo.
- Clear story — Baseline → Alerts → AI actions → Impact. The app structures the narrative for you.
Quick Links
| Step | URL |
|---|---|
| Generate dataset | zeidan.ai/sales-demo |
| Run live demo | zeidan.ai/sales-demo/live |