A bite-size video project by Ed Crotty, Claude Certified Architect at Datacap Systems, Inc.
This is the complete behind-the-scenes plan for a short video (3-5 minutes) where I demonstrate five AI capabilities that most people have no idea exist today. Not in some distant future — right now.
The twist? This entire project — the strategy, the script, the production plan, the social media rollout — was designed collaboratively with AI. You're looking at the output of that collaboration. The video itself will also be edited by AI using a tool called Cardboard.
So really, it's a demonstration of AI capabilities all the way down.
Each demo uses a different AI platform to show the breadth of what's available today.
graph LR
A["1. Photo"] -->|Gemini| B["AI Explains It"]
C["2. Documents"] -->|Claude| D["Instant Report"]
E["3. Sketch"] -->|ChatGPT| F["Professional Image"]
F -->|Gemini Veo| G["4. Video"]
H["5. Pinpad on Camera"] -->|Gemini Live| I["AI Debugs It"]
style A fill:#4285F4,color:#fff
style B fill:#4285F4,color:#fff
style C fill:#D97706,color:#fff
style D fill:#D97706,color:#fff
style E fill:#10A37F,color:#fff
style F fill:#10A37F,color:#fff
style G fill:#4285F4,color:#fff
style H fill:#4285F4,color:#fff
style I fill:#4285F4,color:#fff
| # | What I'm Doing | AI Platform | Why It's Impressive |
|---|---|---|---|
| 1 | Taking a photo of something and asking AI to explain it | Gemini | Instant identification — no Googling, no manuals |
| 2 | Uploading multiple documents at once and getting a full analysis | Claude | An hour of work done in 10 seconds |
| 3 | Drawing stick figures on paper and having AI create a real image | ChatGPT | Napkin sketch becomes presentation-ready |
| 4 | Taking that image and turning it into a video | Gemini Veo | Paper drawing → image → video. Three steps. |
| 5 | Pointing a phone camera at a payment pinpad and debugging it live | Gemini Live | AI sees your device and talks you through the fix in real-time |
Demo 5 is the showstopper. Every retail tech professional has been on hold with support trying to describe what's on a pinpad screen. This demo replaces that entire experience.
The production workflow is designed to be fast and modular — record each demo separately so if one needs a redo, you only redo that piece.
flowchart TD
PREP["Prep Day\nGather docs, draw sketch,\nset up pinpad, dry run"] --> REC["Recording Day\n~30-45 min total"]
REC --> D1["Screen Record\nDemo 1: Photo → Explain"]
REC --> D2["Screen Record\nDemo 2: Files → Report"]
REC --> D3["Screen Record\nDemo 3: Sketch → Image"]
D3 -->|Save the image| D4["Screen Record\nDemo 4: Image → Video"]
REC --> D5["Phone Camera\nDemo 5: Pinpad Debug"]
REC --> TH["Talking Head\nIntro + Outro"]
D1 & D2 & D4 & D5 & TH --> CB["Upload to Cardboard\nAI assembles, trims,\nadds captions"]
CB --> FULL["Full Video\n3-5 minutes"]
CB --> CLIPS["5 Individual Clips\n30-60 sec each"]
FULL --> LI["LinkedIn\nYouTube"]
CLIPS --> SM["Social Media\nDrip Campaign"]
style PREP fill:#f0f0f0,color:#333
style CB fill:#6C5CE7,color:#fff
style FULL fill:#00B894,color:#fff
style CLIPS fill:#00B894,color:#fff
Everything needed to execute this project lives here.
A set of bullet points for each demo — not a word-for-word script, but enough structure to stay on track while talking naturally. Includes a recording checklist so nothing gets missed.
The detailed plan covering video structure, timing for each segment, demo-by-demo technical specs, what to prep, retry protocol if something goes wrong, and a day-by-day timeline.
Ready-to-use prompts for Cardboard (the AI video editor) that tell it exactly how to assemble the raw clips into a finished video. Includes tips from the Cardboard team on picture-in-picture overlays, search-by-content, and pacing.
assets/sketch/— Photos of the hand-drawn sketch (input for Demo 3)assets/docs-for-upload/— The documents used in Demo 2assets/outputs/— AI-generated images and videos from the demosrecordings/raw/— Raw screen recordings and phone clipsfinal/— The finished video and individual clips
This project deliberately showcases three different AI platforms to demonstrate that AI is not one tool — each platform has different strengths.
graph TD
subgraph Claude["Claude (Anthropic)"]
CL1["Long document analysis"]
CL2["Multi-file comparison"]
CL3["Structured report generation"]
end
subgraph ChatGPT["ChatGPT (OpenAI)"]
GP1["Image generation from text"]
GP2["Sketch-to-image transformation"]
GP3["Creative visual output"]
end
subgraph Gemini["Gemini (Google)"]
GE1["Photo identification"]
GE2["Video generation via Veo"]
GE3["Live camera + conversation"]
end
style Claude fill:#D97706,color:#fff
style ChatGPT fill:#10A37F,color:#fff
style Gemini fill:#4285F4,color:#fff
Knowing which tool to use for which job is the real skill — and that's one of the key messages of this video.
| Phase | What Happens |
|---|---|
| Day 1 — Prep | Gather documents, draw the sketch, set up the pinpad, do a dry run of the live camera demo |
| Day 2 — Record | Record all 5 demos + talking head intro/outro (~30-45 min total) |
| Day 2-3 — Edit | Upload everything to Cardboard, review the AI-edited result, request tweaks |
| Day 3-4 — Finalize | Export the full video + 5 individual clips |
| Day 4-5 — Publish | Post to LinkedIn, YouTube, share with RSPA |
The retail technology channel is at an inflection point with AI. Most VARs, ISVs, and partners have heard the buzz — but many haven't seen what AI can actually do with their own eyes.
This video is designed to change that in under 5 minutes. Each demo was chosen to be:
- Immediately understandable — no technical jargon, no setup needed
- Directly relevant — document analysis, visual mockups, and live device debugging are things this audience does every day
- Surprising — even people who use AI regularly don't know about many of these capabilities
The pinpad debugging demo in particular speaks directly to a pain point every person in this industry knows: trying to troubleshoot a device over the phone. Showing AI doing that visually, in real-time, is the kind of moment that shifts perspectives.
Ed Crotty is a Claude Certified Architect at Datacap Systems, Inc. He is an active contributor to the RSPA (Retail Solutions Providers Association), including authoring content on AI misconceptions and best practices.
This project was designed collaboratively with Claude (Anthropic's AI) using Claude Code, demonstrating that AI is not just a tool for the final product — it's a tool for the entire creative process, from brainstorming to planning to production.
Built with AI. About AI. Edited by AI.