The problem
A production/photo studio wanted its senior leaders to actually adopt AI in day-to-day operations, not just hear about it.
A digital-marketing consultant needed a precise, low-volume way to find and prioritize the handful of accounts worth pursuing each cycle.
The approach
For the studio, ran an in-person, hands-on executive workshop backed by a minute-by-minute facilitator script and a shared live build, mapping AI to real repetitive work across its business units.
Shipped a suite of clickable AI prototypes so leaders could see concrete, buildable products against their own workflows.
For the consultant, built an assisted B2B prospecting pipeline that gates prospects on hard criteria, then scores surviving accounts on fit and timing for a non-technical coordinator to act on.
What I built
- A 90-minute workshop program: presenter deck, facilitator run-book with copy-paste prompts, prep/logistics checklist, and participant intake via a Google Form
- Four self-contained, branded HTML prototypes — order-triage support agent, auto-cull/retouch studio, marketing-content generator, and B2B lead/quote assistant — each mapped to a real data source
- Training handouts (prompt cheat sheet, LLM comparison, ways-to-work guide, opportunity map) and narrated, voiced webinar walkthroughs with rendered video
- An assisted prospecting pipeline in Python with pass/fail gates plus separate Fit and Timing scores, a runnable prototype producing a ranked HTML/CSV shortlist, and a live browser configurator to tune gates and weights
- Enrichment APIs called pay-as-you-go in batches, sized for a precision, low-volume need
- Claude API
- Python
- HTML/CSS/JS prototypes
- Google Forms API / Apps Script
- ElevenLabs (narrated webinars)
Result
Leaders left the studio engagement with a shared mental model and prototypes tied to their own workflows, while the consultant gained a repeatable, tunable way to surface and rank a short list of high-fit, well-timed accounts.