Case studies

Small automations. Numbers large enough to matter.

Lume Clinic

Maria Bell, founder

Problem: Lume Clinic had a front desk inbox that looked manageable until flu season. Staff were answering the same appointment, intake, billing, and insurance questions 80 times a day. The clinic did not need a chatbot on the website. It needed an assistant that could sort incoming email and prepare replies without making medical decisions.

What we did: We helped Maria's team map the four safest inbox categories, write approval rules, and train Inbox Triage on 300 past replies. The AI drafts answers for routine admin questions, flags medical topics for humans, and updates a simple tracking sheet so the team can see volume by category.

80 daily emails reduced to 12 manual replies

14 hours saved each week

73% of AI drafts approved without edits

We went from answering 80 emails a day to 12. The other 68 are handled by an agent we trained in the course. I got my mornings back.

Next steps: Lume is adding appointment reminder automation next, but only for patients who opt in. Medical triage stays with humans, where it belongs.

Rocco Ledger

Nate Rocco, solo accountant

Problem: Nate had tried free tutorials for four months. The issue was not motivation. It was that every video assumed a different stack and none started with his actual work. His recurring tasks were client reminders, statement collection, proposal follow-ups, and the small bits of admin that steal a solo operator's week.

What we did: Inside AI Automation Fundamentals, Nate picked one process: month-end document chasing. We built a simple workflow that checks missing files, sends a polite reminder, updates a client sheet, and drafts a second nudge when the first is ignored.

First workflow in production in 9 days

11 clients moved into automated reminders

Month-end admin dropped by 6.5 hours

I tried doing this on my own for 4 months with free YouTube tutorials and got nowhere. With Techi I had my first workflow running in two evenings.

Next steps: Nate is now using ProposalPress to turn discovery calls into fixed-scope bookkeeping proposals before prospects forget why they called.

Fern-Patterson Growth

Lucy Fern-Patterson, agency owner

Problem: Lucy's agency ran paid acquisition for B2B companies and handled inbound WhatsApp messages manually. Leads arrived after hours, especially from founders browsing ads late at night. By morning, intent had cooled and the team was sorting through threads instead of selling.

What we did: We built a transparent WhatsApp qualification bot that asks six questions, scores fit, offers calendar slots, and escalates unusual requests to a human. The agency used the Voice Agents course logic, adapted for chat: clear identity, limited scope, human handoff.

Meetings booked/month grew from 18 to 47

After-hours leads answered in under 30 seconds

Bot runtime cost stayed near $30/month

The WhatsApp bot we built in module 4 books meetings at 2am. That used to be impossible unless I hired someone in the Philippines.

Next steps: The team is adding CRM enrichment and post-call notes so every booked meeting arrives with context instead of a mystery calendar invite.