Pick a depth. Each prompt opens in your AI pre-loaded with the lesson. Click a row to preview the prompt.
An automation nobody can see is an automation nobody will pay for. The proof format that sells this work is short and rigid: a screen recording under a minute showing the flow doing its job for real, paired with a before/after write-up whose arithmetic a stranger can check. This matters now, not later, because your portfolio starts with this module's build — when a prospect asks 'have you done this before?', a 60-second video of a form submission turning into an enriched row and a Slack ping is a better answer than any credential. It also forces honesty on you: if you cannot demo the flow end to end in one take, it is not actually done, and better you discover that here than on a client call.
Record it with Loom (the free plan comfortably covers a one-minute screen recording and gives you a share link) or any screen recorder, using this shot list. Seconds 0-10: the filled-in form on screen, cursor hits Submit — say one sentence: 'A lead just submitted the inquiry form.' Seconds 10-30: the Leads sheet open, the new row appearing with First Name, Last Name, and Company Guess columns visibly populated — 'The automation cleaned the name and guessed the company from the email domain, no human touched this.' Seconds 30-45: the Slack #leads channel with the alert — 'And the owner got pinged eleven seconds after submission.' Seconds 45-60: a plain slide or doc with the math. One take, no editing, mistakes and all; a slightly rough real demo is more credible than a polished fake.
The write-up is four sentences, and here it is in full so you can adapt rather than invent: 'Before: each web lead was re-typed into the tracking sheet by hand and the owner found out whenever someone checked email — about 6 minutes per lead, roughly 90 leads a month, 9 hours of staff time, $225/month at a $25/hour loaded cost. After: the row appears automatically with the name cleaned and the company derived, and a Slack alert lands within about a minute of submission. Time saved: about 9 hours and $225 per month, roughly $2,700 per year. Evidence: run history shows 12 clean runs over 3 days, and the 60-second recording shows one end to end.' Numbers from your own task-2 math go in the blanks; the sentence naming the run history is what makes the claim auditable instead of aspirational.