Capstone 6: A No-Code AI Product
Build and launch a lead-qualification assistant without writing code — intake form, AI scoring, automated follow-up, and a review queue.
Every small business with a contact form has the same quiet problem: inquiries arrive at random hours, a third of them are spam or poor fits, and the good ones go cold while someone finds time to read the inbox. A lead-qualification assistant fixes that. It reads each inquiry the moment it arrives, scores it against your criteria, drafts a reply, and puts it in front of a person to approve.
In this capstone you'll build one without writing code, for a fictional three-person web design studio. You'll write a scoring rubric, build an intake form and a leads table, add an AI scoring step to an automation, draft follow-ups that a human approves before anything is sent, and test the whole thing with deliberately awkward leads before real prospects see it. The same pattern works for any business that qualifies inbound requests: agencies, consultants, B2B sales, recruiting, even rental inquiries.
What you'll build
The studio builds marketing sites for small businesses and gets 10 to 40 inquiries a week. After this build, every inquiry takes this path:
Prospect fills in intake form
│
▼
Leads table (new record, Status = New)
│ automation 1
├─► instant templated acknowledgement email (no AI)
├─► AI scoring step ──► sub-scores, summary, flags (JSON)
│ └─ invalid output or error ──► Status = Error, alert the owner
├─► formulas compute total score and tier (Hot / Warm / Not a fit / Spam)
├─► AI drafting step (Hot and Warm only) ──► Draft reply field
└─► Status = Needs review
│
▼
Review queue (a person reads, edits, and approves)
│ automation 2: Status changes to Approved
└─► send the reviewed email ──► Status = SentFour kinds of tool make this up, and you can swap any of them for an equivalent:
| Part | What it does | This build uses | Equivalents |
|---|---|---|---|
| Intake form | Collects structured answers plus one free-text field | Airtable form | Google Forms, Typeform, Jotform |
| Database and review queue | Stores leads, computes scores, hosts the review screen | Airtable base with an interface | Google Sheets, Notion database |
| Automation platform | Runs the steps when a record appears or changes | Zapier, Make, or n8n | Any of the three |
| Model | Scores and drafts | Claude through the platform's integration | Any capable chat model |
It applies AI Automations with Zapier, Make, and n8n, AI in Spreadsheets and No-Code Databases, and Shipping No-Code AI Tools Safely. If a step feels unfamiliar, those lessons cover it in more depth.
Prerequisites: an Airtable account (or a Google account if you use Sheets), an account on one automation platform, access to a model through that platform (either its built-in AI step or your own API key connected to it), and an email account the platform can send from. Free tiers are usually enough to build and test; check each tool's plan limits before you go live, because automation platforms count every step that runs.