agentleFS
BenPremBenPrem/business-brain/.claude/skills/ai-demo-builder/SKILL.md

ai-demo-builder

Skill1 starsChanged 3 months agoMITView on GitHub

Build working AI-automation demos tailored to a prospect's business to close automation sales — voice receptionist, review-request flow, lead follow-up bot, appointment reminders. Triggers - "build a demo for [prospect]", "AI demo", "automation demo", "voice receptionist demo", "review request demo". NOT for website builds — use website-builder.

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 4 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Don't explain what AI can do — show them.

    line 8
  2. Distinctive:

    Call-event webhook → serverless function: create the calendar event, send a confirmation, log the call summary to the workspace + <TASK …

    line 34
  3. Distinctive:

    Test 3 scenarios (hours question, booking, services list), screen-record, then share the number: "Call this."

    line 36
  4. Distinctive:

    Write-back: after any prospect demo or conversation, log an activity note on the prospect's workspace record and update the matching <TASK …

    line 85
  5. Very specific:

    If a demo asset is client- or public-facing and touches a regulated topic (financing offers, health claims, legal outcomes, insurance), run …

    line 90

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

AI Automation Demo Builder

Don't explain what AI can do — show them. Build a working demo in the prospect's own business context, let them interact with it live, then close. A demo they can call, text, or submit to beats any slide deck.

When NOT to use

  • Website builds → website-builder
  • Cold outreach copy → cold-email

Stack (map each need to YOUR live tools before building)

NeedUse
Scheduled or chat-triggered automationYour always-on agent box (if you run one), or a scheduled cloud agent
Webhook endpoint / form handling / API proxyServerless function (keep the source dir OUTSIDE the publish dir; origin-check + rate-limit anything public)
SMSYour SMS provider (e.g. Twilio)
Email sequenceYour email platform, or direct transactional send
AI generation inside a functionYour LLM router — validate model slugs against its live API first
Voice agentA voice-AI platform (e.g. VAPI) + a provisioned phone number

Demo types

1. AI Voice Receptionist

The prospect calls a number and talks to an AI that knows their business and books appointments.

  • Configure the voice agent with the business's name/hours/services/FAQ — scrape their site or ask the founder for the facts; never invent capabilities
  • Call-event webhook → serverless function: create the calendar event, send a confirmation, log the call summary to the workspace + <TASK SYSTEM>
  • Test 3 scenarios (hours question, booking, services list), screen-record, then share the number: "Call this."
  • Dependencies: voice-platform account + phone number. Not buildable until those exist — say so upfront.

2. Automated Review Request Sequence

Job marked complete → the customer automatically gets a perfectly-timed review ask.

  • Trigger: a message to your automation box or a webhook when a job completes
  • Sequence: immediate thank-you text → 2h review request with the direct Google review link → 3-day gentle follow-up if no review yet
  • SMS via your provider, email variant via your email platform; track asked/reviewed/response-rate in the workspace
  • Demo: walk the sequence using their real business name, city, and actual review link

3. Lead Follow-Up Bot

Form submission → instant personalized AI response + owner notification within 60 seconds.

  • Form → serverless function → LLM generates the personalized reply → email out
  • Owner notification (chat/text/email): lead details, an urgency-signal lead score, suggested talking points
  • Demo: a simple form page, submit a test inquiry live in front of them

4. Appointment Reminder System

Automated 24h (email + text) and 2h (text) reminders with confirm/reschedule replies; cancellations ping the owner immediately.

  • A scheduled job reads the calendar (calendar API or a manual list) and fires the sends
  • Dependencies: calendar access + SMS provider

Picking the demo

Choose by the prospect's loudest pain — missed calls / no reviews / slow follow-up / no-shows. Rough vertical defaults (primary + secondary): dental-medical → 4+1 · trades/contractors → 3+2 · restaurants → 1+2 · legal → 3+1 · real estate → 3+4 · auto repair → 2+4. Their actual complaint overrides the table.

Delivery process

  1. Customize with their real data — business name, hours, actual review link, industry language. This is what separates a demo from a pitch deck. Never fabricate reviews, stats, or capabilities the system doesn't have.
  2. Build into ventures/<automation-venture>/deliverables/demos/<prospect-slug>/<demo-type>/ (mkdir -p). Reusable vertical templates accumulate in deliverables/demos/_templates/ — the second demo in any vertical should take 30 minutes, not 3 hours.
  3. Deploy (if the demo needs hosting): get the founder's explicit green-light first, then deploy with your pinned site ID (netlify deploy --prod --site "<YOUR-SITE-ID>" or your host's equivalent). If a serverless function means the demo only works deployed, tell the founder BEFORE they try it locally and hit the error.
  4. Present live (let them call the number / submit the form) or screen-recorded with narration.
  5. Close: "That's this running 24/7 for [Business]. Setup takes about 1–2 weeks." Pricing is value-based and set by the founder per deal — never quote a number. → run proposal-generator for the formal proposal.
  6. Write-back: after any prospect demo or conversation, log an activity note on the prospect's workspace record and update the matching <TASK SYSTEM> task/stage before closing out.

Compliance gate

If a demo asset is client- or public-facing and touches a regulated topic (financing offers, health claims, legal outcomes, insurance), run regulated-copy-compliance against it before anyone outside the team sees it.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.