The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Back to feedPost
I'm building booking software for pet groomers in Egypt. It is not live yet.
What works today:
A customer opens the groomer's page on their phone, picks their area, their pet and a free time, and books. No call.
The groomer opens their phone and sees today's visits in order. One tap when a groom is done. They change their own prices and it shows on the booking page straight away.
It works for mobile vans, salons, or both.
The video is a 77-second screen recording of the real thing, not a mockup.
What is not there yet: online payment, WhatsApp messages, and many needed features.
It started as a booking app for one business. I rebuilt it so any groomer can set up their own. I want to know whether groomers would actually use it.
The network for creativity
Join 1.25M professional creatives like you
Connect with clients, get discovered, and run your business 100% commission-free
Creatives on Contra have earned over $150M and we are just getting started
Related posts
How do you show a client three website directions on one call without drowning in tabs?
This week we stopped juggling. We built a presenter page with Claude Code and ran the whole call from one tab.
The problem it solves: on a design call you open a tab per concept, resize the window to "show mobile", scroll hunting for the section you meant, and the client loses track of which version is which. The work was fine. The way we showed it wasn't.
What it does, all in one tab:
Direction tabs across the top. The real site runs inside a browser frame, so scrolling, hover states, animations and video all work live.
Desktop or Phone. Phone is a true 390 px view. Desktop renders the real 1440 px layout scaled to fit, so you see the actual design, not a squeezed one.
Full screen in one click, made for screen sharing.
Highlight buttons that jump the site straight to the section you are talking about: the opening, the work grid, the film.
A short page list (home, one project and contact), so the walkthrough has a beginning, middle and end.
The guardrail designers will appreciate: only 4 or 5 pages of each site open. The landing page, one full project and the key pages. Every other link shows "This page comes with the full build" instead of opening. You pitch the real thing without handing over the build before there is an agreement.
It travels too. Packed as a zip, it runs from a double-click with no server and no install, so our partner can run the same demo on his own calls.
Why not just send a staging link? It opens every page, and on a call the client wanders off on their own. This keeps everyone on the same screen.
What I would tell any designer pitching websites: show the site moving. Scroll, hover and motion are exactly what a PDF can't show.
How do you present website concepts today: a PDF, a Figma prototype or a live link?
Locking everything except 4 or 5 pages is the smartest part. It scopes the demo like an MVP, so the client reacts to the core flow instead of wandering into pages that don't matter yet.
The Problem:
Most Shopify stores install a quiz app for $50/month to recommend products. The apps are slow, don't match the brand design, and add to the 15-apps problem.
What I Built:
I built a custom Interactive Ritual Builder directly in Shopify for VERAÉ, my premium textured-hair demo brand - no quiz app.
How it Works - What you see in video:
Start: "Let's understand what your hair is asking for" - 3 simple questions, no overwhelm
Q1: "What would you like your routine to do better?" - Hydration, Definition, Repair, Scalp, Refresh
Q2: "How much structure feels right?" - Simple, Balanced, Complete
Q3: "How often do you usually wash?" - About once a week, About twice a week, It depends
Result: Personalized ritual - Example "Refresh Ritual" - lighter routine for moisture between wash days
Selling: Shows full routine [Wash $28 + Cloud Rinse $29] with total $86 and one-click "Add full routine to cart" - customer builds their own bundle
What I Did:
Built custom diagnostic engine in Liquid - logic analyzes answers and recommends ritual
Built dynamic step-by-step bundling - customers mix-and-match, not forced into one kit
Built conversion architecture - from quiz answer straight to cart with full routine, no friction
Styled to match VERAÉ premium brand - beige, black, editorial fonts - not generic app design
No app installed - faster load, no monthly fee, fully editable in Shopify editor
Tools Used:
Shopify, Custom Liquid, Shopify 2.0 Theme Development, Conversion Architecture, Agency Solutions
Context:
Demo project - VERAÉ is my own self-initiated store to show how I build custom experiences without relying on apps. Not client work. No performance claims.
My Goal:
Fewer unnecessary apps and less manual work. Custom code where it matters, lean app stack everywhere else.
Disconnected or broken systems? Let's fix it.
Most businesses don't need another WhatsApp chatbot.
They need a system that can actually get work done.
Imagine a customer messages your business at 11:47 PM:
"Hi, can I book a consultation for tomorrow?"
A basic chatbot replies:
"Thanks for contacting us. Our team will get back to you."
An AI-powered WhatsApp agent should be able to do much more.
Here's what I'd build instead:
1. Understand intent, not just keywords.
Identify whether the customer wants to book, reschedule, get a quote, check an order, or speak with a human.
2. Connect to live business systems.
Check real availability through a booking API. Retrieve approved information from the CRM. Never invent an available slot or a price.
3. Execute actions safely.
Create bookings, update lead records, send confirmations, and trigger follow-ups.
But only after validating the required information.
4. Remember the right context.
If a customer returns three days later, the agent should understand the relevant conversation history without asking the same questions again.
With proper permissions and data retention controls.
5. Know when NOT to automate.
Refund disputes, unusual requests, sensitive information, and low-confidence decisions should move to a human with the conversation context attached.
And here's the part most demos skip:
What happens when the customer sends the same message twice?
Or the booking API times out after creating the appointment?
Without idempotency checks, retries, and proper state management, an impressive AI demo can become a very expensive operational problem.
That's why I'd build a WhatsApp AI agent around three layers:
→ Conversation intelligence: Understand the request.
→ Workflow orchestration: Validate and execute the right action.
→ Operational reliability: Logging, permissions, retries, monitoring, and human handoffs.
The goal isn't to make a bot sound more human.
It's to make the business respond, decide, and execute more reliably.
I'm curious how business owners see this:
If your WhatsApp could securely complete ONE business task without your team stepping in, what would you automate first?
Bookings? Lead qualification? Follow-ups? Order updates? Something else?
