Premium Site Creation - Workflow Map

0

👤

User Request via Messenger

Entry point - "I want a premium website"

Triggers

Routed to Madame Walicai

Steps 1-5

💃

Madame Walicai - Complete Interview

Handles entire conversation from business name to final questions

📝 Step 1: Business Name Collection

In Conversation

Asks: "What's your business name?"

Actions

Extract business_name from answer

Save to interview_progress.checklist

Trigger auto-scraping (background)

🔍 Step 2: Auto Data Fetching (Background)

Scrapes Automatically

Google Maps: reviews, rating, address, phone, hours

Website: services, about, contact

SERP: web mentions

Saves to

interview_progress.auto_fetched_data

testimonials[], rating, sources[]

🔐 Step 3: Social Media OAuth (In Conversation)

Madame Walicai Sends Links

"Connect your Facebook" (OAuth link)

"Connect your Instagram" (OAuth link)

User Clicks → Fetches Last 365 Days

Facebook: posts (365 days), photos (max 30), about, events

Instagram: posts (365 days), photos (max 30), bio, stories highlights

Merges Into

auto_fetched_data (social_posts[], social_photos[])

⚙️ Step 4: Data Processing & Gap Analysis

LLM Processing Steps

1. Merge: Google Maps + Website + Social (365 days)

2. Deduplicate and categorize all content

3. Extract key information from all sources

4. Generate business description from collected data

Key Information Extracted

Business description & value proposition

Services/products list with descriptions

Team members (from website, social, Google)

Key achievements/milestones

Common customer questions/FAQs

Media Collected for Later Use

Photos categorized (products, team, location, events)

Videos collected (social media, website)

Stored in media library for Phalicai

🔍 Gap Analysis - Determine What's Missing

Analyze collected data vs required fields

Identify missing critical information

Examples: business goals, target audience, unique selling points, pricing

Create list of questions to ask in Step 5

💬 Step 5: Conversational Interview Completion

Generate Contextual Questions Based on Collected Data

Use scraped data to create personalized, conversational questions

Reference specific information from Steps 2-3 in questions

Example: "I saw that your customers love your [service X] on Google Maps. Is that your main specialty?"

Example: "On your Instagram page, I see a lot of [theme Y]. Do you want to highlight that on the site?"

Example: "Your reviews often mention [point Z]. Is it important to emphasize?"

Conversational Gap Filling

For missing critical info, ask naturally based on context

If no pricing found: "I didn't find your pricing online. Do you prefer to display prices or 'quote on request'?"

If no hours found: "Are you open every day or do you have specific hours?"

Make it feel like a conversation, not a form

Validate with User

Present summary using their own language/tone from social posts

"So if I understand correctly: [personalized summary based on their own words]"

Ask for confirmation naturally

Determine Which Fairies Are Needed

Analyze: business_type + all gathered info → fairy requirements

Uses business_type_feature_mapper.py logic

Determines: needs_booking, needs_ecommerce, needs_blog (photos always required)

ASK USER for confirmation - don't force booking/ecommerce if they don't want

Example: "Based on your business type, I recommend adding [booking/e-commerce]. Interested?"

Push Forward

"Great! I'll create your site with all this..."

Set interview_progress.completed = true

Transfer to Phalicai (always - photo collection is mandatory)

Step 6

📸

Phalicai - Photo Collection

File: app/services/fairies/phalicai.py

🎯 Trigger Condition

ALWAYS triggered - photo collection is mandatory

Every business needs photos for the website

🔍 Photo Gap Analysis

Reviews already collected photos from Step 4 (social media scraping)

Determines what photo categories are missing

Required categories: products/services, team/staff, location/storefront, events/activities

Analyzes business_type to determine priority categories

💬 Actions in Conversation

Requests Missing Photos via Social Media

Asks for missing categories identified in gap analysis

Send request via WhatsApp/Messenger - user can take photos now, later, or use camera

User flexibility: snap photos immediately, upload from gallery, or take later

💾 Saves to Database

Table: prospect_photos (only URL saved, not the file itself)

Media uploaded to R2 before website creation

Database stores: photo_url, category, uploaded_at

🔄 When Complete

Update prospects.has_photos = true

Transfer to Gralicai (Step 7) - design consultation is ALWAYS mandatory

Step 7

🎨

Gralicai - Graphic Design Fairy

File: app/services/fairies/gralicai.py

🎯 Trigger Condition

ALWAYS triggered - design consultation is mandatory

Every business needs logo, fonts, and design guidance

📥 Inputs (from Previous Steps)

Business name, industry, target audience

Brand colors from website/social media (Step 4 scraped data)

Existing logo (if found in scraped data)

Photos collected by Phalicai

💬 Actions During Interview

1. Logo Analysis & Improvement

Check if existing logo found in scraped data (photos/website)

If logo found: Describe the logo to the user

Ask: "I found your logo. Do you need any improvements?"

If improvement requested: Ask what specific changes they want

Use HuggingFace tool to improve logo based on user request

Tool: Input existing logo + specific improvement prompt

TODO: Find proper LLM/model on HuggingFace for logo improvement

If no logo found: Generate new logo via HuggingFace

2. Google Font Selection

Think about what Google Font fits their business type

Select font pairing (heading + body font)

Show link to Google Fonts page with the selected font

Example: "I suggest Montserrat for headings. Check it out: https://fonts.google.com/specimen/Montserrat"

Ask: "Is this font ok or do you want something different?"

If different: ask for their preference and suggest alternatives

3. Color Palette Selection

Analyze existing brand colors from scraped data (website/photos)

Suggest color palette based on industry and visual identity

Present color codes (hex) with visual representation

Ask: "These are the colors I suggest. Do you like them or want something different?"

If different: ask for their color preferences or inspiration

💾 Saves to Database

prospects.has_design = true

design_config: logo_preference, generate_logo (bool), font_pairing, color_palette, design_style

If logo upload requested: logo_url (uploaded to R2)

🔄 When Complete

Design preferences confirmed

Transfer to Valicai (Step 8A) - video selection

8A

🎥

Valicai - Video Selection Fairy

File: app/services/fairies/valicai.py

🎯 Trigger Condition

ALWAYS triggered - video review is mandatory

Videos collected by Madame Walicai in Step 4

📥 Inputs (from Step 4)

Videos collected from social media (Facebook, Instagram)

Videos from website scraping

Video metadata: views, engagement, date posted

⚙️ Actions

1. Review Collected Videos

Check each video from Step 4 collection

Determine if video describes the business OR relates to recent action/event

2. Suggest Publication

If video describes business → suggest publishing on website

If video relates to recent action → suggest publishing on website

Otherwise → skip video (don't publish)

💾 Saves to Database

Table: prospect_videos (video_url, category, selected_for_website)

prospects.has_videos = true

🔄 Transfer Conditions

Videos selected for website generation

Transfer to Boolicai (8B) if booking needed, otherwise Calicai (8C)

8B

📅

Boolicai - Calendar & Booking Fairy

File: app/services/fairies/boolicai.py

🎯 Trigger Condition

Determined in Step 5 - needs_booking = true

User confirmed they want booking functionality

Business type: appointments, reservations, consultations

📥 Inputs (from Interview)

business_name

opening_hours (if available)

services[] (to create booking types)

phone, email (for notifications)

⚙️ Actions

1. Create Merchant in Calendar System

INSERT INTO merchants (business_name, contact_phone, contact_email)

Returns: merchant_id

2. Configure Services

Add each service as a bookable item

For each service: name, duration, price (if applicable)

Ask if payment required to book (deposit or full payment)

If payment required: payment_required = true, payment_amount, payment_type (deposit/full)

If no payment: payment_required = false (free booking)

3. Configure Availability

Parse opening_hours → availability_slots

Default: Mon-Fri 9am-6pm if not provided

Set buffer time between appointments (if applicable)

Set max bookings per day/week (if applicable)

4. Payment Setup Note (if payment_required = true)

Inform merchant: Stripe Connect onboarding can be done later

Booking system will be created, payment setup can be completed after website launch

Generate Stripe Connect onboarding link (stored for later use)

Store link in merchants.stripe_onboarding_link

Merchant can complete Stripe setup anytime after website is live

5. Generate Booking Links

WhatsApp: https://wa.me/{phone}?text=Book%20{service}

Messenger: https://m.me/{page_id}

Direct: https://alicai.app/book/{merchant_id}

💾 Saves to Database

prospects.calendar_merchant_id = {merchant_id}

merchants table: business_name, contact info, stripe_onboarding_link, stripe_account_id (null until completed)

services table: service_id, merchant_id, name, duration, price, payment_required, payment_type

availability_slots table: merchant_id, day_of_week, start_time, end_time

booking_contacts table: WhatsApp/Messenger links

🔄 Transfer Conditions

Booking system configured

Transfer to Calicai (8C) if ecommerce needed, otherwise Blalicai (8D)

8C

🛍️

Calicai - E-Commerce Setup Fairy

File: app/services/fairies/calicai.py

🎯 Trigger Condition

AI Analysis: ecommerce = true

OR business_type in: retail, shop, store, e-commerce

📥 Inputs (from Interview)

business_name

products[] (only physical/digital products, not services)

currency (default: EUR)

shipping_zones (if physical products)

⚙️ Actions

1. Product Import/Creation

Check for existing integrations (Shopify, WooCommerce)

Import products if integration exists

OR create products from interview data

2. Payment Setup Note

Inform merchant: Payment integration can be done after website launch

E-commerce system will be created, payment setup completed later

Store note in prospects.payment_pending = true

3. Inventory & Shipping

Set stock levels (if provided)

Configure shipping zones & rates

Set tax rates (based on location)

💾 Saves to Database

Table: products (product_id, name, price, stock, merchant_id)

prospects.has_ecommerce = true

prospects.payment_config (Stripe/Mollie tokens)

shipping_config table: zones & rates

🔄 Transfer Conditions

Products available for website shop page generation

Payment gateway configured for checkout

Transfer to Blalicai (8D) if blog needed, otherwise Final Confirmation (Step 9)

8D

✍️

Blalicai - Blog & Content Fairy

File: app/services/fairies/blalicai.py

🎯 Trigger Condition

AI Analysis: blog_content = true

OR business needs SEO/thought leadership

📥 Inputs (from Interview)

business_name, business_type

services[], target_audience

Auto-fetched data (reviews, social posts)

Social media connections (for posting)

💬 Actions During Interview

1. Analyze Current Posting Frequency

Check how many times per week they currently post on Instagram

Check how many times per week they currently post on Facebook

Review posting patterns from Step 4 collected data

2. Confirm Posting Frequency

Ask for confirmation: "I see you post X times/week on Instagram. Want to keep that?"

Ask for confirmation: "I see you post Y times/week on Facebook. Want to keep that?"

Ask for website blog: "How many times per week for blog posts on your site?"

3. Suggest Content Topics

Analyze what topics they usually post about (from Step 4 data)

Suggest topics based on their existing content patterns

Ask for confirmation: "You usually post about [topics]. Should I continue with these?"

Offer additional topic suggestions based on business type

4. Inform About Flexibility

Tell them: "You can always contact me later to change posting frequency or topics"

Blalicai is available for adjustments anytime after launch

🔄 What Happens Later (After Website Launch)

Generate blog posts based on confirmed frequency and topics

Auto-publish to website blog section

Generate social media variants and post to Instagram/Facebook

Content creation happens automatically based on interview confirmation

💾 Saves to Database

prospects.has_blog = true

content_config: posting_frequency (Instagram, Facebook, blog), topics[]

Note: Blog posts and social content generated AFTER website launch, not during interview

🔄 Transfer Conditions

Posting frequency and topics confirmed

Content preferences saved

Transfer to Final Confirmation (Step 9)

Step 9

👩‍💼

Madame Walicai - Final Confirmation

Before calling website generator

✅ Verify Everything is Collected

Confirm all data collected: business info, auto-scraped data, photos, videos

Confirm fairy configurations completed (Phalicai, Gralicai, Valicai, Boolicai, Calicai, Blalicai)

Review checklist of what was gathered

🔄 Offer Possibility to Go Back

Ask: "Everything looks good? Or do you want to change something?"

Offer to go back to any fairy if user wants to adjust

Options: "Go back to photos", "Change booking settings", "Adjust content topics", etc.

If user wants changes → transfer back to relevant fairy

If everything confirmed → proceed to website generation

Step 10

🎨

Madame Walicai Calls Website Generator

Uses MCP tool to call premium_website_generator.py

Generator Actions

1. Load all interview data + auto_fetched_data + photos

2. Load all fairy data (booking, ecommerce, blog)

3. Generate logo (HuggingFace or text)

4. Build rich context from ALL scraped data

5. Generate pages with AI (home, about, services, contact, shop, blog, booking)

6. Upload to R2 storage

Result

Live site: https://{slug}.alicai.app

prospects.generated_site_slug

✓

🌐

Live Premium Website

Professional multi-page site with all features

🧚 Premium Website Creation - Workflow

📋 Test Scenario

🤖 Testing Method

User Request via Messenger

Madame Walicai - Complete Interview

📝 Step 1: Business Name Collection

🔍 Step 2: Auto Data Fetching (Background)

🔐 Step 3: Social Media OAuth (In Conversation)

⚙️ Step 4: Data Processing & Gap Analysis

💬 Step 5: Conversational Interview Completion

Phalicai - Photo Collection

🎯 Trigger Condition

🔍 Photo Gap Analysis

💬 Actions in Conversation

💾 Saves to Database

🔄 When Complete

Gralicai - Graphic Design Fairy

🎯 Trigger Condition

📥 Inputs (from Previous Steps)

💬 Actions During Interview

💾 Saves to Database

🔄 When Complete

Valicai - Video Selection Fairy

🎯 Trigger Condition

📥 Inputs (from Step 4)

⚙️ Actions

💾 Saves to Database

🔄 Transfer Conditions

Boolicai - Calendar & Booking Fairy

🎯 Trigger Condition

📥 Inputs (from Interview)

⚙️ Actions

💾 Saves to Database

🔄 Transfer Conditions

Calicai - E-Commerce Setup Fairy

🎯 Trigger Condition

📥 Inputs (from Interview)

⚙️ Actions

💾 Saves to Database

🔄 Transfer Conditions

Blalicai - Blog & Content Fairy

🎯 Trigger Condition

📥 Inputs (from Interview)

💬 Actions During Interview

🔄 What Happens Later (After Website Launch)

💾 Saves to Database

🔄 Transfer Conditions

Madame Walicai - Final Confirmation

✅ Verify Everything is Collected

🔄 Offer Possibility to Go Back

Madame Walicai Calls Website Generator

Live Premium Website

Maps Category Mapping

📊 Legend

⚡ Malicai - Content Edits