AI Chatbot Developer (OpenAI + RAG) – Embeddable AI Assistant for Book Companion Tool

Job ID: 40293692

Budget: $3,000 – $5,000 USD

Fluent in English please
This project is for the development of a focused AI assistant, not a generic chatbot.
The goal is to create a knowledge-based financial guidance assistant that answers questions using curated materials from a book and supporting financial education resources.
Please read the full project description carefully before applying. Proposals that clearly demonstrate experience with RAG (Retrieval Augmented Generation) systems and AI chat assistants will receive priority review.

Project Overview
I have just published my first book on Amazon and will serve as one of the primary knowledge sources used by the AI assistant. I have not applied any marketing for the book at this point as I am waiting for the completion of this project.

The project is intentionally focused and lightweight — the goal is a clean Version 1 launch that can expand later a 2.0 version and finally at 3.0 version based on a number of market factors.
As a companion tool for the book, I want to build a simple AI assistant called Money Coach Max.
Max will allow users to ask questions about household money management and receive educational responses based on curated financial knowledge. Max will have a dedicated page on the book website that is under construction now.
The website itself will be built on Wix, so I am looking for a developer to build the Max chatbot system and provide an embeddable chat interface that can be placed on a Wix page.

Budget and Timeline
Budget target: $3,000 – $5,000 depending on approach and experience
Timeline: 2–3 weeks preferred
Goal: clean Version 1 launch
Developers may propose simplified architectures or improvements that reduce cost while maintaining quality.

Core Features Required
1. AI Chat Assistant
Develop an AI-powered assistant called Money Coach Max.
The assistant should:
• Answer questions about family financial management
• Use a friendly coaching tone
• Provide educational financial guidance only
• Avoid giving investment, tax, or legal advice
The assistant should connect to an AI API such as OpenAI.

2. Knowledge Base (RAG System)
Max will answer questions using a curated knowledge library.
We will provide:
• text from the book
• curated household finance educational documents
The developer should implement a retrieval-based knowledge system so that Max references these materials when generating answers.
Preferred approach includes:
• document chunking
• embeddings
• vector database
Alternative implementations are welcome.

3. Embeddable Chat Interface
The developer should provide a chat interface that can be embedded into a Wix page.
The developer will build the functional chat component only.
The overall page design and layout will be created separately in Wix.
The developer should provide a chat module that includes:
• user message bubble
• Max response bubble
• text input field for user questions
• send button or enter-to-send functionality
This component should be delivered in a form that can be embedded or inserted into a Wix page.
The visual styling of the page and surrounding layout will be handled separately.

4. First Name Greeting
Before the chat begins, the user will be asked:
“What’s your first name?”
Max will greet the user by name.
Example:
Hi Sarah, I’m Max. Ask me up to three questions about managing money and I’ll do my best to help.
No account creation required.

5. Three Question Trial Mode
Visitors to the website will be able to ask up to three questions as a trial experience. This trial allows potential readers to experience Max before purchasing the book.
After the third response, Max will display a message such as:
I hope this helped. Readers of the book receive unlimited access to Max.
The question limit should be enforced using:
• browser cookies
• basic IP rate limiting

6. Access to 24/7 Max
Book buyers will receive 24/7 unlimited access to Max.
The Version 1 system needs to include a book access code verification. The system should allow book owners to unlock unlimited access to Max.
The system should include a simple verification tool that allows book buyers to unlock unlimited access to Money Coach Max.
Users will submit proof of purchase through a form that collects:
• first name
• email address
• Amazon order number
Optional: upload screenshot of purchase receipt.
Once submitted, the system should display a randomly selected access code from a predefined list of approximately 10 codes provided by the client.
Users can then enter this code on the Max page to unlock unlimited access.
The system should store the verification submission and allow the user to continue using Max without repeating the verification during the same session.
The goal is a simple and lightweight verification process without building a full account system.
Generate 10–20 different codes like:
STORM2026
MAX2026
JOEPLAN
SAFESEAS
SMARTMONEY
That way if one code leaks online, you can disable it later.

System Overview (Money Coach Max)
The following diagram illustrates the expected architecture and user flow for Version 1 of Money Coach Max. The goal is a clean Version 1 MVP with minimal complexity that can be expanded later.
Max User Flow
User visits Max page


Enter First Name


Max Chat Interface
(User question bubble / Max response bubble / input field)


AI Response Engine (OpenAI API)


Knowledge Base (RAG System)
├─ Book Manuscript
└─ Financial Knowledge Library


Max Response Returned to User

Trial Mode
Visitors may ask up to three questions to experience Max.
User asks question


AI generates response


Question counter increases


After question #3 → trial limit reached
Max then displays a message explaining that unlimited access is available to book buyers.

Book Owner Unlock Flow
After the trial limit is reached, users will see two options:
Buy the Book on Amazon
Unlock Max (Book Owners)
If the user selects Unlock Max, they will be shown a simple verification form.
The form should collect:
• First Name
• Email Address
• Amazon Order Number
• Optional screenshot upload of receipt
After submission, the system should issue one of several predefined access codes (provided by the client).
The user can then enter this code to unlock unlimited access to Max.

Question Logging
Each question asked to Max should be stored anonymously for insight reporting.
Data stored:
• Question text
• Topic category (AI-generated)
• Timestamp
This data will be used to generate insights such as:
“Top financial concerns households are asking about right now.”

Scope Clarification
The developer is responsible for building:
• AI chat engine
• RAG knowledge system
• embeddable chat component
• trial question limit
• verification form and access code unlock system
• anonymous question logging
The client will handle:
• website design (Wix)
• page layout and styling
• marketing copy
• Max character branding
NOTE: It is our assumption that with successful execution of this project, we fully expect to build versions 2.0 & 3.0 working the same developer.