About QuoteFlow AI
Our client is a London print company that receives a steady flow of quote requests by email, for everything from posters and banners to stapled books. QuoteFlow AI is the AI Email Agent we built for them. It works inside the company's own mailbox, day and night, and takes each quote request from the first customer email to a finished PDF estimate.
Technologies Used
We brought together an AI model, the client’s email, and their existing estimating system to give QuoteFlow AI everything it needs. This includes OpenAI to read and understand customer emails, the Gmail API to send and receive messages, and a link to the client’s estimating system so every quote uses their own prices.
Services we offered
Here are 4 technical integrations we made to meet the client’s needs.
-
AI Email Agent
We built an agent with Python and FastAPI that checks the mailbox regularly, spots quote requests, and works out which product the customer needs.
-
Integration with Gmail
We connected the agent to the client’s mailbox through the Gmail API, so it reads enquiries and replies from the company’s own email address.
-
Integration with the Estimating System
The agent sends confirmed job details to the client’s existing estimating system through its API. It then emails the customer the PDF estimate the system returns.
-
Staff Dashboard
We built a React dashboard where the team can see every enquiry, its status and its estimate, with search, status, and date filters.
Got a project you'd like to discuss?
Technical Integration
chilliapple partnered with the client to build an AI email agent that answers quote requests at any hour, while the team stays in charge of special requests.
-
Email Understanding
OpenAI reads each email, recognises quote requests, and picks out details such as size, paper, and quantity.
-
Follow-Up Questions
The agent checks each enquiry against the details that the product needs. If anything is missing, it sends a short, friendly email asking only the questions still open. It never guesses a detail on the customer’s behalf.
-
Conversation History
PostgreSQL stores every conversation, so each reply links back to the original enquiry. Customers are never asked for the same information twice.
-
Automatic Estimates
Once all details are confirmed, the agent prices the job with the client’s own estimating system and emails the customer a PDF estimate.
-
Staff Handover
Requests that need a personal touch, such as a discount, are marked Needs Staff on the dashboard. The customer gets a polite automatic reply saying the team will review their request.
-
Live Dashboard
Status counts show at a glance which enquiries are awaiting the customer, need staff, or have been quoted. From any enquiry, staff can approve or reject the quote, run the agent again, or draft a reply in Gmail.
Our Process
Requirements
Every quote started with an email that staff had to read, follow up on, and price by hand. Customers often left out details like size, paper, or quantity, which meant several rounds of emails before a price could be given. Enquiries sent in the evening or at the weekend waited until the next working day.
In a busy inbox, some enquiries could also be missed, and there was no single place to see each enquiry’s history. The client wanted customers to get a prompt quote at any time, without the team losing control of special requests.
Solutions
We built an AI Email Agent that works inside the client’s mailbox. It reads each enquiry, asks for any missing details, keeps every reply linked to the original email, and prices the job using the client’s estimating system. Anything that needs a personal touch goes to the team on the QuoteFlow AI dashboard.
Results
Customers now get replies at any hour, including evenings and weekends. Every estimate comes from the same pricing system the staff use, and the team no longer chases missing details by email. The Conversation, estimate, and status sit on one dashboard. Staff can now spend their time on complex enquiries and customer relationships rather than routine emails.