Faizan Khan
03 — Case study

Softs Leads AI

An AI system that answers a business’s calls and WhatsApp messages, captures leads and books appointments. I designed how it works from the back end to the front end.

Product
Softs Leads AI
Role
System architecture designer
Team
Three partners, including me
Status
Ongoing, I still work on it
Channels
Phone calls, WhatsApp, web
01 — Overview

Softs Leads AI is an automation product for small businesses. It answers phone calls and WhatsApp messages, replies to common questions, captures the customer’s details and books appointments, so no enquiry is missed.

Three partners, including me, built it. I was the system architecture designer, and I’m still working on the product.

Softs Leads AI homepage design with the headline Turn More Enquiries Into Customers and an AI voice agent card
Homepage design, with a live AI agent a visitor can talk to
02 — The problem

Enquiries are lost when nobody answers

Small businesses get calls and messages at all hours. When the team is busy or offline, the customer waits, or goes to someone else. The business needs every enquiry answered quickly, without hiring someone to sit by the phone.

03 — My role

System architecture designer

  • Designed how the system works from the back end to the front end, and how the pieces connect.
  • Designed the steps to run in n8n workflows, and built security into the design: encryption, secure credentials and access controls.
  • Worked with two partners to build it, and I keep working on the product.
04 — How it works

From a message to a business action

A customer messages the business on WhatsApp, the website chatbot or another channel. n8n receives the message through a webhook or API, identifies the channel and context, routes it to the right workflow and calls the AI and business tools.

The AI services understand the request. A language model detects the intent and writes the reply, using a knowledge base (FAQs, product information, policies) and business rules that control what the AI can do. The business systems then perform the action: capture the lead and update the contact in the CRM, check availability and create a booking, or reach other tools such as tickets and payments.

The customer gets the answer on the same channel, as a chat reply or a confirmation, or is handed to a person when it is needed.

Security and privacy run across the whole flow: encryption in transit (HTTPS/TLS) and at rest where supported, secure credentials, access controls and data minimisation.

AI automation architecture: customer channels feed n8n workflows, which call AI services and business systems, and return a response to the customer, with security and privacy across the whole flow
System architecture, from the customer’s message to the business action
05 — Decision: n8n

One place where the steps live

n8n sits in the middle. It receives every message, identifies the channel, routes it to the right workflow and calls the AI and business tools. Keeping the steps in one orchestration layer keeps the system easy to follow and easy to change.

06 — How a business grows

Fewer missed enquiries, more customers

  • 24/7 support: routine questions are answered at any time.
  • Lead generation: enquiries are captured and prospects qualified.
  • Automation: repetitive manual tasks are reduced.
  • Control and security: the team can track, monitor and step in when needed.

This is what the system is built to do. I haven’t published measured results on this page.