Sellazo AI

AI-Powered Conversational CRM & Sales Automation Platform

Why I Built This

Sellazo started from a genuine engineering challenge: how can technology meaningfully reduce lead conversion time and make sales workflows more intelligent?

What began as a study project quickly evolved into something far more ambitious. I saw in it a rare opportunity to push my technical boundaries across multiple fronts simultaneously, real-time communication with WebSockets, reliable asynchronous processing with queues, AI orchestration at scale, complex third-party integrations, and a sharper product sense around usability and interface design. Building a platform people would actually pay for forced me to think beyond code quality and into system design, scalability, CI/CD pipelines, infrastructure decisions, and product architecture.

This is a project I continue to evolve actively, not just as a technical exercise, but as a real SaaS product with paying customers and growing operational complexity. Each new feature brings new engineering trade-offs worth solving.

The platform is live and fully accessible. You can explore and use Sellazo AI for free at sellazo.com.

The source code is hosted in a private GitHub repository. If you're a recruiter or engineer interested in reviewing the codebase, I'm happy to grant access, just share your GitHub email and I'll add you directly.

Sellazo AI is a multi-tenant SaaS platform designed to transform how businesses capture, qualify, and manage leads through AI-driven conversational experiences.

The platform unifies conversational AI, CRM, workflow automation, scheduling, lead acquisition, analytics, and event-driven integrations into a single scalable ecosystem focused on sales acceleration and operational efficiency.

Businesses can create intelligent AI assistants powered by custom prompts, deploy them across multiple channels, and automate sales workflows based on real-time conversational context, with full support for high scalability, low operational cost, and extensibility across multiple industries.

Core Features

AI Conversational Assistants

Businesses can create fully custom AI assistants configured through structured prompts and instructions. Assistants interact with leads in real time, qualify prospects automatically, answer contextual questions, trigger actions based on conversation understanding, schedule meetings, and collect structured customer data.

The conversational experience was designed to simulate natural messaging interfaces similar to WhatsApp-style interactions, reducing friction for end users.

Dynamic Lead Capture System

A flexible form engine allows businesses to build customizable lead capture forms distributed through standalone shareable links, embedded iframes, website widgets, QR codes, paid marketing campaigns, and social media bio links.

Every form submission automatically becomes a structured lead inside the CRM. After submission, businesses can route leads into traditional success flows or directly into AI-driven conversational onboarding, making lead acquisition transition seamlessly into real-time engagement.

AI Workflow Automation Engine

One of the platform's core differentiators is its AI-driven trigger system. Instead of static form conditions, workflows are configured based on natural conversation context:

• Lead mentions being over 30 years old → send a specific email
• Purchase intent is detected → move lead to a high-priority CRM stage
• Scheduling intent is identified → automatically create a calendar meeting
• Qualification criteria are met → notify sales teams or external systems

The engine supports email triggers, CRM stage transitions, scheduling workflows, webhooks, external integrations, and AI-based contextual decision making.

Conversational CRM

All interactions automatically become structured CRM entities. The CRM includes kanban pipeline management, lead stages, notes and attachments, full conversation history, AI-generated insights, activity tracking, and workflow automation, unifying conversational interactions and operational sales management into a single workflow.

AI Analytics & Natural Language Reporting

An AI-powered analytics module enables users to generate reports and charts through natural language queries:

• "Which campaign generated the most qualified leads this month?"
• "How many leads requested financing?"
• "Show conversion rate by pipeline stage."

The platform dynamically interprets requests and generates quantitative visual reports and insights without requiring manual configuration.

Scheduling & Calendar Integrations

The platform integrates with Google Calendar and Google Meet to enable automatic meeting scheduling, calendar synchronization, and AI-triggered scheduling workflows with automatic Meet link generation. A native scheduling system is also available for businesses that prefer centralized appointment management inside the platform.

Meta Conversion Integration

Direct integration with Meta Pixel and Conversion APIs allows businesses to connect Pixel IDs and access tokens to track and optimize campaign performance. Conversion events are sent to Meta when leads are created or transition to qualified CRM stages, enabling advanced attribution tracking for paid acquisition strategies.

Webhook & Event System

The platform exposes a webhook infrastructure for external integrations, allowing businesses to send leads to external CRMs, trigger external automations, and synchronize operational data across systems. The architecture follows an event-driven model to ensure reliability, scalability, and asynchronous processing.

Architecture & Technology Stack

Frontend

• Framework: Next.js, React, TypeScript
• Styling: Tailwind CSS, Shadcn UI
• Data fetching: React Query
• Key characteristics: Server-side rendering (SSR), dynamic embeddable widgets, reusable component architecture, scalable multi-tenant UI structure

Backend

• Runtime: Node.js, Fastify, TypeScript
• Databases: PostgreSQL, Redis
• Key characteristics: Multi-tenant isolation, queue-based processing, event-driven workflows, high-concurrency handling, AI orchestration pipelines

Redis is heavily utilized for caching, queue processing, event reliability, retry mechanisms, and real-time orchestration.

AI Infrastructure & Context Engine

The platform integrates with OpenAI and DeepSeek through a custom AI orchestration layer that manages prompt composition, structured memory, conversation summarization, context injection, function calling, cost optimization, and model routing.

Key results from the AI layer:

• 80% reduction in token usage through intelligent context compression
• Improved response consistency across long-running conversations
• Scalable conversational memory handling for high-volume sessions
• Reliable cost optimization enabling viable SaaS unit economics

Infrastructure & Scalability

• Frontend: Vercel
• Backend: Railway
• Next milestone: Migration to AWS for horizontal scalability, advanced infrastructure orchestration, improved reliability, and long-term enterprise growth

Engineering Challenges Solved

• Multi-tenant AI isolation at the prompt and data level
• Scalable conversational memory management within token constraints
• AI cost optimization for SaaS viability at scale
• Reliable asynchronous event processing with queue safety and retry mechanisms
• Real-time conversational workflows with low latency requirements
• Dynamic prompt orchestration adapting to business-specific configurations
• CRM and AI workflow synchronization across concurrent sessions

Outcome

Sellazo successfully evolved into a production-ready AI SaaS platform supporting real business operations, with paying customers using the platform in production. Results include automated lead qualification workflows, reduced operational response times, scalable AI-driven customer engagement, and flexible integrations across multiple industries.

Key Learnings

Building Sellazo required combining AI engineering, distributed systems, full-stack development, product design, workflow automation, and SaaS architecture into a cohesive product. The most impactful lessons:

• Efficient context management is the critical constraint for AI scalability
• Queue-based architectures significantly improve reliability in AI-driven systems
• AI cost optimization directly determines SaaS viability at scale
• Event-driven systems provide strong extensibility foundations for automation platforms
• Modular architectures accelerate product iteration and reduce delivery risk

Technologies Used

• Frontend: Next.js, React, TypeScript, Tailwind CSS, Shadcn UI, React Query
• Backend: Node.js, Fastify, TypeScript
• Databases: PostgreSQL, Redis
• AI / LLM: OpenAI API, DeepSeek API
• Infrastructure: Vercel, Railway, AWS