Profile

Danish Khan

Full Stack Developer

IntentPlug - Chat Based Forms
All Projects
SaaS ProductShipped

IntentPlug - Chat Based Forms

SaaS Platform for Businesses to create conversational forms

ClientCatapult
Year2026
RoleFull-Stack Developer
Duration3 months
// Impact

3x

Form completion rate vs. traditional forms

50+

Businesses onboarded

First quarter

40%

Reduction in avg. session time

99.9%

Uptime

Serverless architecture

// Tech Stack
Frontend
Next.jsTypeScriptTailwind CSS
Backend
PrismaPostgreSQL
Infra
Vercel
AI
OpenAI API
// Overview

IntentPlug reimagines web forms as natural conversations. Instead of static, multi-field forms that users dread filling, IntentPlug delivers a chat-based interface that guides respondents through a conversational flow — improving completion rates by up to 3x.

// The Problem

Traditional web forms suffer from low completion rates, poor mobile experiences, and a lack of engagement. Businesses need a way to collect high-quality data without creating friction in the user journey.

// The Solution

We built a drag-and-drop form builder that outputs interactive chat widgets. Each form can be embedded anywhere and supports conditional branching, file uploads, and real-time validation. An analytics dashboard provides funnel insights and response heatmaps.

// My Contribution

I designed and built the entire product end-to-end as the sole developer. This included the JSON state machine schema that powers the form engine, the drag-and-drop builder in Next.js, the standalone ~8KB embeddable chat widget, and the Prisma + PostgreSQL analytics pipeline. I also handled Vercel deployment, CDN configuration, and onboarded the first 10 businesses directly.

// How I Built It
01

Discovery

Audited drop-off patterns on existing client forms, identified friction points, and benchmarked conversational UI patterns across competing tools.

decision

Decided to build our own form engine rather than wrap an existing tool — existing solutions couldn't support the conditional branching depth clients needed.

02

Architecture

Designed a schema-driven form engine — each form is a JSON state machine. This decoupled the builder from the renderer and made conditional logic trivial to implement.

decision

Chose JSON state machine over code-gen because non-technical users needed to edit and extend forms without touching code.

03

Build

Built the drag-and-drop builder in Next.js, the embeddable chat widget as a standalone ~8KB script, and the analytics pipeline on Prisma + PostgreSQL.

decision

Built the widget as a zero-dependency vanilla script rather than a React component — couldn't require clients to have React just to embed a form.

04

Ship

Deployed to Vercel with edge caching on the widget CDN. Onboarded first 10 businesses directly, iterated on feedback, then opened up self-serve.

decision

High-touch onboarding first — spending time with early customers directly shaped the branching editor UX before opening self-serve.

// Key Results
3x improvement in form completion rates vs. traditional forms
Adopted by 50+ businesses within the first quarter
Average session time reduced by 40% with conversational UX
99.9% uptime with serverless architecture
// Learnings & Reflection

The hardest part wasn't the conversational UI — it was making conditional logic feel simple for non-technical users. We went through four iterations of the branching editor before landing on a visual node graph. Lesson: complex power shouldn't require complex UI.

THANK YOU

" First solve the problem.
Then write the code."

~ John Johnson