WarmLane workspace dashboard — real-time outreach performance, multi-channel sequences and per-tenant analytics. warmlane.io →
Two platforms in production with real paying users — one my own, one a client now owns and runs — plus a third in active build. The kind of product teams charge $500K+ and 6 months for. Here's what each one does for its users.
Case Study 01 · My own product
Weeks of visa anxiety, replaced by 10-minute clarity — a live product with paying users. Built solo in 3 months.
See the live product in production.
The Challenge
Visa applications are stressful, fragmented, and opaque. A typical Pakistani applicant for a US or UK visit visa spends weeks researching, second-guessing documents, and worrying about refusal — and still doesn't know their actual approval likelihood until they're sitting at the embassy.
DST.Travel was built to compress that experience into a 10-minute AI-powered assessment with a confidence score, document checklist, and refusal-risk analysis — at a price an average traveller can actually afford (999 PKR / $9 USD).
The catch: building this required production-grade AI workflows, multi-region pricing, secure data handling, payment integration, an admin operations layer, and 400+ SEO pages — all of which typically takes a 3-engineer team 6 months. I did it solo with AI leverage in 3 months.
What I Built
AI Visa Assessment, Country Discovery, and Trip Planner — each with its own wizard flow, paywall mechanic, free preview, and full paid report. Each app has 5–6 step wizards with auto-save, validation, and an inline lead-capture mechanism.
Not a single ChatGPT prompt. Structured prompt engineering accounting for nationality + destination + travel purpose + financial profile combinations. Refusal-risk modelling tuned over months. 13 DynamoDB tables structured for clean data lineage and report generation.
Stripe Checkout with regional currency tiers (USD / AED / PKR / INR / BDT), automatic geo-detection of pricing, webhook handling for order completion, refunds, and dispute management. PCI-compliant by Stripe's standard integration.
Full admin portal for orders, refunds, customer support, and analytics. Separate affiliate portal with referral tracking, magic-link customer report viewer, and automated payout calculations. All built ground-up.
Programmatically generated country-specific landing pages, each with FAQ JSON-LD schemas, internal linking, and search-optimised meta. Indexed by Google, structured for long-tail organic capture.
Tech Stack
Build Timeline
Database schema design, AWS deployment pipeline, auth system, Visa Assessment app v1 with end-to-end paid checkout. Stripe live by Week 4.
GPT-4o integration, prompt engineering, 13-table data schema, Country Discovery + Trip Planner apps. Admin portal v1 for operations.
400+ programmatic SEO pages, affiliate portal, customer report viewer with magic-link sharing, analytics, multi-region pricing, production hardening, launch.
Case Study 02 · Built for a client
More customers, booked automatically — an outreach engine the client owns and runs, instead of renting five tools. Built & deployed solo in ~6 weeks.
WarmLane workspace dashboard — real-time outreach performance, multi-channel sequences and per-tenant analytics. warmlane.io →
The Challenge
Cold outreach fails for operational reasons, not creative ones: messages never reach the inbox, teams duct-tape together five separate subscriptions (a sender, a warm-up tool, a sequencer, an AI writer and a booking app), and none of them truly talk to each other.
The client needed a branded outreach engine they could own and resell — not rent. That meant the whole outbound motion in one place: import an audience, generate copy with AI, warm up and protect sending reputation, send across email, SMS and WhatsApp, auto-reply to responses, and book qualified meetings — all metered, billed and isolated per customer.
The scope is what a funded team usually ships over many months: a multi-tenant SaaS with billing, AI, deliverability and an AWS backend. It was designed, architected, built and deployed end-to-end by one developer in roughly six weeks.
What I Built
One dispatcher orchestrating email, SMS and WhatsApp with a step-by-step campaign wizard, A/B message variants, dynamic merge fields and adaptive per-account pacing. Real-time lifecycle tracking from queued → sent → delivered → opened → replied — with idempotent sending (conditional writes prevent double-sends under Lambda retries).
Gradual account warm-up with automatic throttling, continuous bounce/complaint monitoring and automatic account quarantine on breach. SPF/DKIM/DMARC verification, inbox-placement testing against seed inboxes, and a global suppression list honouring unsubscribes, bounces and STOP requests.
Goal-based copywriting (lead-gen, demo, sales, support) tuned to channel and tone, grounded in a per-customer "AI Brain" knowledge base — company, product, pricing, objections, FAQs. Inbound replies are handled with intent detection and on-brand, human-reviewable responses; interested prospects can be booked automatically.
An account → workspace model (one account, many isolated workspaces, shared billing) with role-based teams (Viewer / Member / Admin), invites and an immutable audit log. Stripe checkout, customer portal, tiered plans and per-channel usage metering / overages — plus a super-admin control plane to provision, impersonate, suspend and bill every client tenant.
A public REST API (v1) with per-workspace API keys and inbound webhooks; an event backbone on SQS / EventBridge; analytics streamed via Kinesis Firehose to S3 / Athena. Calendar booking via Google / Outlook OAuth, structured logging, correlation IDs, CloudWatch alarms, 2FA and security headers — all defined as AWS CDK infrastructure-as-code.
Tech Stack
Idempotent sending under Lambda retries, a per-tenant distributed rate limiter, account-health auto-quarantine, TCPA-safe send windows with per-recipient time zones, and a feature-flagged migration path so the platform can scale 10× without downtime. These are the problems that are easy to fake in a demo and hard to get right in production.
Build Timeline
Auth, accounts, data model, core API and the React app shell.
Campaign dispatcher, sender pools, pacing, tracking and suppression.
Warm-up, account-health quarantine, SPF/DKIM/DMARC and placement tests.
AI Brain, copy generation, reply automation, SMS / WhatsApp, calendar booking.
Stripe billing, plans, usage metering, workspaces, teams, super-admin.
AWS CDK infra, eventing, analytics pipeline, observability, go-live hardening.
Status: functionally complete and operating in the primary AWS region — in final testing, ~85% toward general availability. The remaining work is operational scale-out and hardening, not new product.
The core MVP landed inside the usual 30-day window. WarmLane ran longer because the client kept adding scope as the product took shape — multi-channel sending, deliverability, a super-admin control plane and usage-based billing all grew out of that. Each addition was agreed upfront and priced separately, and because the codebase was built to be extended, adding on stayed clean rather than becoming a rewrite. The goal is always the same: ship the core product to production fast, then grow it from a live, revenue-ready base.
A newly-signed client build, kicking off now: an AI tool that finds local businesses with a weak online presence, auto-generates a branded audit of what's costing them customers, and writes the outreach to win them as clients — packaged as a licensed digital product. Full case study to follow once it ships.
What this means for you
These aren't demos. They're real products with real paying users — delivered in weeks and fully owned by the people who commissioned them. That's the exact outcome I deliver for founders: a live MVP in 30 days under $10K, or a larger production SaaS scoped to your roadmap — built to make money, not to sit in a portfolio.
30-min discovery call. Tell me what you're building. Honest fit assessment.