Project Overview
mTouch Labs designed and built a multi-tenant SaaS platform from zero to production for a B2B startup targeting operations teams. The platform combines project management, AI-assisted workflow automation, and cross-team analytics — reaching 200 paying teams within 3 months of launch.
The Challenge
The founding team had a clear vision but no technical co-founder. They needed a production-grade SaaS product — complete with multi-tenancy, billing, and onboarding — in under 16 weeks before their seed funding runway forced a pivot.
- No existing codebase — greenfield build with a 16-week hard deadline
- Multi-tenancy required from day one with full data isolation per organisation
- Subscription billing with multiple plans, trials, and seat-based pricing
- AI workflow automation was the core differentiator and had to be genuinely useful
- Self-serve onboarding required with no sales team to hand-hold new customers
- Real-time collaboration including comments, notifications, and live status updates from launch
Our Strategic Approach
We adopted a product-engineering hybrid model — leading weekly product reviews with the founders and making scope decisions collaboratively. A feature-flagging system from week 1 let the founders demo specific flows to investors at any point without waiting for a full release.
The Solution We Delivered
A Next.js SaaS platform with a Node.js API backend, PostgreSQL multi-tenant schema using row-level security, Stripe billing integration, and an AI workflow engine powered by the Anthropic API. Real-time collaboration is handled via WebSocket rooms scoped to each tenant.
- Multi-tenant workspace with full data isolation enforced at DB level via RLS
- AI workflow assistant that suggests automation rules based on team activity patterns
- Drag-and-drop project board with custom status columns and automation triggers
- Stripe-powered subscription billing with trials, upgrades, and seat management
- Real-time notifications and live cursor collaboration
- Self-serve onboarding with interactive product tour
- Cross-workspace analytics for operations managers
- Role-based permissions for Owner, Admin, Member, and Guest
Technologies Used
- Next.js — Frontend SaaS app with SSR, middleware-based tenant routing, and dashboard views
- Node.js / Express — REST API with tenant-scoped request context and WebSocket server
- PostgreSQL — Multi-tenant schema with row-level security for full data isolation
- Anthropic API — AI workflow suggestion engine and natural language automation builder
- Stripe — Subscription billing, trial management, and seat-based pricing
- Redis — WebSocket presence, session store, and real-time notification fan-out
- AWS — ECS Fargate, RDS, ElastiCache, and SES for transactional email
- LaunchDarkly — Feature flags for safe progressive rollouts and investor demos
Development Process
- Architecture & Tenancy Design — Designed the RLS-based multi-tenant PostgreSQL schema and tenant routing middleware
- Auth & Billing Foundation — Built authentication with invite-based team onboarding and Stripe subscription integration
- Core Product Build — Built project boards, workflow engine, and notification system in parallel sprints
- AI Workflow Engine — Integrated Anthropic API for pattern-based automation suggestions and natural language rule builder
- Real-Time Collaboration — Built WebSocket rooms scoped per tenant for live updates, presence, and comments
- Self-Serve Onboarding — Built interactive product tour and empty-state guidance reducing time-to-value to under 8 minutes
- Beta & Launch — Ran 4-week closed beta with 20 teams, iterated on feedback, then public launch in week 16
Results & Impact
The platform launched on schedule in week 16. The founders closed their seed round 3 weeks after launch, citing product polish and early traction as key factors.
- 200 paying teams onboarded within 3 months of launch
- Average time-to-value under 8 minutes from signup to first workflow automated
- Seed round closed 3 weeks post-launch
- AI workflow suggestions adopted by 67% of active workspaces
🎯 Key Takeaway
Building a SaaS product is not just an engineering challenge — it is a product and prioritisation challenge. By embedding with the founding team and making scope decisions together, we shipped a product that felt complete on day one, not a half-finished MVP.
Project Gallery
