How We Built a SaaS Platform in 4 Weeks
A founder came to us with working AI technology for analysing customer conversations. The tech was solid. But there was no brand, no pricing, no platform - just raw capability sitting on a server somewhere.
Four weeks later, PingMe was live. Here's how we got there.
The starting point
The founder had built something genuinely useful - an AI engine that could parse customer conversations and surface actionable insights. Think sentiment analysis, topic clustering, and response quality scoring, all in real time.
But "useful AI on a server" isn't a product. There was no way for customers to sign up, no interface to interact with, and no brand to build trust around.
Week 1: Discovery and positioning
Before writing a single line of code, we spent the first week understanding the market. Who actually needs this? What are they paying for today? What language do they use to describe their problems?
We landed on a positioning that resonated: PingMe helps support teams understand what their customers are actually saying. Not just ticket volume and response times - the substance of conversations.
From there, we built out the brand identity. Name, logo, colour palette, typography, tone of voice. Everything needed to feel premium but approachable. The kind of tool a VP of Customer Success would trust.
Week 2: Pricing and product design
Pricing a SaaS product is part maths, part psychology. We designed three tiers:
- Starter - for small teams testing the water
- Growth - the sweet spot for most customers
- Enterprise - custom pricing, custom everything
Each tier needed clear feature differentiation without making the cheaper tiers feel hollow. We mapped out the feature matrix, designed the pricing page, and built the upgrade flows.
Simultaneously, we were designing the core product interface. Dashboards, conversation feeds, analytics views. Every screen needed to surface the AI's insights without overwhelming the user.
Week 3: Full-stack build
With designs locked, we moved into development. The stack:
- Next.js for the frontend and API routes
- PostgreSQL for data persistence
- Stripe for subscription billing
- The founder's AI engine integrated via API
The subscription system was the most complex piece. Sign-up, plan selection, payment, invoice generation, upgrade/downgrade flows, cancellation - each with its own edge cases.
Week 4: Polish and launch
The final week was about the details that separate "works" from "feels right." Loading states, error handling, email notifications, onboarding flows. We also built the marketing site - the page that would convert visitors into trial users.
We deployed on a Friday. The founder sent their first cold emails on Monday.
What we learned
Every founder project teaches us something. PingMe reinforced a few things:
- Brand comes first. The founder had great tech, but without positioning and identity, nobody knew what they were looking at.
- Pricing is a product decision. How you package and price directly shapes what you build.
- Four weeks is enough if you're decisive. The biggest time sink isn't building - it's indecision.
If you're sitting on technology that could be a product, the gap between "working prototype" and "launched SaaS" is smaller than you think. It just takes someone who's done it before.