The PRD said fifteen modules. The product had fifty destinations.

The first problem wasn't design. It was that nobody knew how big the product had become.

The legacy PRD described about 15 modules. Product discovery found roughly 24 sidebar groups, 50+ destinations and around 60 entities.

Buyers, channel partners and vendors exist as entities in the system. That doesn’t make them users. The core audience is the builder-side organisation — sales, CRM and operations, accounts, project management, compliance, admin.

Confusing an entity for a user is how enterprise products end up with navigation nobody needs.

1. 01

   Map what actually existed before designing anything. Designing against an outdated definition of the product would have been faster, and wrong.!

2. 02

   RERA Portal never becomes a write workspace. Let people edit a mirror of a government record and you’ve created two versions of the truth, with no way to tell which is real.!

3. 03

   Don’t decide yet what ships on mobile. Understand the whole product first; the keep/simplify/remove calls come after. Deciding early would have been guessing with confidence.!?

Compliance became its own product area — RERA Compliance Hub, deliberately not nested under Projects. It splits in two. RERA Easy is the operational side: overview, compliances, applications, billing, analytics, tickets — users track applications and supply documents, while the filing itself is handled by the RERA Easy team. RERA Portal is a read-only window onto government data: project overview, profile, complaints, penalties, making MahaRERA information easier to verify and use.

The compliance screen runs five states, not three: inactive, submitted, in progress, error, previous. Error is the interesting one — a failed submission isn’t "not done", it’s worse, because somebody thinks it’s handled.

The Builder Mobile App is a native product, not a responsive layout. The desktop sidebar already needs "Show More" and "Collapse"; turning it into mobile tabs would reproduce the desktop problem on a smaller screen. So: IA first, navigation second. Same product logic, entities, business rules and terminology. Different information architecture, navigation, interaction model and screen composition.

Underneath it, the design system moved to 2.0 and toward 3.0 — atomic structure, tokens, variables, typography, components, patterns, documentation, governance. What it replaced: duplicate and legacy components, external dependencies, inconsistent typography, multiple colour systems, inconsistent spacing, uncontrolled variations, missing component sets, local styles never tokenised, screen-specific duplicates.

Two rules: build the system as infrastructure, not decoration. Reuse before create.

Live and selling. Six AI capabilities at different maturities — WhatsApp AI and Meeting AI shipped; AI Call Analysis, the Briqhaus AI Assistant and AI Marketing Studio still coming; BriqPedia in beta.

Designing a product where half the intelligence doesn’t exist yet is its own problem: the interface has to leave room for capability that isn’t there, without promising it.