Skip to content
Shreyash Agare
Tracklist

A2 · 2026 · Internal tool

Briqhaus CRM

A real-estate CRM for Indian builders: the system sales teams live in all day, with leads, inventory, bookings and the state of every deal in one place.

Scope & Ownership

Design, front end and QA across seven modules, plus the design system that holds them together.

Measurable Impact

Multiple builders run their sales on it. Seven modules shipped in eight months, all looking like one product.

Briqhaus CRM cover

Listen to this one

0:22

Eight months inside one CRM

At a Glance · Executive Summary30-sec read
The Problem

Briqhaus had a legacy CRM in daily use without an underlying design system, causing every new feature to drift in visual consistency and data structure.

My Role & Ownership

Product Design Engineer: Owned end-to-end design, component architecture, front-end QA, and implementation across 7 core modules.

Key Design Decision

Preserved the inherited production layout to avoid disrupting active sales teams, while engineering a unified Figma DS 2.0 component library underneath.

Shipped Impact

Shipped 7 cohesive modules in 8 months (AI agent, knowledge base, projection admin, call analysis, etc.) now running sales for multiple real-estate developers.

Briqhaus is the operating system a builder's sales force lives in: leads, inventory, bookings, payments, and deals. Several developers now run their commercial sales pipelines on it.

7 ModulesShipped in 8 months
154Tables designed
740Screens in system
100 minSaved per agent daily

When I joined, a rough CRM was already in daily use. The mandate was not a cosmetic redesign. It was engineering a scalable design system under the inherited look so that seven ambitious new modules could ship as one cohesive platform.

What I built: Seven modules in eight months

Call analysis in action: transcribes calls, extracts buyer commitments, and saves 100 minutes of manual call review per agent every single day.

Production modules

Briqhaus AI: Natural language agent that queries records, reassigns leads, and runs analytics without requiring CRM navigation expertise.
Click to zoom ⊕
Briqhaus AI: Natural language agent that queries records, reassigns leads, and runs analytics without requiring CRM navigation expertise.
Briqpedia: RERA and legal knowledge base answering out of a verified regulatory corpus rather than generative hallucination.
Click to zoom ⊕
Briqpedia: RERA and legal knowledge base answering out of a verified regulatory corpus rather than generative hallucination.
Projection module: Executive visibility across towers, agreement values, demand generated, and expected collections per construction milestone.
Click to zoom ⊕
Projection module: Executive visibility across towers, agreement values, demand generated, and expected collections per construction milestone.
Banker Snapshot: One-click export synthesizing CRM financials and construction state into investor-ready diligence PDFs.
Click to zoom ⊕
Banker Snapshot: One-click export synthesizing CRM financials and construction state into investor-ready diligence PDFs.
Channel partner campaigns: Tracks broker attribution, commission windows, and revenue conversion by campaign channel.
Click to zoom ⊕
Channel partner campaigns: Tracks broker attribution, commission windows, and revenue conversion by campaign channel.

Feature demo films

I design the feature, write the copy, and edit the walkthrough films:

AI Studio: Generates project brochures directly from CRM asset metadata.
Custom reporting: Instant cross-table analysis and CSV export.
Regulatory QA: Real-time RERA and GST lookup directly inside the deal record.

Core architectural decisions

Projects and towers first, pipeline stages second

Instead of A generic CRM kanban board dragged by sales stage

Generic CRMs organize by deal stage. Indian real estate organizes by building. A salesperson thinks "Tower B, 3BHKs left, who is ready to book", not "show me negotiation stage". Software structure follows the physical asset.

After Unit availability and deal state stopped being two separate lookups.

Information density on purpose: a second click is heard on the phone

Instead of Hiding data behind nested tabs to keep screens looking minimalist

When a sales executive is on a live phone call, a second click creates an awkward pause where the buyer hears them hunt. We optimized for maximum scannability: strict tabular alignment, calibrated color contrast, and immediate access to unit pricing.

After Sales reps find critical pricing and availability while actively speaking with buyers on calls.

Executive overview: Real-time collections, pipeline volume, unit inventory, and overdue payments accessible in one glance.
Click to zoom ⊕
Executive overview: Real-time collections, pipeline volume, unit inventory, and overdue payments accessible in one glance.
Production lead table: Designed against real dense datasets, with multi-stage filtering and quick-action triggers.
Click to zoom ⊕
Production lead table: Designed against real dense datasets, with multi-stage filtering and quick-action triggers.
Payment ledger: Demanded, received, and pending balances tied directly to verified bank reconciliations.
Click to zoom ⊕
Payment ledger: Demanded, received, and pending balances tied directly to verified bank reconciliations.

Built the component design system rather than one-off screens

Instead of Ad-hoc screen designs that required reinventing tables and states per module

Rather than drawing isolated screens, I engineered Briqhaus DS 2.0: standardized data tables, filter bars, status badges, and empty states. Building front-end components myself guaranteed that the spacing and ergonomics in Figma matched the production code.

After Seven complex modules shipped in eight months with complete visual and behavioral consistency.

Impact & Retrospective

Briqhaus CRM now powers operations across multiple commercial builder organizations:

  • 100 minutes saved per agent daily through automated call intelligence and one-click lead updates.
  • Unified 7 modules across 740 screens under one single design system.
  • Next challenge: Streamlining the legacy underlying data models to make upcoming modules even faster to ship.