Case studies

Systems built and rebuilt for critical operations

Production rebuilds, critical system extensions and selected Studio products. Each entry shows the problem, what Grail built, the stack used and the impact we can back up.

Systems and technologies used across this work

TypeScript

Railway

Postgres

From stalled development to production readiness in under 30 days

Grail took over delivery of an inherited care-services marketplace codebase and rebuilt the production platform around a modern architecture, automated delivery controls, granular permissions, durable workflows, and observable operations.

Problem

An inherited platform remained incomplete and unstable after an extended development cycle. Requested changes were slow and frequently created regressions, leaving the business without a dependable path to production.

Solution

We rebuilt the application architecture, delivery pipeline, permissions, workflows, testing, infrastructure, observability, and operating controls as one connected production system.

Stack used

React

TypeScript

Postgres/PostGIS

TemporalWorkOS

AWS

Impact

The inherited repository contained about 1.5M lines of code. More than 400 work items went through rolling client review, and the release repository included 700+ automated test files. The client-agreed production-readiness milestone was reached in under 30 days from Grail taking over delivery.

Project 02 · Crypto banking and transaction infrastructure

Transaction reconciliation and assertion layer

Grail designed and implemented a control layer around an existing .NET crypto-neobank platform, making expected transaction behaviour explicit, inspectable, and testable before launch.

Problem

Incoming and outgoing transactions carried complex expected behaviour. The team needed those expectations made explicit before launch so mismatches could be investigated and system decisions understood.

Solution

We mapped the existing .NET system, wrote down the rules governing transaction flows, and implemented reconciliation paths and assertions around incoming and outgoing behaviour.

Stack used

.NETTransaction rulesReconciliationAssertionsOperational controls

Delivered

Pre-launch capability delivered: the system was mapped, transaction rules were written down as explicit checks, reconciliation paths were covered, and an assertion layer was added for investigation and operational control.

Studio

Purpose-built products and internal applications

Studio is a separate offer for focused product builds. These projects show how we scope, test and ship the first useful release while the client keeps the code and stack.

Studio

Lobang: guarded property matching inside WhatsApp

A WhatsApp-first property workflow that parses listings and buyer requirements, verifies agent details, runs configured relevance and safety checks, and scores potential matches.

Read case study

Problem

Manual property matching made relevance and trust difficult to manage. Automation also had to protect contact details until both parties wanted an introduction.

Solution

We built a service that verifies CEA details, applies relevance and safety checks, scores matches, and asks both parties to confirm before revealing contact links.

Stack used

Temporal

Postgres

WhatsApp Cloud APIEvolution/BaileysFlue

Slack admin

Impact

The prototype demonstrates a marketplace workflow with deterministic checks, admin traces, and double opt-in before contact details are revealed.

Studio

A consumer astrology MVP built for early testing

A Studio concept taken from product prompt to a reviewable web experience, including its interface, content flow, and interactions.

Read case study

Problem

The concept needed to become concrete enough to test before committing to a larger product build.

Solution

We used our internal app-generation workflow to produce the frontend, then reviewed the copy, pacing, interactions, and visual flow.

Stack used

Finite MachineStudio buildFrontend appGenerated UI

Impact

The result is a reviewable consumer-product concept. No production-adoption or user-outcome claim is attached.

Studio

A retail prompt-to-app platform

A retail-focused application builder that turns product prompts into generated interfaces, preview links, and deployable artifacts.

Read case study

Problem

Teams needed working product surfaces before a normal development cycle could make an idea concrete enough to test.

Solution

We built a Finite Machine workflow that captures the specification, provisions a sandbox, generates and edits the application, and packages the result.

Impact

Client-reported at launch: 100 weekly active users. The repository implements the prompt-to-app loop from sandbox provisioning through preview and deployable artifacts.