Case Studies

    Software development case studies

    Real projects, written plainly. Each case study says what the client needed, what AMT built, the stack, and only results that were actually measured. Where a client has not agreed to be named, we describe them instead of naming them.

    Schedule Technical Discovery

    Your first conversation will be with a senior technical partner, not a salesperson.

    Find a case study by what you need

    Find a case study by what you need
    You needStart with
    A chatbot or AI agent in productionTelecom chatbot, banking chatbot, LLM and RAG chatbot, MCP server
    A new SaaS product or platformLast-mile delivery SaaS, fleet management platform, rental relationship platform, patient analytics platform
    To modernise a live system safelyHospitality PMS: AngularJS to Angular, Delphi to .NET, B2B e-commerce and ERP phased migration
    Logistics, fleet or transport softwareFleet intelligence, commodity logistics, airport and stadium food delivery
    Blockchain that solves a real problemSEWA Digital Health Passport on Algorand, supply chain traceability
    Enterprise compliance softwareGRC platform
    Faster, safer releasesPlaywright QA automation
    Software for an NGOLegal aid platform

    All case studies

    • Telecom conversational AI agent

      Industry
      Telecom, Middle East
      What AMT built
      English and Arabic chatbot on WhatsApp, Messenger, web and apps
      Headline result
      50%+ of chats resolved, ~100,000 conversations a month
      Stack
      Angular, Node.js, IBM Watson

    • Conversational AI product

      Industry
      Travel, food and beverage, financial services
      What AMT built
      OpenAI intent detection and RAG for a support chatbot
      Headline result
      More accurate, context-aware answers
      Stack
      NestJS, FastAPI, LangChain, Chroma DB

    • AI chatbot builder

      Industry
      SaaS
      What AMT built
      MCP server that generates chat flows from a prompt
      Headline result
      Faster flow building
      Stack
      Node.js, Python, OpenAI, Anthropic

    • Microfinance banking chatbot

      Industry
      Financial services, Cambodia
      What AMT built
      Transactional chatbot on core banking and CRM
      Headline result
      24/7 balances, transfers, loan applications
      Stack
      Angular, Node.js, Wit.ai

    • AI retrofit planning engine

      Industry
      Energy, UK
      What AMT built
      AI model that turns a postcode into a retrofit plan
      Headline result
      Trained on 20 million properties
      Stack
      React, NestJS, Python

    • Logistics platform

      Industry
      Logistics, Germany
      What AMT built
      Multi-tenant last-mile SaaS with Bluetooth smart locks
      Headline result
      Many operators on one platform, each under its own brand
      Stack
      Native iOS and Android

    • Fleet management platform

      Industry
      Logistics
      What AMT built
      Multi-tenant platform on 12+ microservices
      Headline result
      Unified fleet operations
      Stack
      Java 21, Spring Boot, Kafka, AWS

    • Fleet intelligence

      Industry
      Transportation, Europe
      What AMT built
      Route optimisation and IoT predictive maintenance
      Headline result
      Less idle time, optimised routes
      Stack
      Spring Boot, Kafka, PostgreSQL

    • Commodity logistics platform

      Industry
      Logistics, Malaysia
      What AMT built
      Four apps for booking, dispatch, tracking and e-POD
      Headline result
      Paper replaced, theft risk reduced
      Stack
      Node.js, MongoDB, Kotlin, Swift

    • Airport and stadium food delivery

      Industry
      Travel and venues, US
      What AMT built
      Customer, restaurant, agent and admin apps
      Headline result
      24-hour advance, timezone-aware ordering
      Stack
      Angular, Node.js, Swift

    • Hospitality PMS modernisation

      Industry
      Hospitality
      What AMT built
      Six-phase AngularJS to Angular 21, Delphi to .NET 10
      Headline result
      In progress, parity at each phase
      Stack
      Angular 21, .NET 10

    • B2B e-commerce and ERP migration

      Industry
      Wellness, global
      What AMT built
      B2B marketplace with ERP, migrated in phases
      Headline result
      28% faster fulfilment, 34% higher retention
      Stack
      React, Node.js, Elasticsearch, AWS

    • Enterprise GRC platform

      Industry
      Financial and professional services
      What AMT built
      Risk, audit, BCM, GDPR, SOC and AML/CFT platform
      Headline result
      One platform for governance and compliance
      Stack
      Angular, Node.js, SQL Server

    • Playwright QA automation

      Industry
      Enterprise software
      What AMT built
      Playwright framework linked to Jira, Zephyr and CI/CD
      Headline result
      Earlier defect detection, faster releases
      Stack
      Playwright, TypeScript

    • Legal aid platform

      Industry
      Humanitarian
      What AMT built
      Multilingual, multi-tenant counselling platform
      Headline result
      6+ years of SLA support
      Stack
      Angular PWA, Node.js, Yonyx

    • SEWA Digital Health Passport

      Industry
      Social inclusion, India
      What AMT built
      Blockchain identity on Algorand with DigiLocker
      Headline result
      200,000 to 300,000 members targeted after pilot
      Stack
      Algorand, HashiCorp

    • Supply chain traceability

      Industry
      Biofuels, pharma, food
      What AMT built
      Mass balance traceability on Algorand
      Headline result
      Shared tamper-proof record of origin
      Stack
      Next.js, Node.js, Algorand

    • Rental relationship platform

      Industry
      Housing, US
      What AMT built
      Landlord, tenant and housing organisation platform
      Headline result
      Disputes prevented through shared records
      Stack
      Vue.js, NestJS, Firestore

    • Patient analytics platform

      Industry
      Healthcare, India
      What AMT built
      Comprehensive Health Records with AI tools
      Headline result
      Six years of product engineering
      Stack
      Next.js, Node.js, Python

    How we write case studies

    • Only measured results. If something was not measured, we do not put a number on it.
    • Names with permission. Most clients are described, not named. SEWA and the Algorand Foundation are named because the initiative is public.
    • The stack, always. Buyers check it, so every case study lists it.
    • A buyer's checklist. Each case study ends with the questions to settle before building something similar.

    By industry

    FAQ

    Can I speak to a client reference?

    Ask during technical discovery. References depend on each client's permission, so we match you with the closest project we can.

    Why are most clients not named?

    Many enterprise and public sector clients do not allow their names in marketing. We describe the client instead and keep the technical detail specific.

    Do the results come from AMT or from clients?

    Results come from the delivered project, not from estimates. Where a programme is still running, the case study says so and lists expected outcomes instead of results.

    Does AMT support systems after launch?

    Yes. Several of these case studies are long-term engagements, including more than six years of SLA support for a humanitarian NGO and six years of product work on a patient analytics platform.

    Tell us which of these looks most like your problem. We will tell you what we would do differently for you.

    See all services →