Our group’s work
Aranv is part of the Atharva System group. The projects below were delivered by teams across the group for companies in the United States, the United Kingdom and New Zealand. They show the kind of work we take on: AI agents, business platforms, Odoo, SaaS and eCommerce.
Client names stay private. Each project is described by what it does, the problem it solved and the results that were reported.
Planning something similar? Book a meeting and tell us about your goals.
Selected projects
Six examples of AI, platform and commerce work, with the results reported for each.
Project 01 · Enterprise AI · United Kingdom
Enterprise AI assistant
One conversation for CRM, documents and proposals
A UK company wanted its teams to work across their CRM, document library and Microsoft Teams from a single chat. Work was spread across separate systems, so routine jobs such as preparing a proposal or finding the right document took many manual steps.
The build combined AI agents, search over company documents and secure sign-in to each connected system, so people can ask for what they need and the assistant carries out the steps.
Project 01 at a glance
-
The challenge
Data lived in HubSpot, SharePoint and other tools. Long AI sessions became slow and costly, Word documents lost their formatting when AI edited them, and answers became less accurate when several files were uploaded at once.
-
What was built
Agents on Azure AI Foundry that route requests across systems, HubSpot deal stages that create proposals in SharePoint from branded Word templates, document search with Azure AI Search, sign-in to other tools without leaving the chat, and Teams notifications for approvals and updates.
-
Reported results
Proposal workflows about three times faster, roughly 70% less administrative work, and steady response times in conversations that run for hours.
Technology: React, TypeScript, FastAPI, Azure AI Foundry, Azure AI Search, Model Context Protocol, HubSpot, Microsoft Graph, SharePoint, Teams
Project 02 · HR technology · United States
AI platform for HR support
Confidential employee support with multiple AI agents
A US HR technology company wanted employees to raise workplace concerns in plain conversation instead of rigid forms, while HR teams receive structured cases with clear priorities and deadlines.
Sensitive cases needed extra care. When a complaint involves HR staff, it has to reach someone without a conflict of interest, and every case needs a visible path to resolution.
Project 02 at a glance
-
The challenge
Static forms discouraged honest reporting, complaints arrived incomplete, nothing escalated cases by severity, and HR teams received too many low-priority alerts.
-
What was built
A conversational intake that captures the facts of each concern, a LangGraph and LangChain workflow that routes cases between specialist agents, privacy rules that redirect conflicted cases, and severity-based escalation with deadline tracking.
-
Reported results
About 80% faster complaint intake and documentation, about 70% faster escalation and resolution tracking, and less repetitive work for the HR team.
Technology: React, TypeScript, Python, FastAPI, PostgreSQL, LangGraph, LangChain, NVIDIA AI Endpoints
Project 03 · Medical supply · United States
B2B commerce for a medical distributor
Healthcare ordering and approvals on Odoo 17
A US distributor of medical and rehabilitation equipment serves hospitals, clinics and care providers from a catalog of more than 100,000 products.
The project turned its Odoo store into a B2B buying platform that matches how healthcare organizations purchase: several buyers, approvals, purchase order references and different terms for each type of customer.
Project 03 at a glance
-
The challenge
A basic storefront and a standard cart could not support shared shopping lists, approvals or purchase order tracking, and retail, wholesale and VIP customers needed different onboarding and payment rules.
-
What was built
A redesigned Odoo storefront with quick view, comparison and product labels, shared shopping lists with multi-user approvals and PO references, customer groups for onboarding and payments, ShipStation shipping with live carrier rates, and product bundles.
-
Reported results
About 40% faster bulk ordering, 45% fewer shipping errors and a 30% higher average order value.
Technology: Odoo 17 Enterprise, Python, PostgreSQL, ShipStation
Project 04 · Privacy compliance SaaS · United Kingdom
Privacy compliance SaaS
GDPR compliance software as a service
A London privacy software company wanted to give organizations one place to manage the personal data they hold and their obligations under GDPR.
Over two years, the product grew from research and blueprints into a full cloud platform with an interface that people without a legal or technical background can use.
Project 04 at a glance
-
The challenge
Organizations had to record consent, answer data subject requests, run impact assessments, control international data transfers and report breaches on time, often with spreadsheets.
-
What was built
Product research, documentation and data schemas, a clear interface, data mapping and inventory, compliance automation with gap analysis, breach management and risk assessment.
-
Reported results
The product has since won several industry awards and recognitions for privacy and regulatory technology.
Technology: Ruby on Rails, React, AWS Elastic Beanstalk
Project 05 · Community health · United Kingdom
Social prescribing platform
Live patient data for GPs and care teams
A UK health platform connects people with local services that support their health and well-being, and helps GPs and care organizations refer patients to the right support.
Patient data came from three clinical and service APIs, each with its own structure, so the work focused on bringing that data together and making it useful in one view.
Project 05 at a glance
-
The challenge
Demographic and health data sat in separate systems such as EMIS and SystmOne, which made it hard for GPs to see how a patient was doing over time.
-
What was built
A central database that keeps data from three APIs in step, a real-time dashboard with charts of conditions, medication and other measures, well-being tracking, and an interface that older users find easy to use.
-
Reported results
The platform reports serving millions of patients, with many service providers now part of its referral network.
Technology: Ruby on Rails, Sharetribe, AWS, Sidekiq, MySQL, WebSockets, EMIS and SystmOne APIs
Project 06 · Print-on-demand eCommerce · New Zealand
Custom packaging store
A headless store for custom, sustainable packaging
A New Zealand company sells custom-printed, eco-friendly packaging to brands and retailers around the world, so its store has to handle personalized products at scale.
The store moved from Magento 1 to BigCommerce, with a separate Next.js front end that gives full control over design and speed while BigCommerce manages orders.
Project 06 at a glance
-
The challenge
Magento 1 had reached the end of its life, and standard BigCommerce themes could not deliver the product personalization and design the brand needed.
-
What was built
A full data migration from Magento 1 to BigCommerce, a headless Next.js storefront connected through APIs, product personalization, content from Contentful and Ghost, and Zendesk support.
-
Reported results
The brand reports more than 100,000 customers using its personalized print and packaging.
Technology: BigCommerce, Next.js, React, Tailwind CSS, Contentful, Ghost, MongoDB, Zendesk
Planning a similar project?
Tell us what you want to build and the tools you use today. We will suggest a practical first step.
Case Studies FAQs
-
Why are client names not shown?
We keep client names private. Each project is described by what it does, the problem it solved and the results that were reported.
-
Who delivered these projects?
Teams in our group, which Aranv is part of. Projects can draw on the wider group of developers, designers and testers, while Aranv works with clients in the United States.
-
Where do the results come from?
Each figure is the outcome reported for that project. Results depend on the starting point, scope and data, so we discuss realistic goals for your project during discovery.
-
Can you build something similar for us?
Yes. Book a meeting or send your project details. We will review your goals and current tools and suggest a practical starting scope.
Get in Touch
Cary, NC 27513
Canton, MI 48187