Centre is 2008. Each ring, an epoch.
Each point, something we built.
Now.
Still building.
For eighteen years, every time software changed where it lived, how it was built or what it could do, we moved with it. This is that record, from the first App Store apps to governed AI agents in production.
The real story is three histories happening at once.
Divum began within months of the App Store opening, at the moment software left the desktop and became an economy of continuously evolving platforms. Since then the technology landscape, the product economy and this company have moved together. Read any row left to right.
| Era | Technology shifted | Product economy shifted | Divum’s position |
|---|---|---|---|
| 2008–10 | iPhone, App Store, Android. Six smartphone operating systems competing at once. | Utilities, media, sport and games become apps. | Multi-platform mobile engineering |
| 2010–13 | Native ecosystems mature. Cloud becomes practical. | Ticketing, travel, payments and commerce move to mobile. | Transactional mobile products on AWS and Google Cloud |
| 2013–16 | React, Angular, APIs, responsive web, cloud platforms. | Marketplaces and digital-first startups scale rapidly. | Product engineering and scalable backends |
| 2015–18 | PWA, TensorFlow, deep learning, cheap mobile data. | Commerce, logistics, social and super-app models expand. | PWA, machine learning, data, Jio-scale consumer software |
| 2017–21 | Blockchain, enterprise cloud, APIs, microservices. | Networks of businesses become software-connected. | Multi-enterprise systems and ⦃param⦄ |
| 2022–24 | Foundation models and generative AI. | Software becomes conversational and generative. | Applied GenAI engineering |
| 2025–26 | Agents and model-driven interfaces. | Software performs work rather than presenting screens. | Agentic platforms: Divum.ai and ⦃ai/Studio⦄ |
The date ranges overlap on purpose. Technology waves do not replace each other; they accumulate. Today’s agent still calls a React front end, over APIs, on cloud infrastructure, against enterprise databases whose design goes back fifteen years.
Seven epochs. One company inside all of them.
A conventional timeline shows years. This one shows what changed in the world, what people started doing, what businesses needed, and what Divum built next. Travel through the epochs
What eighteen years taught us.
Each epoch left a lesson that still shapes how we work. These three decide most of our engineering choices today.
Distribution moves faster than organisations.
Every platform shift, from Windows Phone to PWA to agents, rewarded the teams that could ship on the new surface before the organisation had finished debating it. We signed up for early access programmes with Apple, Google, Microsoft, Nokia and Samsung for that reason, and still do.
The visible product is the edge of a larger system.
From the first payment inside a feature-phone app, the work that mattered was behind the screen: APIs, order management, data pipelines, identity. The same is true of AI. The model is a utility; the integration is the work.
Governance is an architecture decision, not a document.
HIPAA, PCI DSS, FDA testing and ISO 27001 taught us that compliance has to be designed in on day one. Our agentic platform treats determinism the same way: as a contract the enterprise can read, version and audit.
Today: applied AI engineering, with the platforms to run it.
Divum is a solutions company. We design and build AI-native products and enterprise systems, and we have productised what we learned into two accelerators that make those solutions faster to ship and safer to operate.
Full-stack GenAI solutions inside existing enterprise systems: voice agents, document intelligence, knowledge and analytics, code migration, and the mobile, web and cloud engineering that has always been our core.
Procurement, logistics, inventory, quality and control towers shared across organisations, with a patent-filed Universal Commerce State Machine. It began in 2018 as one of India’s earliest enterprise blockchain implementations.
Every AI workflow is a versioned MindSchema: traceable, reproducible, cost-bounded and model-agnostic, deployable on an Edge Cluster inside the customer’s own environment with zero data egress.
Proof, from production.
The AI work is not demo-stage. These are outcomes from systems running inside large Indian enterprises, on the same governed runtime.
Counts are Divum’s own records. Each figure on this site carries an evidence level in the archive.
Built alongside the platforms.
Early access programmes and partner status with the OEMs and cloud platforms are how we stayed ahead of each shift, from the first Lumia camera SDK to Azure AI Foundry co-sell.
Regulatory experience across HIPAA, GDPR and DPDP, PCI DSS and FDA system testing. Information security management system certified to ISO/IEC 27001; SOC 2 Type II in progress.
2008 was not the destination. Neither is 2026.