How IOH is rebuilding its IT organization to be AI-native

Telcos have long been held back by transformation projects that take years to deliver. So when IOH set about introducing major changes “our first principle was that we have to be fast and we have to be right first time, says Hitesh Kumar, SVP and Head of Consumer Tech & Digital Innovation, in an interview with TM Forum’s Insight.

“Whether you are merging, transforming or migrating your mission-critical systems, it can no longer be a three-to-five-year game for a telco,” he explains. “Things are becoming irrelevant much faster than we have experienced in the last decade.”

IOH faced that challenge at scale. Over 15 months to September 2026, its consumer IT teams completed the unification of 18 revenue-enabling core platforms following the merger of Indosat Ooredoo and Hutchison 3 Indonesia. The program covered customer care and call center systems, sales and distribution platforms, and the core revenue platforms – charging, billing, catalog and order management – for IOH’s roughly 100 million customers.

At the same time, the company was answering a question common to many operators: “How do we make it faster without changing the people, while ensuring the right cost to serve and the fastest time to deliver?,” explains Kumar. Its answer was to prepare those same teams to become AI-native during the program: “Same people, new skill sets.”

Changing behavior, not processes

“It’s all about people,” Kumar emphasizes. “Once your people are used to using AI, future problem solving becomes much easier for your company. Without that, your organization will evolve too slowly.”

Rather than impose new processes, IOH let employees discover the benefits of AI for themselves, without fear of failure or cost. “We are doing it organically, not in a provocative way. It is more behavioral than about processes,” he says.

Between January and July 2026, more than 50 IT engineers were encouraged to “tokenmaxx” – to use as many AI tokens as they needed – and to solve broader business problems with AI rather than within their functional silos. At the end of the period, instead of cutting budgets, IOH asked them to show the value of their token usage.

“Using AI to deliver outcomes is a mindset change, and we measured it very scientifically,” says Kumar. IOH compares the cost of internal development with the cost of using an external partner, and tracks how far people now work across the business rather than within a single function. By IOH’s measures, internal delivery now costs significantly less than external quotes for comparable work, and the cycle time for a medium-sized development has fallen from three weeks to three days.

From IT silos to a forward-deployed model

Two tools underpin the change. The first is a set of guardrails covering security, data privacy, spend limits and code review. The second is an internally developed AI skills library: employees describe a business problem in natural language, and the library supplies the skills needed, from architecture and scaling to storage and pipeline management.

“If you have never done coding in your life, you would still be able to do full-stack development using IOH’s skills for front end, Java back end, pipeline management, IaaS, event-driven architecture, testing and more,” claims Kumar. “Use the right skills and you will be able to get your feature out at scale for 100 million customers, validated by our autonomous, home-grown testing suite. We also focus on strong eval sets when switching inference from one model to another.”

As a result, traditional IT roles are blurring. “With AI, the right skills and the right guardrails, you are no longer a developer, or a scrum manager, or a delivery manager, or a tester. You are somebody who can work across the delivery chain,” he says. “We are changing our operating model from a traditional IT vertical to a forward-deployed operating model. Working horizontally is the new way for all of us.”

Building in-house and resetting partner relationships

The second major shift is in how IOH works with its partners. “With AI, building a product is no longer a differentiator,” says Kumar.

“With our partners, every change we make to our systems now has to demonstrate clear efficiency gains. We are still working with our partners, but our focus has shifted to building more of the experience layer ourselves, alongside our partners, because AI is now capable enough for us to build it at scale.”

IOH set itself the principle that 99% of software code has to be generated by AI. “Engineers will write very little code and will instead focus on complex problem solving,” says Kumar.

The results are already visible. IOH has developed Flash, a commercial product catalog with AI embedded at both design time and runtime, which allows new products to be launched in the market instantly – a task one person can now handle end-to-end, according to Kumar: AI has delivered “a better output with greater flexibility, fully aligned with our specific needs as an operator.”

In another case, a developer with six months’ experience built a new content aggregator platform that allows flexible bundling of non-telco services. Each new partner can now be integrated in just three days. “Within the last six months, the team has shipped four to five platforms that are creating an impact for the business – rolling out products much faster, improving the end-to-end purchase experience, and enabling hyper-personalization and localization at scale,” says Kumar.

Don’t wait for perfect foundations

Kumar’s advice to other telcos is not to get stuck trying to perfect their foundations first. “The rate at which AI is changing means few telcos can uplift their foundations fast enough,” he says. Instead of a long data-unification program, for example, a federated data layer combined with MCP (Model Context Protocol) “helps you kickstart your journey into being AI-native.”

What IOH does insist on is clean, AI-generated documentation for every problem it solves. In this way its knowledge base survives in the agentic space as people or LLMs change over time.

“Once we have that scale of more and more agents working for us, everything else will follow,” says Kumar. “Every telco will run on two workforces: people and agents. The winners will be the ones who learn to manage both.”