New Delhi, Aug. 5 -- Artificial Intelligence (AI) adoption is shifting from generic automation to industry-specific, AI-driven workflows. While sectors like banking and retail are already advanced, others with low digital maturity are still catching up. The focus now is on embedding intelligence into everyday processes, often within legacy systems.
AIonOS, an AI-first digital transformation company, that was founded as a joint venture between InterGlobe Enterprises (led by Rahul Bhatia, founder of IndiGo Airlines) and Assago Ventures (spearheaded by industry veteran CP Gurnani), builds platforms tailored to industry needs while working with existing infrastructure. In a conversation with TechCircle, Arjun Nagulapally, CTO at AIonOS, explains how the company approaches AI deployment, where enterprises are seeing real gains, and why some sectors remain behind. Edited Excerpts:
Each industry has its own operational DNA. How does your company tailor AI implementations or agents to fit the specific workflows of two very different sectors, like logistics and telecom?
Traditionally, software was a boxed product, people bought machines with Windows pre-installed and used apps like Word or PowerPoint for general tasks. It was a one-size-fits-all model. Later, this moved to cloud-based SaaS applications from companies like Salesforce or ServiceNow. These platforms offered core capabilities with some industry-specific extensions, but they were still largely generalized solutions.
From around 2020 to 2023, AI began to shift that model. The focus is now moving toward truly industry-specific solutions because different industries have different priorities. Patterns start to emerge when you map out who the key stakeholders are, what metrics you want to move, and what capabilities you need to support those outcomes. Instead of building another horizontal software platform, we've focused on solving specific problems by building solutions aligned to those needs.
AI isn't just another layer of software, it changes how people interact with systems. Users no longer have to adapt to the software. Now, software needs to understand people. The role of developers is shifting from writing complex code to translating human language into machine-executable outcomes.
This changes the focus from technology to outcomes. It's not about writing code; it's about moving business metrics. So the two main pillars are, defining the right business metrics for key personas and, enabling employees and customers to interact with systems naturally through AI.
Every department follows a predictable set of standard operating procedures, steps, systems, and outcomes. Our product suite abstracts and supports those activities with intelligence layered on top. For example, IntelliConverse, part of our customer experience stack, understands user intent in different languages, maps that intent to business outcomes, and connects to back-end systems without friction.
This becomes a horizontal platform that we adapt for each industry. For instance, a user might call a hotel to book, cancel, or modify a reservation. These are clear intent-driven workflows. In healthcare, we work with a large U.S. hospital where patients call to check appointments, ask about medications, or confirm dietary instructions. These calls are also structured around predictable tasks.
Internally, similar patterns exist across HR, IT, sales, marketing, and operations. Employees might apply for leave, request assets, or run operational dashboards. These activities can also be standardized, layered with AI, and adapted to specific industry needs.
We bring deep domain expertise and build vertical-specific solutions. When customers deploy our IntelliMate Agentics suite, they can see results quickly, not in months or years, but often in weeks. From first conversation to a working prototype, customers see their business metrics start to shift.
What kinds of efficiencies are enterprises really aiming for with AI today? Where are the biggest gains actually coming from?
If a technology helps people work more efficiently, like eliminating the need for paper invoices, signatures, shipping, and manual scanning, it frees them to focus on more strategic tasks. In areas like route planning and demand forecasting, where global supply chains are under pressure from tariffs and sourcing issues, this shift is essential.
It's not just about setting static business rules or picking optimal routes. It's about having an AI-driven approach that continually evaluates data in real time. The system should understand whether the data aligns with business goals and automatically adjust to data changes to support better decisions.
In logistics, having timely access to the right information, accurate demand forecasting, efficient route planning, and smart warehouse operations is critical. Much of this now involves automation and vision-based systems that ensure compliance and streamline operations. These efficiencies aren't just theoretical, they're being realized across logistics, travel, and hospitality.
The broader goal is to shift human effort from repetitive work to strategic decision-making. Even a 10% improvement in decision-making speed can have a big impact when scaled across millions of transactions. Companies like MoveIn and UPS ship globally, so any operational improvement matters. For airlines like Indigo, faster responses to customer queries-such as providing updates via WhatsApp during disruptions-reduce the need for human intervention, lower costs, and improve customer experience.
These gains, while they may seem small in isolation, compound significantly when applied consistently and at scale.
Let's take telecom as an example, where networks are dense and real-time decisions are crucial. What role are AI agents playing in areas like network optimisation and customer service?
There are always "yesterday's problems" issues the industry is already dealing with. Many companies respond by offering quick fixes, often packaged as AI solutions. We do that too when needed, but that's not our focus.
Telecom is going through a major transformation. No provider today is just offering phone services. They've expanded into technology services, cloud, infrastructure, TV, wireless internet, and more. The industry has moved from landlines to mobile, to bundled enterprise services, and now to AI-driven customer engagement. Telecoms are becoming tech companies, and we help them take the next step in that shift.
Our role is to build systems of truth for these providers. Take a small business wanting an Airtel landline. Right now, they visit the website, enter a number, and wait for a callback. That callback leads to a conversation about needs, hiring plans, and bandwidth. It's reactive and slow.
Now imagine a setup where the system already knows the business is a customer, sees their recent activity, like a LinkedIn post about opening a new office or a marketing campaign, and uses that context. An AI agent can connect the dots, interpret the signals, and send a lead to a field agent. That agent then reaches out with a relevant, timely offer. It's proactive, not reactive.
We're working with telecoms, technology vendors, and the Open Digital Architecture (ODA) to make this possible. Recently, we showcased the first full lead-to-order AI agent at TM Forum's Digital Transformation World. It cut down the process from days to minutes, turning a cold lead into a hot one fast.
This is the next phase of evolution, where intelligence is embedded, business logic is learned (not coded), and operations are automated at scale. That's where our domain expertise comes in. We co-create solutions with our customers through our AI Foundry. AI Foundry brings together people who understand both the technology and the industry. Instead of waiting for customer requirements, we experiment with what's possible, build proof-of-concepts, and show value early. We help business stakeholders see results on real data so they can make decisions faster and with confidence.
How do you manage integration with legacy systems in sectors like telecom or healthcare, where the infrastructure is outdated but still critical?
Technology has always evolved with a strong emphasis on backward compatibility. This has been a constant across every major shift. When we build AI-native systems, we also ensure they can integrate with older technologies.
We don't just build new products, we work with outdated systems too. We make them AI-ready and compatible with the agent-based applications we build. This requires investing in resources to upgrade legacy tech so it can interoperate smoothly with new layers.
When we engage with customers, we're upfront: they don't need to immediately replace their software. We handle the transition without disruption. For example, think about an airline. You can't ground a flight just to upgrade a legacy system. Like nightly baggage system maintenance, upgrades must happen without downtime.
AI helps with this. It can sit on top of legacy systems, keeping the business running while we upgrade the backend. There are proven strategies to handle this, and we've already applied them, for example, with one of the world's largest aircraft manufacturers. We automated their entire flight onboarding process as part of a broader modernization effort.
We use this experience as a repeatable model. We call it ALS-DLC, and it's the foundation of how we modernize legacy systems at scale.
GCCs are playing a bigger role in digital transformation. How do you see them shaping AI adoption across the sectors we've discussed?
Global Capability Centres (GCCs) have evolved. Earlier, the approach was to scale operations first, often through backend IT outsourcing-and then optimize later. That model is changing.
GCCs were traditionally focused on reducing costs and improving efficiency through outsourced support functions. But now, there's a shift toward co-innovation with customers. For example, with AI foundry models, we're seeing critical business functions being outsourced, not just support roles, where value is delivered upfront through AI and domain expertise, not by scaling headcount.
This shift makes deep understanding of the business domain essential. You can't drive innovation with people who lack domain knowledge or the ability to apply technology meaningfully. That's where GCCs are heading: moving from support centers to innovation hubs.
The ongoing layoffs across services and product companies aren't about human inefficiency. They reflect changing decisions on where to apply technology versus human talent. Meanwhile, GCC hiring is increasing, especially for roles that require domain expertise and the ability to lead industry transformation. There's still a shortage of such talent.
The future of GCCs lies in becoming global innovation hubs, delivering faster, cost-effective innovation. Technology and economics will continue to shape this direction. The ability to innovate quickly and affordably is what gives companies a competitive edge. GCCs can be the central point enabling that innovation at scale.
Looking ahead, which sector do you think is least prepared for the shift to AI, and what's holding it back, tech debt, regulation, mindset, or something else?
Industries with low levels of digitisation are often the ones most ready for disruption. Take agriculture in India, for example. Technology adoption in this sector is still limited. Farm sizes are small, and many farmers use outdated, non-scientific methods. In regions where the first wave of technology adoption hasn't occurred, catching up becomes more difficult.
Several factors contribute to this. One is the broader ecosystem-regulation, government support, available technology, and whether industry players see value in modernising. Another is the expected return on investment. Agriculture has significant potential to benefit from technology. Drones can be used for pest surveillance, autonomous machines can handle tilling, and sensor-based systems can guide decisions on fertiliser use and crop placement. We're already building these kinds of models for clients in the Middle East and Asia-Pacific.
AgriTech is starting to grow, but its progress depends on whether governments and key stakeholders decide to invest in and support these innovations. In contrast, sectors like knowledge work have already seen deep technology penetration. Workers in these fields understand its value because they've used it. The same applies to industries like banking, consumer goods, and retail. These are already undergoing disruption, whether in how they reach customers, manage brand visibility, or deliver near real-time experiences.
Ultimately, the pace of change depends on where an industry sits on the digital adoption curve. Those already well along will continue advancing quickly, while others will need to catch up.
Published by HT Digital Content Services with permission from TechCircle.