By Sriram Subramanya, Founder, Managing Director & CEO, Integra Software Services
Every enterprise I know today runs on more software than it did five years ago. CRMs, HRMS platforms, content management systems, collaboration tools, project trackers, compliance dashboards, and a growing shelf of custom-built internal applications. Each of them was bought or built to solve a real problem, and each of them usually does what it was chosen to do. Yet, most C-suite leaders would say that the team feels less productive today. The paradox is that every new system has solved an individual problem while collectively making work more complex.
The reason is rarely that any single system is failing. Rather, it is how poorly these systems have been integrated. Information resides across several databases, analysis is carried out using different tools, and employees are left to deal with varying assumptions and biases while making decisions.
The cost you do not see on any invoice
Fragmented information systems create a productivity cost that never appears on P&L statements. It shows up as the pauses inside every workflow. Take, for instance, an HR manager approving a 10-day study leave under company policy. The process starts with finding the relevant policy document on the intranet, updating records on the HRMS, and notifying the employee through email or an internal messenger. Each step happens on a different system, even though the workflow itself is straightforward.
A 2023 Gartner survey found that nearly half of digital workers report that they struggle to find the information they need to do their jobs well. In an Indian enterprise context, where teams routinely span multiple offices and often multiple countries, that number may well be higher.
Other costs, equally true but not as commonly cited, include employees abandoning internal search engines and defaulting to querying their co-workers or even producing documents that already exist.
Different teams often duplicate efforts in solving the same problem, simply because they cannot see what others have already done. Decisions are made based on incomplete information, since there is no single platform that provides all the information needed. Reviews and audits cost more time than necessary because data is spread out on different platforms.
Each cost by itself may be viewed as a small friction. However, taken together and over the period of a year, these costs accumulate significantly.
Why the fragmentation grew in the first place
The fragmentation we are living with comes from many good decisions made independently over time, not from poor ones.
Each system was bought because it solved a real problem at the time. However, when it came to buying tools, the idea was never to build the information architecture. As a result, the systems evolved in isolation from one another, and what started as an inconvenience has gradually become a challenge that affects the learning and decision-making processes of the enterprise.
What happens when you attempt to scale AI on top of a fragmented stack
This is where the older fragmentation problem is quickly becoming a much more expensive one. AI has changed the economics of fragmented information systems. Earlier, disconnected systems mainly impacted employee productivity. Today, they also limit what AI can do.
Every enterprise AI initiative depends on access to reliable context. An AI assistant, agent, or workflow cannot reason over information it cannot find. If knowledge is scattered across disconnected applications, stored in inconsistent formats, or hidden behind isolated systems, AI works with only a fraction of the enterprise’s intelligence.
The organisations seeing meaningful productivity gains from AI are the ones that identified information architecture blind spots and invested effort in improving data quality and making knowledge discoverable, more than those with the most advanced models. AI simply amplifies the quality of the environment it operates in.
It becomes far more difficult for organisations that neglect this foundation. AI models churn out convincing outputs, relying on partial data or failing to find the necessary data altogether. Recently published findings from Forrester are clear on this issue; enterprise AI adoption is hindered by fragmented knowledge systems, regardless of how much the models are tuned.
From my experience at Integra, where I have worked with publishers and enterprises using AI models in their processes, I have seen this pattern repeatedly. AI systems do not factor in the investment made in technology. They simply work with the information available to them.
What actually works, from what I have seen
I would offer three observations from three decades of building content technology and services for global enterprises.
First, fragmentation is a governance issue. Nothing short of costly consolidation efforts will solve the problem where there is no clarity about which system is the source of truth. The issue of governance may sound like a small part of the discussion, but it is actually the key element.
Secondly, the consolidation that works is layered. The most successful organizations that I have observed rarely undergo the ordeal of ripping and replacing all their systems. Instead, they create an experience layer above the legacy stack that brings coherence to their systems and unifies search, surfaces necessary data at the right time, and puts all the governance rules in one place.
The third observation is particularly relevant for CIOs. Investing in information hygiene is dull, un-glamorous, and compounding. These same three traits keep it from winning in the annual budget conversation, yet they are exactly what make it so high-yield as an investment.
The productivity challenge is about to get tougher
The productivity cost of fragmented information systems was tolerable when the alternative was slow decisions and duplicated work. It becomes far less tolerable when the alternative is AI systems that cannot see the information they need to be useful.
The point is that irrespective of whether the top firms driving the next wave of digital productivity in India have access to the latest tools or the most substantial budget to adopt AI technology, it is important that we invest in coherent information systems that allow employees and AI to access, trust, and use enterprise knowledge effectively.