From AI Experiments in 2025 to Enterprise Scale in 2026 Why Data Foundations Will Decide the Winners Current state of AI initiatives

Current state of AI initiatives; organizations have spent billions of dollars on AI initiatives. As per published reports, only 5% of the initiatives are delivering value, and the payback period for most is 2–4 years, as opposed to the 7–12 months expected for technology investments.

The reasons cited for such a low success rate are: lack of clear objectives, lack of a governance model for AI initiatives, and, more importantly, poor data quality, fragmented workflows, and data governance issues.

It is said that modern organizations are connected but still disconnected. This paradox exists because these organizations have digital systems, networks, and integrations, but their data, processes, and people are not connected. Siloed data, inflexible processes, reactive compliance and governance, combined with poor data quality, have either stalled or slowed AI adoption across organizations.

These disconnected organizations lose nearly $13 million annually, experience failure rates of up to 65% in their digital transformation initiatives, and realize slower revenue growth.

Industry Need

The industry is looking for ways to unify data, processes, and compliance without rebuilding its entire IT infrastructure. It seeks software systems that can adapt to changing business goals without requiring re-engineering for every change. The industry is also ready to move from automation to orchestration.

Why are AI initiatives not giving the desired results? – Decoding DPP

DPP stands for Data-Process-People

Data drives decisions, and bad data results in poor or sub-optimal decisions. At the core of this is critical master data such as suppliers, products, bills of material, customers, employees, and assets. Often this critical data, also called foundational data, is found to be incomplete, inaccurate, inconsistent, and unverified. Adding non-compliant data to this list makes it even more challenging for organizations to rely on data for decision-making.

Apart from data challenges, workflows are broken and often include many manual steps. Information is not captured at the source, and multiple systems collect data that must then be consolidated to create a single source of truth for downstream applications.

All of the above results in reduced operational efficiency, delayed service delivery, manual steps to verify and correct data, higher compliance risk, all of which lead to higher costs and lost revenue.

Re-imagined master data management platform

Prevention is better than cure; all of us know this. Similarly, data quality should not be an afterthought. Irrespective of the source of the data, it must go through the same rigor of quality, compliance, and completeness checks. This means organizations must build systems or data guards to prevent bad data from entering, because any bad data that does enter the organization must be reviewed and cleaned - which costs time, money, and is often very people-intensive.

If systems and processes are built to prevent bad data from entering, the time, effort, and money spent by organizations on cleaning the data will be eliminated. This is the modern, reimagined data management approach.

What is a re-imagined MDM?

• Unified data, not siloed
• Verified, validated, compliant data
• Verification at source - i.e., verify PAN, GSTN, bank account, geo-location, presence, documents, signatures, etc., at source
• Low-code/no-code platform with preconfigured domains or workflows
• AI-driven governance: detects duplicates, anomalies, and inconsistencies automatically
• Cloud-native but cloud-agnostic
• Multimodal data support
• Process-aware and process-rich
• Orchestration, not automation

Why is it important now?

AI is redefining the way decisions are made; those who are not ready to embrace AI will either fall behind or perish. Just as the internet redefined access and communication, AI will redefine context, intelligence, and adaptability. In such an environment, data is the foundation of digital transformation. Multimodal data must be captured, extracted, verified, and stored using a next-generation platform. The cost of manually processing multimodal data such as voice, video, images, and geospatial data  - through back-office resources will increase multi-fold, making it unviable for organizations to even capture the data.

Building data confidence

Organizations that adopt a modern data platform report faster onboarding, fewer process exceptions, cleaner analytics, and better collaboration between business and IT. Most importantly, they achieve something rare: data confidence - knowing that every report, workflow, and decision rests on a single, trusted version of the truth.

Because when data stops being a problem and starts being a strength, transformation stops being a buzzword and becomes reality.

Importance of Context Graph

The next wave of enterprise value creation will come from those who integrate multimodal data, processes, and intelligence into adaptive, business-ready systems.

The context gap, which exists in various forms such as emails, documents, meeting minutes, and audio/video recordings of meetings, needs to be captured.

Today, we know what decision was taken, but we don’t know why it was taken. For example, the system of record provides information about the discount given to a customer but does not provide the reasons behind it (i.e., the context graph is missing).

The AI platform can scan emails, documents, images, audio/video recordings of communications, meeting notes, etc., and prepare the context graph for the decisions taken by the organization. Over time, this will help organizations remain consistent in their decision-making process.

The context graph enriches the data by capturing the underlying intent and reasoning - the “WHY” behind the “WHAT.”

Experiments to Scale

2025 was the year of awareness and experiments; 2026 is likely to be the year of scale. Those who are data-ready will start seeing success in their AI initiatives. Apart from business alignment, governance structures, and KPIs, organizations will have to ensure 100% data quality and process agility. They need to migrate to next-generation master data management and business process transformation platforms and make sure that data and processes work cohesively. AI initiatives will not succeed without clean data and integrated processes.

Author: Suresh Anantpurkar, Founder & CEO at Manch Technologies