Why Ignoring Data Is Derailing AI-Driven Digital Transformation

Hari Sonnenahalli is a thought leader and seasoned enterprise architect at NTT Data Business Solutions (NDBS).

In the race of digital transformation, artificial intelligence (AI) is often positioned as a game changer. From personalized customer experiences to predictive analytics, organizations are investing heavily in AI-driven solutions to gain a competitive edge. Yet, despite this surge in investment, many companies still struggle with something as basic and critical as producing reliable sales forecasts.

So, what’s holding them back? It’s not the lack of technology—it’s the lack of data readiness.

Data Isn’t Just The Fuel—It’s The Engine

AI isn’t a magic bullet. It depends on accurate, timely and complete data to function correctly. If that data is fragmented, outdated or biased, the AI outputs, no matter how advanced the algorithm, will be flawed. And as Thomas Redman, author of People and Data: Uniting to Transform Your Business, argues, “If your data is bad, your machine learning tools are useless.”

Why Companies Overlook The Data Problem

Despite the clear connection between data and successful AI outcomes, many organizations underinvest in data quality and infrastructure. Why?

Data work is invisible. Data cleaning, governance and labeling are behind-the-scenes efforts that rarely get recognition or budget.

Vendors drive the hype. Many AI tools are marketed as plug-and-play, encouraging the false belief that robust outcomes require minimal data effort.

Leadership has blind spots. Despite the surge in AI investments, most organizations still struggle to become truly data-driven. Although many implement advanced technologies, few have successfully embedded data into their core decision-making processes or company culture.

A technology-first approach is taken. Many companies prioritize purchasing AI tools over building a foundational data strategy.

The result? A shiny AI implementation on top of a fragile data core.

The Hidden Cost: Inward-Looking Sales Forecasting

Sales forecasting should be one of the most obvious beneficiaries of AI. With the right data, companies could anticipate buyer behavior and model demand fluctuations to allocate resources more efficiently. But instead, many forecasts remain outward-looking, internally biased and narrowly scoped.

The core issue is that many sales forecasting systems are built on fragmented, siloed internal data. Imagine a B2B SaaS firm using an AI-powered forecasting tool. CRM data is inconsistent, marketing signals are missing, and churn indicators aren’t tracked. The AI ends up mirroring internal bias, reflecting how the business sees its pipeline, not how the market is moving. The result? A forecast trapped in an echo chamber.

To break this cycle, organizations must shift from AI-first to data-first thinking. That means laying the groundwork with data infrastructure before expecting AI to deliver insight or automation.

Practical Steps For Change

It’s imperative to understand where your organization stands in terms of data maturity. Below is the framework that I’ve used to lay the foundation to build tailored data strategies.

Unify your data architecture.

AI models rely on accessible, cross-departmental data to deliver meaningful insights. Achieving this requires a seamless data strategy that begins by unifying data from sales, marketing, finance and operations. In my experience with enterprise teams, cloud-based data lakes and lakehouse architectures broke down silos and enabled real-time, shared access. Scalable hybrid models further ensured consistency, whether teams worked remotely or on-site.

Why It Matters: Seamless AI outcomes rely on data that’s not trapped in legacy systems or owned by one function.

Establish strong data governance.

In enterprise architecture, the data architecture model ranks data governance just after business strategy, yet it’s often overlooked. Governance goes beyond compliance—it builds trust and helps prevent issues like redundancy and poor synchronization, especially when handling massive CRM datasets across ERP systems. Without it, operations are at risk.

Data quality is equally crucial. Even advanced AI fails with flawed inputs. In a retail project, we removed thousands of duplicate CRM entries through automated cleansing and real-time validation, which improved forecast accuracy. Monitoring consistency and completeness made the data reliable for AI.

Why It Matters: Governance builds trust in data and reduces the risk of AI making flawed or biased decisions.

Enable real-time data pipelines.

In today’s market, speed is a strategy achieved by combining historical data models with real-time data feeds. An example is the real-time analytics platform I worked on for the San Francisco 49ers executive team that enhanced game-day decisions by streaming data from ticket scans, concessions and foot traffic into a live dashboard. This mirrors the real-time data pipelines I’ve built in commercial projects, where external signals like customer behavior or pricing trends feed directly into forecasting models, enabling rapid, informed decisions.

Why It Matters: Sales forecasting and predictive analytics need live data, not weekly reports.

Implement interoperability and API-first designs.

Too often, data systems are built in isolation, limiting integration and flexibility. Many companies stick to outdated technologies out of habit or lack of exposure to modern solutions. Although legacy systems can function, embracing newer, API-driven architectures brings greater agility and feature richness. These modern frameworks, like Due API, are lightweight and reduce development cycles, positively impacting timelines and budgets. In one project, I implemented an API-first design using SAP to streamline the production order process and improve efficiency.

Why It Matters: Seamless doesn’t mean centralized—it means connected and cooperative.

Encourage data literacy across the business.

A data strategy only succeeds when people understand how to use it. The 49ers analytics platform illustrates how front-line staff and executives alike can make real-time decisions when data is democratized and accessible. I’ve seen similar success by leading workshops where cross-functional teams learned to interpret dashboards, ask the right questions and use low-code tools to extract insights. This not only improves decision making but also fosters a true data-first culture.

Why It Matters: AI is downstream from data. You can’t automate what you can’t understand.

Looking Forward: Data-Driven Forecasting As A Competitive Edge

Organizations that treat data quality and integration as strategic priorities can unlock AI’s real value. These companies can forecast more accurately, adapt faster and deliver smarter experiences. Until that shift happens, AI will often underdeliver, and forecasting will likely remain tedious.

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?