IBM predicts enterprises will see AI returns in 18-24 months with strong data foundations

Siddesh Naik emphasized that successful AI adoption requires more than model deployment, noting that around 80% of effort goes into data preparation, while returns on smaller use cases are faster than full-scale, long-term AI transformations

IBM India’s Data and AI Software head, Siddesh Naik, has projected that enterprises can expect significant returns on their artificial intelligence (AI) and generative AI (Gen AI) investments within 18 to 24 months—provided they adopt a robust data strategy and infrastructure.

Speaking to a business daily, Naik emphasized that a comprehensive AI adoption journey involves much more than simply deploying models. “You need the right use case and data foundation. For smaller, siloed use cases, returns may come quickly, but a full-scale AI transformation takes time,” he said. According to Naik, about 80% of AI project efforts are dedicated to data preparation, with only 20% focused on AI model development.

The key challenge, Naik explained, lies in creating a streamlined and standardized data ecosystem. This includes collecting data from diverse sources, ensuring data quality, transforming it into uniform formats, and implementing governance policies to track data lineage. Without these critical steps, AI adoption risks remaining fragmented and inefficient.

Naik advocates a phased approach to building this foundation: starting with data collection and quality assurance, followed by setting up data pipelines and ETL (extract, transform, load) processes, then integrating modern data platforms like lakehouses for reporting, and finally applying governance and lineage frameworks. “A step-by-step approach ensures a scalable and sustainable AI strategy,” he added.

People, process key to scaling

Industry experts echo these sentiments. A Boston Consulting Group report last year highlighted that scaling AI depends largely on people and processes, including change management, talent acquisition, and workflow optimization. Data quality and management are crucial technological factors, while AI model performance remains a top priority.

Earlier this year, Wipro CTO Sandhya Arun warned that many AI proof-of-concepts (PoCs) fail to deliver business value without strong policies and data foundations, often ending up as mere experiments.