Establishing a robust, technology-driven data foundation is essential for enterprises aiming to harness AI’s full potential and navigate complex data challenges effectively, as Mark Venables explains
In today’s evolving digital landscape, AI has become a critical tool for organisations seeking to innovate and gain a competitive edge. However, while AI’s potential is widely acknowledged, many enterprises face significant obstacles in harnessing it effectively. A recent insight produced by MIT Technology Review in partnership with Snowflake, identifies a considerable challenge: establishing a solid data foundation supporting AI at scale. Data complexity, silos, and governance limitations are cited as common barriers, preventing companies from fully capitalising on AI’s transformative capabilities.
Prasanna Krishnan, Head of Collaboration and Horizon at Snowflake, reinforces that for AI to deliver actionable insights, organisations must develop a resilient, technology-driven data strategy. “A strong data foundation brings data together, controls access, and enables insights,” she says. This holistic approach requires overcoming common barriers such as fragmented data sources, governance constraints, and the need for efficient, scalable infrastructure.
The challenge of data silos
For many enterprises, data silos remain a key challenge in building a comprehensive AI strategy. Dispersed across different departments and stored in multiple formats, data often resides in isolated systems such as ERP, CRM, and IoT platforms. This fragmentation limits the organisation’s ability to create a unified data view, undermining attempts to use AI meaningfully. “Many companies have data stored in separate systems or formats, which makes it difficult to analyse or use together effectively,” explains Krishnan.
A unified data strategy is critical in resolving this. By consolidating structured, semi-structured, and unstructured data within a single, accessible framework, organisations can perform meaningful, cross-functional analyses. As Krishnan emphasises, “A strong data foundation supports each stage of the data journey—not only centralising data but enabling its application.”
Furthermore, unstructured data – documents, emails, and customer interactions – represents a wealth of potential insights but is challenging to analyse with conventional methods. While many enterprises already capture and store unstructured data, tapping into its full value requires advanced capabilities such as natural language processing (NLP). By embedding NLP into their data strategy, companies can derive new insights from previously untapped sources, significantly improving decision-making and customer insights.
Ensuring governance and access
Even with a unified data foundation, effective governance remains essential. Enterprise AI requires strict control over data access to ensure compliance with regulatory standards, particularly in the finance, healthcare, and manufacturing sectors. As noted in the MIT Technology Review Insights report, more than half of surveyed executives express concerns about data security and governance, emphasising that it is a priority as they adopt AI solutions.
“Even with all data centralised, companies still need control over who accesses it and must protect sensitive information,” says Krishnan. “A robust governance model involves setting permissions at granular levels, enabling organisations to manage access while ensuring that sensitive data remains protected. Transparent access logs and data lineage tracking further reinforce compliance, helping companies meet regulatory requirements while securing their assets.”
Data governance is not solely about compliance, however. It also enables organisations to broaden data access within the company, empowering employees across functions to make data-driven decisions. “A strong data foundation not only controls access but also ensures that insights are accessible to those who need them,” Krishnan continues. “This data democratisation encourages innovation and agility, equipping employees with real-time information to enhance operations and customer experiences.”
From data to insights
With consolidated data and governance in place, organisations can transform information into actionable insights. However, many enterprises struggle to extract its full value even with the correct data. Data quality issues often hinder AI applications, leading to unreliable insights that can undermine confidence in AI-driven decisions.
According to the MIT Technology Review Insights report, over 50% of executives cite data quality as a top challenge in deploying AI. The data feeding AI systems must be clean, reliable, and current for meaningful insights. “If your data isn’t reliable, then your insights won’t be either,” says Krishnan, reinforcing the importance of data quality in AI strategies. “Another essential factor is the timeliness of data.”
For organisations operating in fast-paced environments, up-to-date information is crucial. Real-time data-sharing capabilities allow different departments to access the latest insights without delays associated with traditional data transfers. Ensuring timeliness is as critical as data quality,” Krishnan adds, noting that real-time access enables organisations to respond proactively to market changes or operational demands.
Integrating AI-powered tools such as NLP and machine learning (ML) into the data foundation enhances accessibility, enabling users to interact with data in more intuitive ways. NLP allows employees across departments to query data using natural language, reducing dependency on technical expertise and fostering a culture where data-driven decisions become the standard.
Scaling AI across the organisation
AI adoption often begins with pilot projects, yet organisations must scale these initiatives across the enterprise to realise their potential. Scaling AI effectively involves addressing new challenges, such as increased data loads, computational demands, and the requirement to uphold governance standards as data use grows.
“Scaling AI typically starts with experimentation,” Krishnan says. “A single team might run a pilot application, but expanding that to an entire organisation involves new challenges,” For AI to become integral across functions, a flexible, scalable infrastructure is essential. This means building an architecture that can adapt as demand increases without disruptive changes. “Scaling AI requires infrastructure that allows for seamless growth, much like increasing electricity usage as demand grows,” Krishnan adds, highlighting the need for elasticity.
Organisations may face compatibility issues in a multi-cloud environment as they attempt to integrate AI efforts across departments using different platforms. Cross-cloud compatibility supports scalability by allowing data and applications to function seamlessly regardless of cloud provider. This approach enables the organisation to operate across various cloud environments, avoiding the challenges that typically arise from cloud-specific constraints.
Controlling costs in AI initiatives
As with any large-scale project, the financial implications of AI adoption require careful consideration. The MIT Technology Review Insights report identifies cost as a significant barrier, with 48% of business leaders listing it among their primary concerns. Generative AI and machine learning applications require substantial computational power, and storage and processing costs can add up quickly.
“Cost control is crucial when implementing AI,” Krishnan notes. “Organisations benefit from adopting a flexible, consumption-based pricing model that aligns costs with actual usage, particularly valuable in the experimental stages of AI adoption. When exploring AI, companies often start by experimenting with models. This requires the ability to test different options without high upfront costs.” As the organisation identifies the most effective approaches, it can scale up its usage, ensuring cost efficiency while gaining insights into the ROI of various AI models.
Organisations can also manage costs by optimising their choice of AI models, selecting those that best meet specific needs without excess computational burden. Krishnan notes that experimenting with multiple models and selecting the most suitable one is a cost-effective way to maintain performance without overspending.
Future-proofing the data foundation
As AI continues to shape the business landscape, building a resilient, adaptable data foundation has become essential for long-term success. A future-proof data strategy supports each stage of the data journey, from ingestion to AI applications and insights sharing, preparing organisations to grow alongside emerging technologies.
“A strong data foundation is not only built for today but designed to evolve as AI capabilities expand,” Krishnan says. “By unifying data within a centralised, technology-driven framework, organisations gain a foundation capable of supporting AI at scale across departments and use cases. This approach also allows for collaboration across the organisation and with external partners, such as suppliers and clients, which extends the benefits of data-driven insights.
“Integrating data from existing systems, such as ERP and CRM, enhances this adaptability without requiring disruptive changes to current workflows. For instance, an organisation might have data in an ERP system, CRM data, and IoT data stored in various silos. A future-proof data foundation can pull these sources together into one centralised location.” This approach allows organisations to enrich their insights with data from across functions, creating a more accurate understanding of operations and customer behaviours.
Preparing for an AI-driven future
As AI becomes more ingrained in organisational strategy, the demands on data infrastructure will continue to grow. The MIT Technology Review Insights report highlights the strategic importance of a technology-driven data foundation in meeting these demands. With an approach focused on unifying data, enforcing governance, and ensuring quality and scalability, organisations can leverage AI to its full potential.
A resilient data strategy does not simply address today’s needs but prepares organisations to meet future challenges and opportunities. As Krishnan summarises, “A strong data foundation isn’t just about addressing current needs, it is about building a platform that supports the organisation’s growth and innovation for years to come.”
By embracing the proper data practices, tools, and governance standards, companies can transform data from a passive asset into an active growth enabler, empowering teams to make faster, smarter decisions and innovate confidently. In this AI-driven era, organisations that invest in their data foundations are prepared to navigate today’s challenges and positioned to define the future of enterprise AI.
