Why the key to scaling AI is building an effective data strategy

By Andrew Weaver, Principal Product Specialist, Databricks.

ChatGPT has taken the world by storm, bringing the capabilities of generative AI to public prominence unlike ever before. In reality, the core components of ChatGPT have been around for a while, but few companies are at the maturity level required to combine them effectively. As this use of AI increases in the modern world, it’s essential to determine how both AI and machine learning (ML) applications can be scaled to meet specific business needs. The key to this is building a solid enterprise data strategy.

Most leaders recognise that AI has the potential to change how organisations work but while pockets of success bring value, it is scaling these across the organisation that can be game changing. Generating organisational-wide value through AI can not only bring competitive advantages, it can also help businesses be more efficient during a tough economic climate during which we all need to be able to do more with less.

So with this in mind, let’s examine not only the challenges and rewards of scaling AI, but also the key steps that leaders can take to ensure they are creating the most business value.

Challenges and rewards of scaling AI

Data and AI strategies are closely interlinked. Without ‘good’ data, AI cannot learn – therefore, leaders have to first make sure they are managing their data properly before AI use cases can be truly scaled across the organisation. According to a recent MIT report, which surveyed CIOS, CTOs, Chief Data and Analytics Officers – IT leaders are in agreement about this. More than three-quarters (78%) of the executives surveyed - and almost all (96%) of the leader group - say that scaling AI and ML use cases to create business value is their top priority for enterprise data strategy over the next three years.

For most businesses, scaling individual AI successes across the enterprise is the exception rather than the norm. However, most executives surveyed in the MIT report are looking to embark on a major expansion of AI use cases in all core functions in the next three years. More than half expect AI use to be widespread or critical in their IT, finance, product development, sales, and other functions in the next two years.

There are significant rewards that can come from this. While most leaders will look to pursue a wide variety of use cases, a primary benefit of AI will be on the top line, increasing the returns from revenue-generating uses. In addition to automating key processes that save time and money, investing in development of AI can enable organisations to look towards innovations that can improve experiences for both employees, customers and prospective customers.

But the process of scaling AI isn’t without difficulty. For large organisations, it can be challenging to implement a data strategy across disparate teams with different job roles and objectives. Also, organisations that are still working with legacy data architectures, such as data warehouses, may not even be in a position to think about AI, let alone use it to deliver value. Looking forward, organisations that are not able to take full advantage of AI will take a backseat to those who can, making it crucial for business success to take a deeper look at the processes for managing their data.

The first step: building a strong foundation for data

In order to scale AI, data has to be easy to discover, manage, maintain and distribute. The first step is ensuring that you are building a strong foundation for data from the beginning. Legacy architectures such as data warehouses are costly to maintain, prevent data from being shared expediently, and can result in outdated or duplicated information being distributed. In the current economic context, businesses cannot afford for information silos to form, as this is kryptonite to innovation and scaling AI use cases.

A good example of this would be UK retailer, Marks and Spencer (M&S). In 2019, M&S launched its data strategy, with a vision to be the most data-driven retailer in the industry, empowering every colleague to make faster and more informed decisions. Its strategy focused on four key pillars: centralising its data into a unified cloud platform, improving data quality to increase colleague understanding and trust in data, growing data capability to develop and attract talent and delivering business value through scaling data science and AI. This move was transformative, with data insights allowing the retailer to understand its customers’ preferences – from their past purchases, online behaviours and responses to marketing communications and offers. To tackle issues with employee attraction and retention, M&S has also invested in training its employees in data science, having built several programmes aimed to improve data literacy throughout the organisation.

The MIT report also found that, when asked which aspects of their company’s data strategy is most in need of improvement in order to support their AI goals, speed of data processing was a priority for respondents. With this in mind, employing a platform that easily stores data for analysis, and allows for the timely and accurate flow of data, will be key in laying down the path for successfully scaling AI and ML use cases. Modern architectures, such as a lakehouse, bring together the benefits of both data lakes and warehouses, removing much of the complexity typically associated with these legacy systems. Employing a lakehouse will reduce the number of different platforms needed, easily storing data for AI and ML use cases and making rolling out a data strategy far easier.

The impact of an effective data strategy on AI

In the coming years, AI is undoubtedly going to change the way we live our lives, and the key to whether that is for the better, whether from an corporate, or even societal point of view is having a solid foundation of good data that we understand and trust.

Technology is already there with the power to transform every job on the planet - but in order to take full advantage of these innovations, leaders must understand how their data is either helping or hindering them on the path towards achieving their business objectives.

By Matt Aldridge, Principal Solutions Consultant, OpenText Cybersecurity.
By Josh Mesout, Chief Innovation Officer, Civo.
By Cedric Mauvielle, Vice President, Global IoT and Partnership, Console Connect.
By Lori MacVittie, F5 Distinguished Engineer.
By Vijay Raman, Vice President, Products & Technology at ibi.
By Anthony Mullen, VP Analyst at Gartner.