As with anything new, often we are drawn to the latest innovations, without fully considering the underlying fundamentals and what is required to leverage these innovations. In recent times, ‘Artificial Intelligence’ and ‘AI’ have become established buzz phrases – umbrella terms encompassing a range of different technologies, which are then applied across many different use cases. Consequently, for many businesses, it is more appealing to focus on showier but more limited prototypes, rather than investing in underpinning capabilities to enable longevity.
Coordinated approach
Generating useful and sustainable value from anything and everything that comes under the AI umbrella requires some significant groundwork; a co-ordinated approach to ensure cohesion across organisational culture, data governance and technological strategy. Assuming the necessary work has been undertaken to assess the need for AI in the first place, which is also paramount, there are crucial steps that must be considered well before AI and machine learning (ML) tools can be consistently exploited to deliver Information Advantage.
While not a unique issue, the need to lay those foundations is particularly profound in the areas of Defence and National Security, where scaling any digital tool is challenging. The path from proof-of-concept to production-ready tools is financially treacherous, even without consideration for evergreening and support. While there are many enablers that organisations could use to scale and maintain AI effectively, arguably there is one that should be considered essential – the data platform.
Data platforms
Data platforms provide the necessary infrastructure to support data applications, turning raw data into a strategic asset by leveraging a single source of truth. Trustworthy and reusable data is a vital prerequisite to any analytics, especially AI, and combining that data needs to be done in a cyber-secure and access-controlled manner.
The term ‘platform’ can be interpreted in many ways. Within Leonardo’s UK cyber business, we refer to a ‘data platform’ as a digital fabric built with a microservice architecture, which enables ingestion of various data sources, ensuring they can be seamlessly accessed, integrated and scaled. Without a capable secure data platform, it is impossible to gather data insights, create reliable AI systems or control risk within the digital system. Data platforms are also a key component in creating, managing and enforcing data governance, one of the biggest challenges faced by many organisations.
In an age where there is a lack of digital SQEP (Suitably Qualified and Experienced Personnel), automating data management tasks, and bolstering the platforms’ intelligence and operational efficiency can significantly reduce the resource required to do the time-consuming and mundane tasks like data cleaning and database identification. The platform will need to evolve with changing operational requirements, so version control of features, along with the ability to track data lineage, will significantly enhance the overall pipeline quality and systems’ ability to adapt at pace.
The unsung hero
Attempting any form of big data analytics – especially using AI/ML – without a suitable data platform in place, will result in a series of proof-of-concepts suitable only for use in isolated environments, rather than an operationally-appropriate tool that can consistently adapt to changing operational requirements and remain fit for purpose in a time with a rapidly transforming digital landscape.
Data platforms are the underappreciated bedrock powering the rise of AI. The ability to manage, process and scale data effectively, along with providing the necessary compute, tools and frameworks, are all critical to delivering successful and secure digital applications.