How agile approaches are transforming Defence data projects
In the Defence sector, quick, data-driven decisions can make the difference between success and failure on operations. Traditionally, Defence IT projects have concentrated on large-scale initiatives that take time to show results. Now, a shift towards agile working methods is changing this. This blog explores how smaller-scale data projects and a focus on "early release of value" is the answer.
Breaking down data silos
One of the biggest challenges in Defence is data silos. Different parts of the organisation store their data separately, making it hard to access and share information. Successful projects tend to be stand-alone. Andy McLannahan, Defence Director at Made Tech, explains how ‘the Ministry of Defence (MOD) is federated and the information architecture has grown organically meaning often work independently,’ This slows down decision-making because data isn’t easily available across teams and platforms. The UK's Data Strategy for Defence flagged this issue in 2021, noting that
although Defence generates vast amounts of data, much of it stays locked away in different units. Improving data sharing across platforms could allow MOD teams to better predict resources and respond quicker to threats. Rather than trying to fix everything at once, agile projects focus on smaller steps to unlock the value of data. For example, the Defence Information Platform is consolidating over 25 platforms into one, making it easier to find and manage data. McLannahan explains that even on ambitious projects delivering small, quick wins, rather than waiting years for a big project to complete, is more likely to bring success.
Using AI and agile to speed up decisions
As more data becomes available, Defence organisations are looking at ways to use AI to make faster, better decisions. By 2030, it’s estimated that up to 30% of Defence processes could be automated using AI. But to tap into this potential, agile methods are needed for rapid testing and scaling of AI tools. A ‘fail fast, learn fast’ mindset is key to quick innovation. By conducting pilot projects and iterating based on feedback, Defence can refine AI models quickly without risking large failures. This also helps create a culture of continuous learning.
Shifting to early release of value
In the past, Defence projects followed a waterfall model – large, multi-year plans with all the value delivered at the end. Now, initiatives like Defence Digital Foundry are shifting to agile methods, delivering value in stages rather than waiting for one big project cycle to finish. It’s estimated that the MOD could save billions of pounds by adopting agile approaches across its operations. Agile methodologies align equally well with NCSC’s modern security practices. McLannahan explains that security is now ‘baked in’ from the start of a project rather than being checked at the end. This approach supports safer iterative delivery of new applications and systems.
Addressing data security
While AI offers significant potential to improve decision-making and automate processes, it also introduces new security risks. For example, large language models (LLMs) trained on Defence data could unintentionally leak classified information, or AI systems could be manipulated by adversaries to provide false or misleading outputs. Organisations are understandably cautious about sharing sensitive data with LLMs. Agile approaches help tackle this challenge by allowing defence teams to test and refine AI models in smaller, more secure steps.
The future is agile
The key to delivering value quickly, and maximising the advantages of AI lies in adopting agile methods. This approach will be critical if we want to stay responsive to the ever-evolving challenges in defence. For more insights, visit our data and AI services pages or take a look at the government Defence AI Playbook.
