VIVAERO

Change agent

PTC’s business development director, David Hughes explores how increased funding is set to accelerate artificial intelligence’s hold on the aerospace and defence (A&D) sector. However, with speed comes the risk of firms accruing ‘technical debt’ if they move too quickly. 

The much-heralded Sovereign AI Fund opened for business earlier this month at Wayve’s London headquarters. Backed by £500 million from the Department for Science, Innovation and Technology, it marks the biggest indicator yet of the UK government’s desire to grow Britain’s 6,000-strong artificial intelligence (AI) community and lock in some of the brightest minds in the business. 

For any new technology to be truly successful, there must be a solid business case and the A&D sectors - bound by their own unique sovereign requirements - seem very easy bedfellows. It’s not difficult to see why. The power of AI to leverage proven data to deliver productivity gains and speed to market is enormous and these two highly technical and regulated industries have massive economies of scale.

After all, most primes and OEMs boast multi-site, global operations and run programmes that can last for many years. These highly complex cycles can feature a multitude of products and variants, meaning the cost of engineering change across design, manufacturing and maintenance, can run into millions.

“This very scenario can typically take months to analyse, yet if you get AI right, I could see the big A&D companies completing initial change impact analysis in a matter of hours,” begins David Hughes, director at digital transformation specialist PTC. “Can you imagine the financial and productivity benefits this would bring? It has the potential to transform the sector through the deployment of Agentic AI to achieve specific goals for planning, reasoning and acting with vastly reduced human intervention. Of course, humans will remain in the loop in this sector, but more from a governance perspective. This is what all the firms in A&D will be working towards and it will happen soon - naturally, with varying degrees of success.”

Aerospace will be seen as a priority sector for the Sovereign AI fund after consistently being flagged as a tier 1 beneficiary due to it sitting at the critical intersection of civil aviation, defence and space and export-controlled technologies.

In short, AI applied by the airframe manufacturers, the powertrain developers and the component producers directly impacts national security, defence readiness and supply chain resilience.

“The race is on, and we’re seeing that from the number of enquiries our teams are receiving about how our Digital Engineering portfolio can accelerate the AI journey, adds Hughes. “It’s exciting, but we also need to be cautious. Without strong data foundations in place, organisations risk creating technical debt rather than gaining a competitive advantage.

“Whilst A&D manufacturers are eager to adopt AI-driven tools, we must learn from the mistakes of previous technology waves. For example, during the rapid adoption of new systems in the early 2000s, many companies prioritised speed over sustainability, leading to a proliferation of disconnected custom-built applications. 

“These systems were often difficult to manage, scale or maintain, leaving businesses burdened with long-term complexity. The results were often a plethora of software applications with disconnected, inaccessible information silos.”

The adoption of AI carries a similar risk profile today warns Hughes: “The relative ease of building AI tools, which will obviously be accelerated by the £500m Sovereign fund, means businesses run the risk of creating ad-hoc capability that is difficult to maintain and scale – and we’ve been here before. 

“Hopefully we’ve learned the lessons of the past and can avoid building costly technical debt and focus instead on building capability that can be scaled and maintained into the future.

“Businesses have enterprise systems of record that are domain specific. Real value has always been derived when accessing these different sources of truth and bringing that information together in a rational way. Traditionally, this has always been dependant on many humans being in the loop. 

“AI has the potential to significantly scale value here, but this sector is of course heavily constrained by sovereign security issues, particularly the defence sector. A good example is the use of large language models (LLM), and the emergence of publicly available reasoning engines. Quite understandably, this market may not tolerate this model, so we could find A&D primes hosting their own LLMs, whether they be generic LLMs or domain-specific, to ensure they access domains that carry minimal security risks.”

Data is the real currency...According to Hughes, the critical factor in any successful AI strategy is the quality of the underlying data foundations and the data that organisations have accumulated over time. Those data foundations need to be rock solid for Artificial Intelligence to have a true benefit, otherwise organisations will just deliver poor results much faster.

For A&D companies, awash with data at every turn, it poses the real question – what is a data foundation and what data is stored there? In response, it helps to sub-divide data into two broad categories – persistent data and transient data.

“Persistent data refers to systems of records. If we think of PLM and ERP as two main systems of record for design and build, persistent data will define what the product is and how to build it and must endure over a long period of time. 

“This includes data that defines the product - such as CAD, BOM structures and associated artifacts - and data that defines how production is executed and records what happens. Item masters, supplier info, production orders and financial record are some examples.

“This data must be traceable, version controlled, auditable and re-usable across the complete lifecycle over long periods of time. There are of course other authoritative systems of record that would be included.

“Transient data, on the other hand, is supporting data that represents ‘current state’. For engineering, this would include draft design explorations and supporting data not ultimately approved. For manufacturing, this would include shopfloor execution signals such as machine uptime, production tracking for WIP, real-time data that can be used to understand current shift performance and identify bottlenecks.

“Once the process has been completed, the data is typically historized for supporting production analytics, rather than forming part of the system of record. So, in short, ‘persistent’ is data that must remain true over a long period of time regarding the product and operations (system of record), whilst ‘transient’ is data pertaining to what is happening right now, to achieve the output (system of process).”

PTC works with many companies globally on their digital transformation initiatives, which includes UK A&D manufacturers. Much of this support focuses on consolidating product data, strengthening governance and ensuring core systems are fit for purpose - therefore setting the stage for Artificial Intelligence to be deployed responsibly and securely.

In addition to helping customers with ‘overarching’ agentic AI requirements, it has also been busy embedding AI directly into its portfolio offerings. For example, Creo, its CAD solution, has had AI embedded into its generative design features for some time and real value is also being derived from the later releases of Windchill, its PLM solution, with capabilities such as AI-driven parts rationalisation.

The solutions are not just targeted at primes and OEMs. There’s probably an even bigger argument that SMEs should be exploring AI quicker, as they don’t tend to have the human resources in play that the larger businesses can call upon.

“There’s definitely an appetite from all levels of the supply chain to look into agentic AI and lots of money flowing into the sector will accelerate adoption...there’s no question about that,” concludes Hughes. “If this is done with strong data foundations and at a pace that consistently mitigates the possibility of ‘technical debt’, then AI can truly elevate the digital thread to a new level in A&D.”

www.ptc.com