At a recent roundtable convened by the Royal Academy of Engineering and McKinsey, senior R&D and engineering leaders discussed how AI is reshaping workflows, expanding design possibilities, and enabling more integrated engineering models across sector.
The themes raised by the discussion were clear: AI is beginning to reshape R&D not through incremental efficiency gains, but by enabling more integrated engineering models, unlocking new design possibilities, and strengthening engineering teams.
Yet leaders also emphasised that realising this value requires disciplined implementation, careful problem definition, and organisational readiness. Technology alone does not create impact. The insights from the discussion offer a practical roadmap for engineering executives looking to capture AI’s potential in R&D.
Expanding the frontier of R&D productivity
Re-imagining workflows
Participants emphasised that AI can do more than streamline individual tasks. By absorbing and interpreting complex, high-volume data – such as video, simulation outputs or multidimensional product canvases – AI enables workflows that were previously impractical or resource-intensive. The opportunity lies not in automating what exists, but in redesigning how information flows across the R&D lifecycle embedding new and improved ways of working.
To capture this value, leaders stressed the importance of a clear understanding of how today’s workflows operate. Without that clarity, organisations risk reinforcing inefficiencies, rather than improving performance.
Accelerating design
Examples shared during the roundtable illustrated how deep-learning surrogates are dramatically compressing vehicle product design cycles. Instead of manually evaluating hundreds or thousands of possible variants, AI is testing and refining tens of thousands of design permutations in minutes in specific use cases. This broadening of the design space – coupled with rapid optimisation –means organisations are pursuing higher-performing solutions faster.
Crucially, participants emphasised that this is not about accelerating innovation, not replacing engineers. AI does not replace validated engineering processes – human judgement remains essential to ensure that AI-generated designs reflect and respect real-world constraints, from regulatory and safety requirements to manufacturability and customer needs.
Bridging scientific and engineering disciplines
Many engineering challenges require cross-functional expertise, tackling physics and chemistry questions to enhance material composition and performance, and bringing together mechanical, electrical, chemical, and digital disciplines. Leaders noted that AI is helping companies to better translate between models and data structures to help teams understand how decisions in one domain influence outcomes in another. Participants viewed this capability as a catalyst for more holistic, system-level innovation.
Applying AI to solve real business problems
A recurring theme of the discussion was the importance of disciplined scoping and implementation. Leaders cautioned against viewing AI as a universal solution, but as a tool that must be carefully scoped, rigorously tested, and deployed to address clearly defined challenges.
Clarifying problem definition
Successful AI initiatives begin with a rigorous understanding of the problem and the workflows surrounding it. Leaders must be clear about the objective and desired outcome from solving the problem, and the timeframe on which a return on investment is expected. Teams must confirm whether the receiving functions are prepared to act on AI-generated outputs and whether workflows will need to change accordingly. Without this alignment, acceleration in one stage can cause bottlenecks downstream.
Choosing the right level of capability
Not every challenge requires the most advanced tool. Simpler techniques often solve well-defined problems more reliably. Participants noted frequent cases where teams defaulted to the 'new shiny thing', without a clear problem definition. Participants also discussed the importance of being a 'smart AI customer' by asking questions about assurance platforms and data provenance to ensure the right product fit for the problem.
Maintaining engineering rigour
Participants emphasised that while AI accelerates iteration, testing and standards processes should not be hurried. Engineering teams making use of this technology must deeply understand the logic and processes underlying AI models, to maintain assurance and testing standards, and avoid wasted time unpicking mistakes or backtracking to understand outcomes. Leaders stressed the need for heightened standards of validation, robust testing and cross-checking throughout, and a thorough understanding of the logic and process before implementation.
AI as an enabler of engineers
Technology alone does not generate impact. The roundtable reinforced that organisational readiness – leadership conviction, team alignment, and stakeholder trust – is core to delivering value.
Embedding AI in how teams work
Realising the benefits of AI requires changes to roles, decision-making processes and expectations about how R&D teams operate. As with many changes, leaders must articulate how responsibilities evolve, what new skills are needed, and how teams should work with AI-enabled tools. Participants discussed the thinking tools that can help teams consider and articulate impact, risk and preparedness for teams, and the need to repeatedly revisit such questions to confirm and clarify. Without this clarity, even promising pilots fail to scale.
Bringing stakeholders along
For AI-enabled approaches to succeed, stakeholders – executives, regulators, and customers – should understand both the performance impact and the rationale behind them. Roundtable participants highlighted the importance of pairing strong measurement and evidence of performance with a compelling narrative, often in bounded innovation spaces e.g. sandboxes. This combination can build “safe enthusiasm”: confidence to innovate without compromising rigour.
What this means for UK engineering leaders
AI offers significant potential to accelerate innovation, unlock productivity improvements, and strengthen the competitiveness of the UK’s R&D ecosystem. To capture this opportunity, engineering leaders should focus on four priorities:
1. Anchor AI in real business problems to create value, redesigning workflows, not just individual tasks
2. Bring together cross-disciplinary teams to leverage AI’s integrative capabilities and drive R&D innovation
3. Deeply understand current workflows to ensure these are improved and streamlined as AI is deployed
4. Articulate both the narrative and measurement behind AI-driven approaches to build stakeholder trust
AI is not about the technology alone. It’s true impact will come from the judgement and expertise of engineers who understand how to apply it effectively. With thoughtful leadership and disciplined execution, AI can become a powerful enabler of the next wave of R&D innovation in the UK.