At McKinsey’s second annual R&D Leaders Forum, more than 50 senior executives in innovation, R&D, and engineering convened to explore how AI is reshaping the future of product development and scientific research. This year, the focus of discussions moved from theoretical AI capabilities to practical applications—on using AI as a collaborative tool to accelerate innovation and streamline development in R&D and engineering across industries such as automotive, aerospace, heavy industrials, and medtech, where it aids in virtual simulation, regulatory processes, and molecular discovery.
While gen AI has demonstrated considerable success within the software sector, the hardware domain has remained a persistent challenge for many organizations. AI is increasingly impacting R&D and engineering operations, however, with established companies adopting it across a handful of use cases. Recent advancements in foundation models have shown competitive performance in engineering design and scientific simulation tasks. Additionally, new agentic architectures are facilitating autonomous processes in lab experimentation, product simulation, and iterative testing, positioning AI as an active collaborator within teams, rather than a passive assistant.
Rewards for the bold
AI-powered R&D presents a compelling opportunity to reverse longstanding productivity slowdowns—but only for those organizations taking bold steps to reskill teams, overcome cultural and technical barriers, and rethink legacy processes. Early adopters that moved quickly are widening the gap in AI outcomes versus those whose AI uptake has lagged. This impact is seen across operations KPIs, as shown in the chart below. This article explores how organizations can accelerate progress, scale AI successfully, and reach full maturity within hardware engineering domains.
Three major themes emerged from this year’s discussions.
1. Collaborating with AI to accelerate the product development cycle
AI is revolutionizing discovery and design by reshaping R&D processes through innovative methodologies. In virtual laboratories, AI agents collaborate with human scientists to address complex challenges, autonomously generating queries, analyzing literature, and brainstorming with each other to accelerate breakthroughs.
Notable scientific AI examples include the development of nanobodies for COVID-19,1 where domain-specific agents showcased the innovation potential of a “Virtual Lab”: an AI-human research collaboration that performs sophisticated, interdisciplinary scientific research. The Virtual Lab created a novel computational nanobody design pipeline and proposed nanobody candidates to test experimentally.
Additional research from McKinsey’s Life Sciences Practice highlights R&D as a leading area for gen AI adoption in medtech. Tailored AI tools have improved productivity by 20 to 30 percent by streamlining the creation of documentation and other manual tasks. This shift enables employees to redirect time toward higher-value activities such as research, innovation, and strategic problem solving.
In mechanical engineering, gen AI enables engineers to produce 3D component designs via natural language prompts. For instance, AI can significantly reduce the detail design time for steel structures such as building frameworks from several weeks to a few hours. Meanwhile, deep learning surrogates (DLS) trained on high-fidelity simulation data provide fast and accurate approximations of structural, aerodynamic, or thermal systems, allowing for rapid iteration and halving development time. This method is increasingly used in new-product introduction workflows in automotive, power generation, and aerospace. Companies adopting this AI-supported approach could see a reduction in rework of more than 20 percent, as DLS technology has the ability to enhance accuracy by processing large volumes of complex engineering data, and identify patterns and correlations while minimizing errors.
2. Generative AI is transforming product ideation
Gen AI continues to reshape industries, with nearly half a trillion dollars of estimated impact in innovation potential from AI applied to R&D.2 The focus now shifts to how organizations can strategically deploy this technology to boost productivity, reduce manual effort, and prioritize high-impact activities.
One promising application is in product ideation and concept testing. Rapid consumer research, focused on key expectations and experiences, can inform initial design elements, which gen AI can then transform into visual renderings or product concepts. From early-stage idea generation to AI-assisted formulation, gen AI serves as a thought partner, identifying emerging trends and accelerating go-to-market execution. For example, Nestlé has leveraged AI to generate new product concepts aligned with consumer preferences, accelerating innovation efforts.3
These consumer insights can be transformed into actionable business strategies, by using large language models (LLMs) to create synthetic personas—AI-generated proxies for real consumers trained to reflect human behavior and preferences.4 These personas can be deployed to test hypotheses, evaluate new product ideas, and uncover pain points, reducing reliance on traditional consumer panels and shortening research cycles. Research shows that synthetic personas have replicated human feedback with 70 to 95 percent accuracy5 in reflecting consumer responses, enabling companies to pinpoint feature-level preferences and design more niche, targeted offerings with a strong product-market fit.6
Elsewhere, in molecular discovery, AI can leverage curated data from vast chemical libraries to map molecular structures to physical properties and identify innovative material candidates. Downstream, the automation of chemical synthesis, material handling, and wet lab experiments allows targeted screening. By incorporating AI into computational chemistry research, AI can learn the underlying physics and atomic interactions to more efficiently select for desired properties and aid in use cases such as finding new therapeutics, refrigerants, or cytotoxic compounds.
By leveraging gen AI to simulate and synthesize consumer feedback, companies can unlock data-driven innovation faster, creating products that deliver greater value to their target audiences.
3. AI is augmenting engineers and teams
AI is rapidly becoming a force multiplier for engineers and R&D teams—automating repetitive tasks, streamlining workflows, and synthesizing insights from vast and complex data sources. One clear example is found in technical documentation: AI can now generate up to 90 percent of a finalized specification document by leveraging previous specs, templates, and engineer input—particularly in highly regulated sectors such as pharmaceuticals and aerospace manufacturing. In the medical field, AI can generate initial protocols by analyzing historical trials, earlier phase data, and standardized templates, significantly reducing manual workload and accelerating development cycles.
These improved processes enable engineers to use AI to simulate design alternatives and conduct root-cause analysis by detecting patterns in historical data and accelerating iteration. Acting as a collaborative partner, AI amplifies human expertise and creates capacity for teams to focus on innovation and strategic problem solving.
Agentic AI takes this further, autonomously simplifying multistep workflows like literature reviews, experimental design, and interpretation. These systems break down silos by linking discovery, analysis, and execution in a continuous loop. As this matures, engineers will shift from task execution to orchestration, fundamentally reshaping the innovation process.
Taking the next step
The leaders at this year’s event agreed that AI is a competitive differentiator and that integrating it into the product development life cycle will be the next big thing. The automation of processes that were once manual can make all the difference. One leader asserted: “If you are not moving on this now, you will be left behind.”
It is therefore essential to address the challenges of adoption—rethinking workflows, developing talent, and refining operating models. Moreover, the consensus at this year’s meeting was unequivocal: AI demands strategic investment and purposeful integration, with an ongoing series of strategic build-versus-buy decisions and experts customizing AI agents and applying specialized knowledge. To maintain a competitive edge, companies must be able to adapt and combine open-source and licensable models to their specific needs. As AI continues to evolve, one thing is clear: Those who lead in adoption will lead in growth.
Here at McKinsey’s Product Development Practice, we are looking forward to another year of exponential progress. And we're offering follow-up workshops to support your AI efforts in engineering and R&D. Contact our team here: rd_leaders_forum@mckinsey.com
Be part of the conversation with McKinsey Talks Operations. Learn more and register for events with our Operations experts here.
The authors wish to thank Jennifer Hou, Najaf Mooraj, Ruijun Zheng, and Trevor Thorburn for their contributions to this blog post.
1 Kyle Swanson et al., “The Virtual Lab of AI agents designs new SARS-CoV-2 nanobodies,” Nature, July 29, 2025.
2 “The next innovation revolution—powered by AI,” QuantumBlack, AI by McKinsey, June 20, 2025.
3 Veeral Shah, “Unlocking new opportunities with gen AI,” Nestlé, June 2024.
4 “How beauty players can scale gen AI in 2025,” McKinsey, January 6, 2025.
5 “How we compare results—generative AI for causal experimentation,” Subconscious.ai, accessed September 10.
6 Jessica Moulton, Rob Cain, and Roger Roberts, “Fortune or fiction? The real value of a digital and AI transformation in CPG,” McKinsey, October 3, 2024.