BLOG
Reinventing innovation for chemical research & development
5-MINUTE READ
August 19, 2024
I'd like to thank Paul Böhm for his valuable collaboration on this blog post.
To achieve the UN Sustainable Development Goals, the chemical industry requires innovation. For example, efforts could include developing new chemical processes to reduce emissions, discovering novel energy generation and storage solutions, and devising innovative food preservation techniques to combat hunger. There is no lack of innovation opportunities, but there are constraints on funding innovation. So, a significant improvement in innovation effectiveness is essential.
The chemical industry is a crucial enabler for solving these challenges, but doing so will require reinventing its approach to innovation. That means a fundamental change in the processes, organization and capabilities of today’s chemical company.
For example, a shift from:
- lab-based experiments to quantum-powered in silico experiments
- captive (in-house) R&D units to innovation ecosystems
- labs staffed with PhDs and technicians to innovation hubs that bring together chemists, data scientists and customers
Figure 1: Key changes in R&D at chemical companies
The journey toward generative chemistry
We expect the reinvention of innovation to be driven by three distinct levers: automation of lab experiments, in silico chemistry, and intelligent analytics that combine wet lab results, in silico insights and customer needs. This combination defines the journey toward generative chemistry. It replaces traditional trial-and-error approaches with technology-enabled, highly-focused, predictive and prescriptive experimentation.
Figure 2: Chemical R&D evolving along 3 dimensions
Three approaches driving the reinvention of innovation
Let us take a closer look:
In today’s chemical labs, 70–80% of tasks are still performed manually. The anecdote that a typical lab repeats an experiment every 3 years has some truth in it. The root causes are simple: Long-established lab environments with old equipment and unharmonized, disconnected systems hinder efficient exchange and transparency of data and insights. Consequently, there is high potential for automating simple, repetitive tasks (e.g., sampling).
Nowadays, the “lab of the future” is no longer just a vision. It has become a reality, enabled by technologies that can lead to a fivefoldi increase in lab throughput and that are ready for implementation. In these modern R&D labs, connected, automated and cloud-based lab-execution platforms change the planning, execution and analysis of experiments, providing faster, more accurate and safer processes. Real-time connectivity among instruments, data and scientists will facilitate full coordination of activities across labs and partner ecosystems. With this approach, experiment throughput can be increased fivefold while delivering unmatched accuracy and reproducibility of results.
Trial and error in the lab is a slow and costly approach, yet it remains common practice in chemical R&D. High performance computing (HPC) and quantum computing are transforming this practice by providing the computing power needed to run simulations of highly complex systems and reactions.
Much quicker than in a physical lab environment, these in silico experiments make it possible to rapidly test a wide range of formulations or chemical reaction pathways. These simulations can lead to cost reductions of 30–50%ii from early ideation to successful lab-scale product development. Chemical companies are looking for ways to use these capabilities for molecular modeling and for simulation to investigate molecular structures, properties and complex reactions. They also hope to use in silico experiments to achieve specific properties of materials.
HPC and quantum computing are a more cost-effective alternative to trial-and-error experimentation in a physical lab. However, computational chemistry won’t completely replace physical experimentation. Instead, it allows researchers to focus lab work on those experiments that are highly likely to be successful based on in silico pre-studies.
Digital technologies, particularly generative AI, are revolutionizing research and analytics in R&D. Today, they can search, analyze and interpret content from various sources to evaluate technical feasibility and market potential of new ideas. They can evaluate and draw conclusions from results of complex analysis like nuclear magnetic resonance or infrared spectroscopy. They can even create text to describe technologies or experiment outcomes. Digital technologies can increase average success rates of R&D projects by up to 70%iii and reduce the manual workload of highly skilled research professionals.
For example, AI can perform complex content searches not only for text but also for concepts and relevant information. It can search structured data and unstructured data, such as market reports and scientific literature, to quickly validate the feasibility and attractiveness of new ideas. AI-based analytics can also predict potential chemical properties, define new chemical formulations and provide recommendations for experiments to create chemicals with desired properties.
Across industries, companies are pursuing gen AI for R&D as a competitive advantage. Our research finds that “reinventors,” who outperform peers on financial and non-financial metrics, see R&D as the top area where they expect to use gen AI to reinvent how they operate in the next three years.
Exploring the realm of possibility
So, what does good look like?
Let’s envision the art of the possible by following a chemical R&D manager, Lauren, throughout her day:
Figure 3: A chemical R&D Manager’s day
As you can see from Lauren’s role, R&D’s use of generative chemistry will shift her lab’s experiments from a traditional trial-and-error approach to tech-enabled, highly-focused, predictive and prescriptive experimentation.
Generative chemistry will further reinvent the R&D function of chemical companies with remote monitoring of lab processes, globally connected labs and more (see Figure 4).
Figure 4: Generative chemistry reinvents R&D
What chemical companies need to do
Some companies in the chemical industry have already started their journey to reinvent innovation and have achieved new performance frontiers with higher innovation effectiveness. More innovations, more differentiation from competition, more growth and value are the outcomes.
To turn the vision into reality, the first step is defining a company-specific transformation path.
Such a transformation path typically comprises:
- Digitalizing, consolidating and structuring your R&D data to set the foundation for the creation of valuable insights through machine learning and AI.
- Connecting your R&D labs and harmonizing your technology landscape of systems and instruments.
- Building the capabilities your R&D organization requires to achieve its future ambitions and developing a clear pathway through strategically hiring and upskilling.
- Seeding AI and in silico technologies. Implementing prioritized use cases and testing feasibility and impact.
- Deploying at scale. Rapidly scaling up successful use cases across labs, geographies and business units to maximize value and to hardwire digital R&D within your organization.
- Tailoring your processes and R&D operating model to maximize the full value and potential of digitalization. Integrating in silico experimentation and AI-based search and analysis as fundamental steps in your standard R&D process.
Typically, this transformation requires a reallocation of innovation budgets, which is often hard since it means a break with “more of the same.” A shift from traditional budget uses to building digital and AI capabilities is required.
Innovation always includes taking calculated risks. Although transforming innovation may seem risky, the greater risk lies in ending up with non-competitive innovation and missing growth opportunities. Act now to position your company on the trajectory toward transforming innovation and a new level of performance.
Sources
i Estimate based on Accenture project experience and market research.
ii Ibid.
iii Ibid.