The next level in chemicals R&D digitalization: three paths on the road to generative chemistry

Today, chemical companies are feeling pressure from many quarters. They face an economic slowdown, increased costs for materials and energy and intense global competition. At the same time, they need to meet ambitious lower-carbon and net-zero goals to become more sustainable in terms of both operations and products—all of which will require significant investment.

The challenges the industry faces vary, but they share a common solution: innovation. In an analysis conducted by Accenture, we found that in the next two to three years, the industry’s planned investments in excellence programs will not be enough to grow enterprise value. Instead, these efforts will simply “balance out” rising costs, allowing companies to do little more than tread water. As a result, the opportunity to achieve growth will lie in innovation and the creation of new offerings. Not surprisingly, we found that across industry investor communications, innovation is cited as a vital contributor to future competitiveness and growth.

All of this is putting the focus on R&D—the heart of innovation in the industry. The growing need to do business in a more sustainable way puts the R&D function under pressure to develop more sustainable products and processes. In times of increasing economic pressure and rising cost, the R&D function will need to find ways to work faster and more efficiently to effectively develop the innovations needed to keep up with the growing demand for sustainability. And this transformation of R&D will rely heavily on technology.

Digital technology is already disrupting R&D in the industry, thanks to everything from quantum and high-performance computing (HPC) to automation, the extensive integration of labs, systems and data and increasingly powerful analytics. In particular, artificial intelligence (AI) and even Generative AI are playing a widespread role in this trend and are increasingly woven into a range of R&D processes, from the initial search for new ideas and the planning of experiments to the analysis of results, filing of patents and full-scale plant design.

Technology-driven change will continue to be the norm in R&D in the coming years. Based on our analysis, we see the evolution of R&D in the industry moving forward along three technology-enabled pathways: experiment automation, in-silico experimentation and intelligent analytics (See Figure 1).

Figure 1. Three pathways of chemical R&D’s further evolution

To take advantage of this evolution, chemical companies will need to understand and exploit changing technology and leverage it to transform R&D. This will mean making fundamental changes in R&D processes, organizations, skills, capabilities and culture—in a very real sense, re-inventing R&D. By successfully doing so, companies will ensure that they can innovate with the speed, precision and cost-effectiveness needed to grow.

Experiment automation

In the R&D lab of the future, connected, automated systems will be used to plan, execute and analyze experiments faster, more accurately and with increased safety, due to less human interaction with potentially dangerous processes and materials. The chemical industry is exploring a range of possibilities on this front. For example, cloud-based platforms are being used to combine data and search engines for automated research, experiment design and experiment analysis. Cloud-based access and control is making it possible to remotely manage fully automated laboratory facilities in order to plan, execute and analyze physical experiments.

Increasingly, these various automation systems will be brought together on comprehensive lab-execution platforms. These can provide customizable lab functions and AI/machine learning (ML) capabilities, and seamlessly connect instruments, data and scientists in real time to enable the end-to-end orchestration of automated activities across labs and partner ecosystems. Such platforms can support a wide range of technologies and tools, which means companies don’t need to rely on a single system or vendor, giving them the flexibility to easily change systems and bring on new technology as needed.

The automation of physical experiments has the potential to bring a range of benefits. Our analysis indicates that it could help reduce the lab workforce by up to 30%, while virtually eliminating human exposure to hazardous substances through the use of robots. It also has the potential to increase experiment throughput fivefold, while providing high levels of accuracy and reproducibility.

In-silico experimentation

HPC and quantum computing are revolutionizing chemical-industry R&D. Cloud-based HPC is making it possible to perform chemical calculations 50% faster than traditional computers, and this technology continues to improve. At the same time, emerging quantum computing systems are expected to eventually offer computational speeds that are 150 million times higher than those provided by HPC. Both technologies are making it possible to run simulations of very complex reactions. These in-silico experiments can be performed much more quickly than physical experiments, making it possible to rapidly test a wide range of possible chemical formulations. In the chemical industry, companies are looking for ways to use these capabilities for molecular modeling and simulation to investigate molecular structures, properties and complex reactions; and to achieve specific properties of materials or substances, such as strength, conductivity or catalytic activity.

In-silico experimentation not only accelerates the work of R&D, it can also cut the cost of developing a chemical product by 30% to 50%, according to our research. It also opens the door to finding effective solutions to important challenges. For example, Accenture has worked with a consortium of partners to explore the use of ML and massively parallel computing to simulate the breaking of chemical bonds of perfluoroalkyl and polyfluoroalkyl Substances (PFAS), which have lately gained media reputation as “forever chemicals,” achieving the highest known accuracy for this to date.

Intelligent analytics

Backed by growing computational power and Generative AI, data analytics is becoming a game changer for chemicals R&D. This is especially evident on the front end of the R&D process, where analytics can be put to work in content searches that uncover potential innovations and in patent searches that help R&D understand the competitive landscape.

For example, the emerging “deep meaning” search of content leverages AI to perform complex searches not only for text, but for concepts and relevant information. This method 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 prescriptive recommendations suggesting which experiments to run to create chemicals with the desired properties.

When combined on a platform, these technologies can be used to bring together, prepare and store vast amounts of chemical-development data from a broad variety of internal and external sources for analysis. AI can then be used to fill in gaps in that data, creating a more complete picture of what will be needed to conduct planned experiments. For example, AI-based analytics could be used to identify new ingredients for improving composite materials; for recommending polymer recipes that will provide specific attributes requested by a customer; or to develop formulations that remove toxic ingredients from adhesives.

Overall, our research indicates that AI-enabled analytics have the potential to increase success rates with experiments by as much as 70%, thanks to better experiment planning and outcome predictions. By enabling a large scale “fail fast” approach, these tools could reduce the cost of innovation projects through the proof-of-concept phase by 50%, while also allowing a 100% larger project portfolio in stage 1 of the innovation process. And they may enable R&D to reduce manual effort in the analysis of scientific and research literature by 80%.

Figure 2. Intelligent automation

Heading toward generative chemistry

Each of these three technology-enabled pathways will bring a stream of expanding benefits for R&D. We expect those technologies to converge and eventually enable “generative chemistry,” fundamentally changing the nature of the R&D function of today. In R&D of the future, labs will be largely automated, managed remotely, and linked with other internal labs and external innovation partners. Experiments will typically be planned by computers and run virtually, with only the most promising possibilities then being tested physically in the lab. And AI will be woven throughout R&D operations, bringing new levels of efficiency and speed to processes.

R&D’s embrace of generative chemistry will bring a shift from the traditional trial-and-error approach to technology-enabled, highly focused predictive and prescriptive experimentation. Similarly, it will enable R&D to move from tight planning and control to more far-ranging discovery and emergence of innovations, with the ability to rapidly, accurately and cost-effectively explore a very wide variety of experimental options.

What companies need to do

The evolution of digital R&D will continue in the coming years, but there are six initial steps that chemical companies can take now to position themselves to succeed:

1.???? Get your data ready. Cleanse data and transfer it to the cloud to enable seamless data sharing and analysis.

2.???? Build up your computing power. Invest in cloud-based or hybrid infrastructure to support flexible, scalable in-silico chemistry.

3.???? Staff your digital R&D unit. Hire data scientists and AI, HPC and quantum-computing specialists to strengthen your digital chemistry capabilities.

4.???? Adjust your R&D process. Make AI-based research and planning and in-silico experiments standard parts of your innovation process.

5.???? Connect and automate your labs. Ensure that lab equipment, documentation and analysis data are interconnected across lab operations to support automation and collaboration.

6.???? Nourish your innovation ecosystem. Connect with stakeholders across your value chain to share data and accelerate innovation.

Innovation has always played a key role in the chemical industry, but it is becoming more important than ever. That means R&D should begin now to prepare for this digital future and be ready to take advantage of the technology-powered possibilities that will help it stay ahead in the race to innovate.

I'd like to thank Paul Bohm and Michael Henkel for their valuable collaboration on this article.

My posting reflects my own views and does not necessarily represent the views of my employer, Accenture.