Artificial Intelligence (AI)-augmented research has put tens of drugs into clinical pipelines. And it is only a matter of time before an international regulatory authority approves a drug substantially created using machine learning (ML)-backed technologies. AI is being used to optimize synthetic routes, predict pharmacokinetic properties, identify drug target sites, and generate novel molecular structures. While these step changes are accelerating innovation, a bigger leap is necessary to take full advantage of AI/ML.
The next substantive change in pharmaceutical R&D warrants a convergence of physical experimentation and AI-augmented digitalized workflows in a seamless, mutually reinforcing loop.
The design-make-test-analyze (DMTA) cycle is one of the major workflows that impact R&D productivity. Each step of the DMTA cycle is dependent on the output of the previous step and cycle times, and relies heavily on effective coordination and collaboration. Digitalized, AI-enabled DMTA cycles can transcend fractured workflows to accelerate R&D.
Improving DMTA Cycle Efficiency
Every new chemical or biological product, whether novel or reformulated, goes through iterative DMTA cycles until the desired properties and performance are achieved.
The DMTA cycle can be thought of as individual experiments. The scientist will first think of how to Design the experiment. In synthetic work, this includes deciding the reagents, starting materials, and reaction conditions necessary to yield the desired product(s) with minimal impurities; and suitable test methods to characterize product identity, purity, and yield. This designed experiment is fully executed in the Make step, and analytical techniques are used to characterize the resulting products in the Test step. Finally, in the Analyze step, the results are interpreted to determine if the experiment is the best route/process, or if the product needs further synthetic elaboration.
Manual Data Translation is the Enemy of AI Enablement
In most organizations, valuable data is scattered across instruments and systems. Scientists spend a significant amount of time moving, organizing, analyzing, and reporting data to ensure seamless flow. Scientists interpret analytical results and make informed decisions by assembling information from multiple sources. At best, those decisions are recorded in their electronic notebook. There are many other transitions in pharmaceutical DMTA cycles, where manual steps lead to inefficiencies and increased risk. In the absence of digital continuity between data and decisions, AI enablement is limited to the steps of a workflow.
A process in which feedback from each digitalized, AI-enabled DMTA cycle informs the next—without the limitation of manual transposition or translation of data—must be the goal. This is not about removing scientists from research; it’s about eliminating bias and ensuring data continuity.
A scientist’s value in the AI-native lab is their ability to apply their experience and knowledge to evaluate decisions and provide refinements, as necessary.
An AI-native lab requires organizations to use automation and integrated systems to eliminate manual data translation steps. Workflows and data flow must be fully digitalized. Only then can AI be effectively assimilated into pharma R&D.
Contract research labs can follow suit and adopt best practices from AI-native R&D organizations to remain competitive. Labs may determine the relevant AI-based technologies to incorporate into their offerings to accelerate bench-level innovation and productivity.
Virtuous DMTA in Drug Discovery
Traditional molecular design employs AI models trained on high-quality, contextualized experimental data to generate candidate structures, propose synthetic routes, and suggest priority experiments. These designs flow into automated or semi-automated experiment execution systems, where data is captured at every step and flows continuously to the next steps. Virtuous DMTA in drug discovery is when hypotheses, designs, experiments, and outcomes are all digitally linked.
Continuity differentiates virtuous DMTA from the traditional cycle. Experimental outcomes are not just reported; they are immediately contextualized with reaction conditions, analytical metadata, and prior decisions. This allows AI systems to learn not just from success, but from failure—adjusting predictions, refining models, and informing the next design iteration. Failed experiments are just as important as successful ones for AI model training. Scientists remain central to this loop, using their expertise to validate AI-driven recommendations, introduce domain insight, and determine when exploration should give way to directed experimentation.
Over time, drug discovery organizations will prioritize the speed at which they learn over experiment throughput. In other words, how quickly experiments can be turned into actionable knowledge would matter more than the number of successful experiments it took to break through and/or optimize.
Virtuous DMTA in Pharmaceutical Development
AI-enabled virtuous DMTA in pharma development supports informed decision-making across process optimization, formulation development, stability studies, and more, where process robustness, cost, and regulatory readiness are paramount.
Digitally connected, AI-enabled workflows allow development teams to evaluate optima systematically, instead of relying on spreadsheets, reports, or static datasets. AI models can predict how changes in process parameters may affect quality attributes, while integrated data systems ensure traceability across development phases. This continuity supports knowledge transfer, preserves institutional learning, and reduces risk.
Virtuous DMTA in development is about creating an environment where scientists can interrogate data, test hypotheses, and adapt strategies with confidence. When digitalized workflows and AI operate as a continuous feedback loop, development becomes more resilient, predictable, and efficient.
A Digital, AI-Enabled Future for Pharma
As pharmaceutical organizations are pressured to deliver better therapies faster, this AI-native virtuous approach to DMTA enables them to free knowledge trapped in data silos, accelerating the journey from molecule design to market. As valuable partners, contract labs must identify where they need to adopt technologies and upgrade their tech stack to provide a competitive advantage.
Stay up-to-date with the industry beats and get expert insights. Register your lab to join the Contract Laboratory global network.
If you’re looking to outsource pharma R&D testing services, submit a quick request right away!
Andrew Anderson is senior director, Product Portfolio, at Revvity Signals. Before joining Revvity, through the acquisition of ACD/Labs, he led product and technology management at ACD/Labs as VP of Innovation and Informatics Strategy. He previously worked in technology scouting at PepsiCo and has served in a variety of management, business development, and strategic partnership roles in the scientific software industry. Andrew earned a BSc in chemistry, an MBA from San Diego State University, and started his professional career as an analytical chemist and regulatory CMC specialist at Pfizer.