Overview
The client, a leading US-based pharmaceutical innovator, sought to accelerate early-stage drug discovery by unifying AI-driven molecule prediction, automated synthesis workflows, and real-time data visibility into a single cloud-native platform. Their existing processes were fragmented across multiple disconnected systems, creating inefficiencies and delaying scientific decision-making. By leveraging modern cloud enablement for life sciences, the objective was to build a modular, scalable solution that could integrate their AI prediction engine, streamline synthesis documentation, automate orchestration with the laboratory, and provide real-time inventory insights—ultimately boosting small molecule synthesis throughput by at least 50%.
Our client
Our client is a US-based pharmaceutical R&D company recognized for its strong focus on technological innovation and automation in small molecule drug discovery. With a fully automated laboratory and an in-house AI-based molecular prediction engine, the organization is committed to modernizing discovery workflows and accelerating experimentation through scientific informatics platforms. Their emphasis on integrating advanced informatics and automation makes them a frontrunner in next-generation pharmaceutical research.
Client’s challenge
Despite their advanced capabilities, the client’s discovery workflow was hindered by fragmented systems and siloed processes. Molecule prediction, retrosynthesis planning, inventory checks, and automated synthesis orchestration operated independently, resulting in delays and manual validation at multiple stages. Scientists lacked seamless access to retrosynthesis pathways, real-time inventory availability, and automated coordination with the laboratory—challenges commonly addressed through lab informatics and ELN/LIMS integration. These inefficiencies slowed decision-making and prevented the organization from fully leveraging their AI and automation investments.
Client’s goals
The client’s goal was to develop a modular, scalable, cloud-based solution that would:
- Integrate with their in-house AI-based molecule prediction engine
- Standardize and document reaction synthesis workflows using robust scientific data management systems
- Automate synthesis orchestration with the lab
- Provide real-time inventory visibility and feasibility checks
- Accelerate small molecule synthesis throughput by at least 50%
Our approach
Excelra collaborated closely with the client through a multi-phased, agile delivery model to ensure smooth adoption and minimal disruption:
Discovery & wireframing
- Conducted workshops with scientists, engineers, and IT teams to gather requirements, review existing systems, and define workflows.
- Created BRDs, technical architecture, wireframes, and ERDs, along with a multiphase rollout plan prioritizing business-critical features.
Design, development & testing
- Defined epics, user stories, and acceptance criteria; designed intuitive UI/UX and developed backend APIs and database schemas for ELN, SDMS, pathway prediction, and inventory modules.
- Integrated Ketcher for molecule visualization, configured pathway prediction and synthesis validation APIs, and built project, library, and user management
- features. Conducted comprehensive unit, data flow, and end-to-end testing to ensure reliability and performance.
AWS cloud implementation
- Set up AWS cloud infrastructure (S3, databases, storage) with secure, SSO-ready authentication and scalable multi-tenant architecture.
Workflow
Excelra collaborated closely with the client through a multi-phased, agile delivery model to ensure smooth adoption and minimal disruption:
Delivered systems and workflow for enhanced research efficiency
To streamline the end-to-end process of molecule design, synthesis, and evaluation, we implemented an integrated workflow that leveraged automation, advanced analytics, and secure governance to ensure efficiency, accuracy, and compliance.
Data capture and registration
Researchers registered new molecules through a secure digital platform integrated with an electronic lab notebook. Automated validation against internal compound libraries ensured consistency and traceability, forming a strong foundation for AI-driven drug discovery workflows
Predictive modelling
An AI/ML-driven retrosynthesis engine analysed historical reaction datasets and chemical rules to propose optimal synthesis pathways. These pathways were ranked based on confidence scores, enabling researchers to select the most feasible route. Alternative synthesis options were also provided to support flexibility and risk mitigation.
Automated synthesis execution
Once a pathway was selected, the system interfaced with a robotics-enabled laboratory. Connected to inventory and scheduling systems, the platform ensured reagent availability and orchestrated automated synthesis steps. Real-time monitoring of critical parameters such as temperature and pH was performed, and analytical assays (e.g., spectroscopy and chromatography) were triggered automatically upon completion.
Assay and data analysis
An integrated analytics module captured assay results and linked them to the original synthesis record. Compliance-ready reports were generated for quality assurance, while interactive dashboards visualised pharmacokinetic and physicochemical profiles. This enabled rapid decision-making and supported downstream development activities.
Feedback and optimisation
The workflow incorporated an AI-driven optimisation engine that compared predicted versus actual outcomes. Based on this analysis, the system suggested parameter adjustments for future runs and continuously refined predictive models through iterative learning, ensuring ongoing improvement in accuracy and efficiency.
Governance and security
Role-based access controls and audit frameworks safeguarded sensitive data throughout the process. Every action—from molecule registration to assay reporting—was logged to maintain transparency and compliance with industry standards for data integrity and regulatory requirements.
Figure: Automation workflow
Our solution
Excelra delivered a cloud-native, modular drug discovery platform with the following key capabilities:
- Unified data management: Centralized molecule registration, combinatorial library ingestion, pathway validation, and results tracking.
- End-to-End Automation: CSV/RDF export for validated pathways enabling fully automated synthesis execution in the lab.
- Real-Time inventory visibility: Automated integration with internal LIMS, supplier catalogues, and automation lab inventory to flag unavailable materials early.
- Scalable APIs: Enabled future integrations and easy scale-up through a robust API-first design.
- Multitenant architecture: Allowed the client to manage multiple customers and synthesis orders from a single platform.
You can explore a similar success story in our AI-powered Drug Discovery Case Study, where scalable data pipelines accelerated molecule prediction workflows.
Figure: Custom roundtrip workflow
Key results
- 50% Faster synthesis: Reduced cycle times by half, accelerating go/no-go decisions in drug discovery.
- Improved data integrity & FAIR compliance: Standardized metadata capture ensured reusability across projects and teams.
- Reduced rework: Inventory feasibility checks minimized failed synthesis attempts and saved material costs.
- Streamlined collaboration: Scientists, automation lab, and inventory teams worked from a single source of truth.
Conclusion
This engagement was a success story of combining AI, automation, and cloud-native architecture to transform small molecule drug discovery. Excelra’s expertise in workflow design, API-based integrations, and cloud deployment positioned us uniquely to deliver a solution that not only met but exceeded the client’s expectations.
