Throughout its 125-year history, the client has grown into one of the world’s largest biotech companies, as well as a leading provider of in-vitro diagnostics and a global supplier of transformative innovative solutions across major disease areas.
The Challenge
Drug discovery is a complicated process. It requires high precision to analyze big data sourced from various instruments, smart wearables, or across labs – all of which pile up as operational inefficiencies that can cause delays and cost escalations.
The client ran its lab workflow on legacy systems requiring high manual intervention. This was time-consuming and hindered scale-up. A single experiment took a fortnight and 20 scientists; the client could only run 80 weekly experiments. Rerunning these experiments was a costly proposition.
Adding to this complexity was a shortage of medical writers and a reliance on manual clinical protocol drafting, which further delayed the process. This increased the risk of errors, which could ultimately lead to failure to secure regulatory approval at the final stage.
The client commissioned Persistent to help it automate the lab workflow, enabling its scientists to visualize the experiment automate the generation of clinical protocols and see its real-time status to draw timely insights that could lead to faster turnaround at lowered costs.
The Solution
Persistent’s team of Google experts, 5 onshore and 20 offshore, created a platform that automated the client’s lab workflow, helped its scientists visualize the process, and analyzed data in real-time.
Our team leveraged some of Google Cloud Platform’s (GCP) best-in-class solutions, starting with GCP Foundation to quickly set up an enterprise-ready foundation on Google Cloud, including billing and identity management with Google SSO for authentication. We established a secure protocol to connect source and ingestion to GCP and automated data fetching from sequencer pipelines. We leveraged GCP’s Cloud Storage to store unstructured raw data and plugged-in compute engines for on-demand data processing.
We also worked closely with the client to understand its scientific, regulatory and operational needs to design an AI-powered protocol generator: AiScribe. Built on Google’s AI stack, AiScribe drafts clinical trial protocols guided by Digital Trust principles. The underlying framework objectively evaluates the quality, accuracy and compliance of AI-generated content. It also assigns an objective confidence score to AI-generated content based on syntax, grammar, citations, traceability and explainability, accelerating turnaround times and improving operational efficiency.
Benefits
- 60% faster protocol authoring, accelerating time-to-market for clinical studies.
- 40% improvement in citation quality, enhancing scientific rigor and regulatory compliance.
The Outcome
The real-time processing of data and automated workflow helped the client:
- Slash experimental costs by 60%.
- Scale-up experiments by 125 times, from 80 per week to 10,000.
- Improve turnaround time with reduced manual errors.
