There are multiple challenges of applying advanced technology Artificial Intelligence (AI) methods such as Natural Language Processing (NLP), Machine Learning (ML) and Robotic Process Automation (RPA). Availability of the skilled resources and availability of data (lot is better) are the most rated challenges. In addition, life sciences being a highly regulated and compliance driven industry, challenges related to change management adds to the complexity.
Techsol has been focusing on applying various AI methods to some specific challenges in the Pharmacovigilance and Medical Affairs/ Information processes. We have skills, AI tools/ algorithms and access to data to make automation a reality. Utilizing our unique approach (Risk Assessment based AI adoption) coupled with our skilled resources (AI enabled SMEs and data science engineers), we have successfully delivered automation at various client locations.
Our automation platform leverages various industry standard AI tools, frame works and platforms such as UiPath, IBM, AWS etc. Our team consists of certified consultants, data science engineers and SMEs.
Our AI Capabilities for Enabling Process Automation
With our medical information domain expertise, our team has engineered practical AI solutions for the following key focus areas:
- Medical Inquiry Fulfillment Suggester
- MedComm Retrospective Quality Assurance
- Auto-Redaction of Patient Identifiers
- Medibots for medical information self-service
- QC of IVR Intake and Call Recording
The use of AI in pharmacovigilance is aimed to towards mining valuable information across disparate sources to interpret meaningful patterns for informed decision making. Following are some of the areas that Techsol has been actively involved for developing AI applications:
- Automated Literature screening
- Intelligent Case Intake and Data Entry
- Data-driven Retrospective Quality Assurance
From our experience, our conservative approach is structured around enabling native process automation followed by AI-enhanced capability that leverages omni-channel engagements while keeping human in the loop all the time (human in the loop).
A Risk-based phased AI adoption plan that involve three stages. All the use cases identified for automation are categorized in to three phases based on regulatory compliance risk – Low, Medium, High and Very High.
- Native automation phase: In this phase, we will help client leverage configurations, automation rules and default automation native functionality available with an existing data management system(s).
- Semi automation phase: In this phase, AI technologies such as NLP, RPA, basic ML or combination of these shall be utilized to deliver automation of low to medium risk use cases. There shall be multiple sub phases and iterations within this phase while client is realizing automation in a step by step manner
- Full automation phase: This phase involves usage of various advanced AI methods such as NLP, ML, Deep Learning (DL), RPA etc.
Our typical AI project model involves an engagement involves up to 24 to 36 months in a phased manner beginning with a native automation and by executing multiple small pilots.