skip to Main Content

PIPMG hosted a webinar on the “Potential of AI to Benefit Pharma Projects” by Edward Ho of Mely.ai

What problems in Project Management could AI solve? Can we be inspired to identify opportunities by seeing what other industries are doing?

Dr Janette Thomas presented an Introduction to Data Analytics and AI Potential for Pharmaceutical Project Management

Prof Paul Boudreau gave an introduction on AI in Project Management,

Edward Ko of Mely.ai gave a presentation on Vision, Mission, Tech, Industrial Application Maely.ai Case Studies, are there Pharma Synergies. The four use cases presented by Edward were

  1. Using Computer vision on construction design drawings
  2. Smart document management on supply chain and logistics
  3. Intelligent Knowledge Engine in Project Management
  4. Virtual assistance with sentiment analysis, for all verticals.

 

Following the presentations we had breakout discussions to answer questions

  1. Construction/Computer Vision (CV);    How to apply CV into clinical trials end point measurement? Any other applications?
  2. Supply Chain and Logistics/Data Extraction; Any value from extracting data from heavy documents in clinical documentation, regulatory dossiers?
  3. Project Management/Searchability of Knowledge. In which areas does the knowledge management would benefit the most? Regulatory, clinical, manufacturing? Any value from collecting automatically historical data into a searchable database?
  4. Communication/Virtual Assistance (VA). How would you use VA to manage your projects?

 

Points form the Discussions include;

  • There may be only half a dozen key documents that would provide sufficient information about a project. So data extraction from documents could be doable.
  • May need to have consistent format for documents.
  • For a small company the amount of information is going to be small compared to a large company. So sharing information is going to be important. EG Prof Chas Bountra, a previous PIPMG meeting speaker, highlighted that knowing where not to look is important – all the failed compounds so we don’t repeat past mistakes.
  • Data quality may be an issue but supervised learning may circumvent issues.
  • Predicting the length of projects  and resources required could be useful.
  • Highlight early when projects are going off track. Highlight the impact of risks as they are occurring.
  • Potential to streamline the regulatory dossiers. Documents can be audited to ensure Precision and accuracy. By training the models this can be significantly improved.
  • Are the regulators themselves using AI techniques to examine the documentation?
  • Could we use a VA to summarise discussions and keep track of the learning to share across projects?
  • Most valuable information is within companies, should we be able to share information as anonymised insights for future projects, perhaps through a Data Trust?
  • IBM Watson has compiled a lot of public information from Medline.
Back To Top
X
X
X