Top 7 AI Lessons Learned
By Katie Anderson, Chief Editor, Pharmaceutical Online
AI can undoubtedly bring a efficiency to pharma manufacturing, but each company’s journey with AI integration is unique. After two days of presentations on technology integration at ISPE’s 2025 Pharma 4.0 conference, a closing panel gathered to discuss best practices and lessons learned with AI adoption. As an icebreaker to start the panel, Teresa Minero from LifeBee asked each presenter to share their top recommendations and major issues to avoid with AI. Each panelist’s insightful answers are shared here.
1. Don’t Be Afraid of Change.
Paul Hanson, Ph.D., Head of Lifecycle Management, Innovation, and Strategy, Takeda Pharmaceuticals
Top Recommendation: Things are changing so fast. Don’t be afraid to kill your darlings.
Issue to Avoid: Avoid having a very narrow mindset. Think beyond the moment.
2. Take Baby Steps.
Sachin Sontakke, Head of Data & Analytics, Gilead Sciences
Top Recommendation: Take baby steps. Think big but take baby steps.
Issue to Avoid: Don’t start many use cases at the same time.
3. Challenge The Status Quo.
Pooja Arora, HCLS Advisory & Strategist, ThoughtWorks
Top Recommendation: Meet the users where they are. Also, challenge the status quo.
Issue to Avoid: How good a POC is; it is equally a trap.
4. Start With One Use Case.
Vanessa Fernandes, MD, Quality Project Lead, Sanofi
Top Recommendation: Don’t start many use cases at the same time.
Involve pilots at early stage during your project to improve. Build in parallel a community of champions; they can help us in the engagement phase.
Issue to Avoid: Don’t wait for perfection. Focus on business requirements. Don’t develop something without a real goal.
5. Focus On The User.
Marc Jordá, Principal Data Scientist, Aily Labs
Top Recommendation: User needs are the most important thing. You can provide something entirely new that they didn’t have before.
Issue to Avoid: Be mission first or impact first rather than AI first.
6. Use Electronic Batch Records.
Ross Fitzgerald, Production Coordinator, Recordati Ireland
Top Recommendation: In an ideal world, your organization would already have electronic batch records in place. It’s important to get operators involved, validating insights early and often.
Issue to Avoid: Don’t underestimate the manual work that you will need to do if you don’t have electronic batch records.
7. Focus On Your ROI.
Christian Gay, GxP Compliance & Validation Expert, Aizon
Top Recommendation: Find a small project you can solve quickly that has some return on the investment.
Issue to Avoid: Make mistakes and improve otherwise.
Adjust To Your Experience
Though the conference provided a wealth of knowledge and tips when it comes to incorporating new technology, these top recommendations and issues to avoid sum up the approach to technology in pharma manufacturing perfectly. Keep them in mind as you go on your own journey with AI, and adjust based your own experience.