Artificial intelligence is reshaping the pharmaceutical industry, offering powerful solutions to improve marketing, medical affairs, and operations. Success, however, depends on strategic implementation, tracking and acting upon measurable results, and a keen focus on ethical and practical considerations. This article provides actionable insights for pharmaceutical professionals, through real-life examples.

Setting the Vision: A Targeted Approach to AI Implementation

AI initiatives must align with organisational goals and address specific challenges. The following real-life examples highlight how leading life science companies have effectively implemented AI to achieve tangible results:

  • Pfizer’s Charlie Platform(1): Pfizer’s generative AI platform leverages GPT-based language models and machine learning to automate regulatory reviews, content creation, and marketing segmentation. It ensures compliance and efficiency by training the system on pre-approved content categorised by therapeutic areas.
  • Merck’s myGPT Suite(2): Merck’s proprietary tool uses Retrieval-Augmented Generation (RAG), a technique that enhances chatbot responses by retrieving additional context from internal documents. It automates workflows and supports task management, while employee training in prompt engineering maximised adoption and results.
  • Bayer’s AI for Real-World Evidence (RWE)(3): Bayer developed an AI tool that helps researchers quickly find answers in large healthcare databases, like electronic health records (EHR) and insurance claims data. By using RAG, it improves accuracy when translating research questions into database queries. This makes it easier for epidemiologists to study patient populations, disease trends, and treatment patterns without needing deep technical expertise in database searching.
  • Prime’s Study on Systematic Literature Reviews (SLRs)(4): . Prime explored the use of generative AI in conducting SLRs and found good agreement between the human reviewer and the AI
  • Plain Language Summaries (PLS) by Pfizer(1): Using its “MAIA” AI platform, Pfizer created a PLS for a sickle cell study and compared the acceptability of the AI generated PLS vs a PLS written by a medical writer. Patients found the AI-generated summary as easy to understand as the human written version.Pfizer also generated a video version of the PLS using the AI platform Synthesia and 58% of patients preferred the video PLS compared to 29% who preferred a written PLS, demonstrating the potential of AI’  in enhancing the accessibility of clinical data for patients.

Measuring ROI: Demonstrating Value

Pharmaceutical companies must tie AI initiatives to measurable outcomes. Key performance indicators (KPIs) include:

  • Time Saved: Automating regulatory reviews saved Pfizer 30% of the time compared to manual processes.
  • Engagement Growth: Novo Nordisk’s AI-driven email campaigns achieved a 24% higher open rate.
  • Cost Reductions: AI automation can save significant manpower costs, as seen in systematic literature reviews and content creation.

Example KPI: A $200,000 investment in AI tools could save $300,000 annually through a 40% faster content review process, coupled with improved engagement and reduced compliance errors.

Ethical Considerations in AI for Pharma

AI must be implemented responsibly to avoid unintended consequences. Key considerations include:

  1. Bias in healthcare data: Foundation AI models trained on internet content (such as ChatGPT, Gemini, Midjourney, etc.) contain biases that must be recognized and mitigated. Health-specific models can also have incomplete or biased datasets, potentially reinforcing disparities. Solutions include using diverse datasets and fairness-aware algorithms.
  2. Transparency: RAG systems, such as Merck’s myGPT, can provide explainable outputs by referencing source data. However, the creative and slightly unpredictable nature of LLMs means that identical queries may produce varied responses.
  3. Governance: Establishing AI ethics committees ensures accountability and alignment with ethical standards.

Challenges and Mitigation Strategies

While AI promises transformative benefits, challenges remain:

  • High Initial Costs: Start with pilot projects to demonstrate ROI before scaling. Encourage employees to experiment with generative AI to supplement their own roles.
  • Workforce Implications: Reskill employees for high-value roles complementing AI.
  • Over-Reliance on AI: Maintain human oversight to ensure accuracy and avoid errors.

Conclusion: A Vision for Pharma’s AI Future

AI is not just a tool but a transformative force amplifying human expertise. Through strategic implementation, measurable outcomes, and ethical considerations, pharmaceutical companies can unlock AI’s full potential. By embracing innovation thoughtfully, the industry can achieve better outcomes for patients, providers, and businesses alike.

Key Takeaways

  1. Align AI Initiatives with Organisational Goals: Ensure that AI projects address specific challenges and align with the company’s overall objectives. Identify common, repetitive tasks or projects that suck resources and are unfulfilling for employees – these are often good use cases for AI.  This strategic alignment is crucial for achieving tangible results and maximizing the impact of AI.
  2. Measure and Demonstrate ROI: Tie AI initiatives to measurable outcomes such as time saved, engagement growth, and cost reductions. Use key performance indicators (KPIs) to track progress and demonstrate the value of AI investments. This helps in justifying the investment and scaling successful projects.
  3. Address Ethical Considerations: Implement AI responsibly by addressing potential biases in healthcare data, ensuring transparency, and establishing governance structures like AI Ethics Committees. This ensures that AI is used ethically and maintains trust with stakeholders.
  4. Upskill through tailored training – deliver executive inspiration sessions through to hands-on deep skills enablement for solutions oriented colleagues and customer-facing teams.

 

References:

  1. ASH Publications. Use of Generative Artificial Intelligence for Medical Writing in Hematology. Blood. 2024;144(Supplement 1):7668. Accessed January 22, 2025.
  2. Merck Group. Introducing MyGPT Suite at Merck. LinkedIn. Accessed January 22, 2025.
  3. ACL Anthology. Clinical Natural Language Processing 2024. Accessed January 22, 2025.
  4. ISPOR. Generative AI in Health Economics and Outcomes Research. Accessed January 22, 2025.

 

Contributing Authors:

David Logue, Partner, Life Sciences, Baringa Partners

Chris Finch, Vice President, Earthware Limited

Richard Lucas, Innovation Director at Tangent90

James Turnbull, Founder of Camino

Sam Pygall, Regional IT Lead @ MSD

Terry Levi, Founder, Embrace.AI for Pharma

*Generative AI was used as part of the process of creating this article, building on the strengths of human and AI collaboration.