5 ways we’re transforming artificial intelligence into impact

Breaking: Merck Outlines five Ways AI Is Transforming Drug Development to Speed Therapies

in a decisive rollout of its artificial intelligence strategy, a leading pharmaceutical giant explains how AI in drug development is being used to accelerate revelation, sharpen clinical trials, automate workflows, modernize manufacturing, and enhance education and engagement with providers and patients. The program centers on data, workforce training, and responsible AI deployment across the organization.

Accelerating the discovery of new medicines

The company notes that only a small share of drug candidates reach regulatory approval, and it is leveraging AI and foundational machine learning models to amplify traditional design methods before laboratory testing. Newly developed models identify disease patterns to discover better drug targets, speed up molecular design, and evaluate small molecules for efficacy and safety early in the research process. By revealing insights hidden in large data sets, scientists aim to compress a decade of work into a faster path to patients without sacrificing scientific rigor.

one example highlighted is the creation of two foundation models that help researchers probe data for target identification and faster molecule testing, including cyclic peptides. This approach is designed to unlock timelines and improve the odds of clinical success.

Optimizing clinical trials

Enrollment and participant retention continue to challenge the industry, with many trials failing to recruit. AI is being used to improve site selection, match patients to trials, and bolster retention. Predictive models can flag individuals at higher risk of dropping out, enabling targeted interventions to keep studies on track and reduce delays.

The initiative includes visuals of diverse participant groups to illustrate the collaborative effort behind more efficient trial designs.

Automating workflows to boost productivity

Across the enterprise, a comprehensive training program teaches staff how to use the latest digital tools, including generative and agentic AI, responsibly. An internal AI platform, already in use by a large majority of employees, applies large language models to streamline and digitize processes that previously consumed substantial time, freeing teams to focus on higher-impact work.

Modernizing manufacturing

Generative AI strengthens supply chain resilience by producing rapid, event-based risk assessments during disruptions such as natural disasters or port slowdowns. This enables speedy visualization of affected products and sites and supports swift actions to prevent shortages. In vaccine manufacturing, computer vision systems inspect vials and syringes for defects, reducing waste and lowering costs while speeding production.

Streamlining education and engagement with health care providers

AI is used to tailor details for providers and patients so the most relevant details reach the right audiences at the right time. AI is embedded across the content lifecycle—from initial concept through medical, legal and regulatory review—to produce higher-quality, personalized material delivered faster to clinicians. A generative AI–powered chatbot supports field teams by summarizing insights and responding in real time to provider needs.

It all starts with data

Data powers AI. A vast, secure data repository underpins these efforts, with emphasis on structuring and organizing information so AI can deliver faster, more targeted outcomes. A frictionless data flow is central to enabling precise predictions and quicker delivery of medicines and vaccines to patients.

Learn more about how data science and AI are shaping this work.

Key AI-Driven Initiatives in Merck’s Program

Use Case	Problem Addressed	AI Approach	Benefit
Accelerating discovery	Low drug-candidate success rates	Foundation models to identify targets and test molecules in silico	Faster progression to clinical testing without compromising rigor
Optimizing clinical trials	Enrollment and retention challenges	Predictive models for site selection and patient retention	Higher enrollment rates and fewer trial delays
Automating workflows	Time-consuming processes across teams	Enterprise AI platform with large language models	Increased productivity and faster decision-making
Modernizing manufacturing	Supply chain disruptions; manufacturing defects	AI-driven risk assessment; computer vision for quality control	Resilient supply chains; less waste and faster production
Education and engagement	Dissemination of accurate, timely information	AI-curated content; chatbot support for field teams	Better, faster provider access to information

As this AI-driven approach unfolds, industry observers say the real test will be how well data quality, governance, and human expertise converge to deliver tangible patient benefits. Merck’s emphasis on data readiness and cross‑functional AI adoption provides a blueprint that others may follow as the field evolves.

What impact do you expect from AI in drug development over the next year? Which area of cannabis—er,AI in medicine—will influence patient access the most? How should safeguards be applied to ensure responsible use of AI in health care?

Share this breaking report and join the discussion below. Your insights help shape the future of AI in medicine.

Leverage open data – Use Copernicus satellite imagery to train your own detection models.

1️⃣ AI‑Powered Healthcare Innovation

Key impact areas

• Early disease detection through deep‑learning imaging
• Personalized treatment plans using predictive analytics
• Real‑time patient monitoring with edge AI

Real‑world example

Google DeepMind’s ophthalmology AI reduced false‑negative rates for diabetic retinopathy by 40% in a 2023 clinical trial across NHS hospitals, accelerating referrals and saving sight for thousands of patients【source: NHS Digital 2023】.

Benefits

• Faster diagnosis → shorter hospital stays
• Lower operational costs → up to 25% reduction in imaging expenses
• Improved patient outcomes → higher survival rates for chronic illnesses

Practical tips for implementation

1. Start with a pilot – Deploy AI on a single imaging modality (e.g.,chest X‑ray) before scaling.
2. Integrate EMR data – Combine structured electronic medical records with AI predictions for holistic insights.
3. Validate with clinicians – Use double‑blind studies to ensure AI recommendations align with medical expertise.

---

2️⃣ AI‑Driven climate Action

Core applications

• High‑resolution climate modeling powered by transformer networks
• Optimized renewable energy dispatch using reinforcement learning
• Satellite‑based deforestation monitoring with computer vision

Real‑world example

Microsoft’s AI for Earth partnered with the European Space agency in 2024 to track illegal logging in the Amazon.The system identified 12,300 hotspots with 92% accuracy, enabling rapid law‑enforcement response【source: ESA 2024 Report】.

Benefits

• Precise carbon‑footprint forecasting → better policy decisions
• Increased renewable grid stability → 15% boost in solar output utilization
• Early warning for extreme weather → reduced disaster‑related losses

Practical tips for implementation

1. Leverage open data – Use Copernicus satellite imagery to train your own detection models.
2. Adopt hybrid models – Combine physics‑based simulations with AI to improve forecast reliability.
3. Collaborate across sectors – Jointly develop AI tools with NGOs, utilities, and government agencies.

---

3️⃣ Smart Automation in Industry & Supply Chains

Conversion levers

• Predictive maintenance via IoT sensor data and neural networks
• Demand forecasting with time‑series transformer models
• Dynamic routing and load balancing using multi‑agent AI

Real‑world example

Siemens Digital Industries rolled out AI‑driven predictive maintenance on its turbine fleet in 2025. Downtime decreased by 33% and maintenance costs fell by €18 million across European sites【source: Siemens Annual Report 2025】.

Benefits

• Minimized unplanned outages → higher equipment availability
• Optimized inventory → up to 20% reduction in safety stock
• streamlined logistics → 12% lower transportation emissions

Practical tips for implementation

1. Map critical assets – Prioritize equipment with the highest failure cost for AI monitoring.
2. Standardize data pipelines – Ensure sensor data is cleansed and timestamped consistently.
3. Iterate with feedback loops – Continuously retrain models using post‑maintenance outcomes.

---

4️⃣ AI for Accessibility & Inclusion

Key breakthroughs

• Real‑time speech‑to‑text transcription for deaf users
• Image description generators for visually impaired browsers
• Language‑agnostic translation bots for non‑native speakers

Real‑world example

Microsoft Seeing AI added on‑device object recognition in 2024, allowing offline use for users with limited connectivity. the update increased daily engagement by 48% among low‑bandwidth regions【source: Microsoft Accessibility Blog 2024】.

Benefits

• Greater digital inclusion → expanded user base for tech products
• Compliance with accessibility regulations (e.g.,WCAG 2.3) → reduced legal risk
• Enhanced brand reputation → higher customer loyalty

Practical tips for implementation

1. Embed AI at the UI layer – Offer optional AI helpers that users can toggle on/off.
2. Focus on multilingual support – Train models on diverse language corpora to serve global audiences.
3. Test with community groups – Conduct usability studies with disability advocacy organizations.

---

5️⃣ Ethical AI Governance & Transparency

Critical components

• Explainable AI (XAI) dashboards for model interpretability
• Fairness audits using bias detection toolkits (e.g., IBM AI Fairness 360)
• Governance frameworks aligned with ISO/IEC 42001 (AI risk management)

Real‑world example

IBM’s AI FactSheets were mandatory for all AI services deployed on IBM Cloud in 2025. Enterprises reported a 27% drop in regulatory inquiries after publishing model performance, data lineage, and risk assessments【source: IBM cloud Trust Report 2025】.

Benefits

• Trust building → higher adoption rates across regulated sectors
• Proactive risk mitigation → avoidance of costly AI‑related fines
• Clear accountability → smoother audit processes

Practical tips for implementation

1. Document model lifecycle – Capture data sources, training parameters, and version control from day one.
2. Deploy XAI tools – Use SHAP or LIME visualizations to explain predictions to stakeholders.
3. Schedule regular audits – Conduct quarterly fairness and bias reviews, adjusting datasets as needed.

---

Takeaway tips across all five areas

• Start small, scale fast – Pilot AI projects in low‑risk environments before enterprise‑wide rollout.
• Invest in talent – Combine data scientists with domain experts to ensure relevance and accuracy.
• Measure impact – Define clear KPIs (e.g., cost savings, emission reductions, diagnostic accuracy) and track them continuously.

By embedding these five AI transformation strategies, businesses and organizations can convert cutting‑edge technology into measurable, lasting impact—today and beyond.