The AI Frontier: reimagining the tools and processes of key functions in the Pharma and Life Sciences industry

Current activities and processes encounter challenges that demand a nuanced approach and strategic thinking. Hot topics include the governance and ethical concerns in the wake of the rapid rise of AI and Machine Learning (ML) in the mainstream. The ethical considerations revolve around the responsible use of these technologies in decision-making and the need to ensure transparency and accountability. Additionally, the healthcare industry faces the critical challenges of ensuring supply chain sustainability, dealing with geopolitical shifts and adapting to the digital transformation of data. The imperative to digitise data – rather than merely streamlining processes – introduces concerns with regard to data accuracy, security and privacy. Furthermore, the healthcare sector is increasingly aligning with the principles of the green and circular economy, with the emphasis on sustainable practices from product development to administration to patients. Balancing innovation with ethical considerations and sustainable practices is an ongoing challenge that demands harmonious integration of regulatory frameworks and industry best practices.

Regulators and legislators are navigating the landscape of AI in the healthcare sector with a strategic and proactive approach, while recognising the potential risks. In the United States, the Biden administration is launching a dedicated federal organisation dedicated to assessing the safety of AI. Operating under the National Institute of Standards and Technology (NIST), this organisation is set to play an important role in developing guidance for regulatory rulemaking and enforcement methods.  Meanwhile, across the Atlantic, the European Union (EU) has taken a significant step towards the regulation of AI with the EU AI Act. This legislation defines AI systems with high risk to the health and safety or fundamental rights of natural persons, while mandating conformity assessments, risk management and human oversight. AI applications with particular relevance for the pharmaceutical industry include those relating to drug development, clinical trials, pharmacovigilance etc. Moreover, the EU AI Act supports the establishment of a European Health Data Space, which will facilitate the access and use of health data for AI training and innovation while ensuring data security and privacy. The European Medicines Agency (EMA) and Heads of Medicines Agencies have launched a joint AI workplan to 2028, outlining a strategy to harness AI for medicines while mitigating the risks. The workplan covers four key areas: guidance, policy and product support; AI tools and technology; collaboration and training; and experimentation.¹

Looking at the EU’s pharmaceutical legislation reform, the EMA should be able to capitalise on AI to improve decision-making in the regulatory process. For example, as outlined in Paragraph 60 of the proposed regulation for union procedures for the authorisation and supervision of medicinal products for human use, the EMA could support its judgement with the analysis of health data including real world data (RWE). The vast volume of data generated by the healthcare industry could undergo processing and utilisation through AI and ML algorithms, marking a strategic step towards regulatory practices that are more informed and driven by data.

AI offers numerous opportunities to streamline activities and processes in Life Sciences, revolutionising key aspects of the industry. 

Let’s explore some practical applications:

AI and early drug discovery programmes 

• Discovery: AI can be used to generate and screen large numbers of potential drug candidates based on various criteria such as chemical structure, biological activity, pharmacokinetics, toxicity and similarity to known drugs or targets. This can help reduce the time and cost of synthesising and testing thousands of compounds and identify novel or optimal molecules that may otherwise be overlooked by human intuition or conventional methods.
• Data processing: AI can also be used to analyse and integrate various types of data such as genomic, proteomic, transcriptomic, metabolomic, phenotypic and clinical data, to identify and validate new drug targets, biomarkers, pathways and mechanisms of action or disease. This can help improve the understanding of the biological and molecular basis of diseases and the effects and interactions of drugs, and enable more precise and personalised medicine to be used.
• Designing delivery systems: AI can also be used to design and optimise drug delivery systems such as nanoparticles, liposomes, micelles or antibodies, which can enhance the stability, solubility, bioavailability, specificity and efficacy of drugs and reduce their side effects and toxicity. This can help overcome some of the challenges and limitations of conventional drug formulations, and allow more targeted and controlled drug delivery to the desired tissues or cells.
• Simulation: AI can also be used to simulate and predict the outcomes and impacts of drug discovery experiments such as in silico, in vitro, in vivo and clinical trials, using various models, algorithms and data sources. This can help improve the accuracy, reliability and reproducibility of the results and reduce the ethical and practical issues of animal and human testing, as well as the risks of failure and attrition. For example, in the light of the Swiss law on animal research and welfare (LRA)  , which aims to protect the dignity and well-being of animals used for scientific or educational purposes, as well as to promote the development of alternative methods, AI can be such an alternative method. 

AI and clinical trials

• Trial design: AI can help design more efficient and adaptive trial protocols by using data from previous trials, real-world evidence and literature to optimise parameters such as sample size, inclusion and exclusion criteria, endpoints and statistical methods. AI can also help identify potential risks and mitigate them through adaptive design strategies such as interim analyses or adaptive randomisation methods.
• Recruitment: AI can help recruit and retain more diverse and representative participants, by using Natural Language Processing (NLP) and computer vision to analyse online and offline sources of information such as social media, electronic health records and medical images, in order to identify and engage eligible and interested candidates. AI can also help monitor and support participants throughout the trial by using chatbots, wearable devices and mobile apps to provide reminders, feedback, education and motivation, and to collect data on adherence, outcomes and adverse events.
• Trial management: AI can help monitor and manage the quality and integrity of the trial data by using machine learning and anomaly detection to identify and correct errors, inconsistencies and outliers in the data, and to flag potential fraud, bias or protocol deviations. AI can also help automate and streamline the data management and reporting processes by using NLP and data visualisation to generate and summarise the data in a standardised and user-friendly format.
• Analysis of trial data: AI can help analyse and interpret the trial results by using advanced statistical and causal inference methods to estimate the treatment effects, confounding factors and subgroup interactions, and to account for missing data, multiplicity and heterogeneity. AI can also help generate and test new hypotheses by using data mining and knowledge discovery to identify novel patterns, associations and insights from the data, and to suggest follow-up experiments or trials.

AI and regulatory affairs

• Regulatory intelligence: AI can help collect, analyse and synthesise large amounts of data from various sources such as regulatory guidelines, databases, publications or social media, in order to provide insights and recommendations for regulatory strategy, planning and compliance. For example, AI can help identify relevant regulatory requirements, best practices or precedents for a specific product, indication or market, monitor changes and updates in the regulatory landscape or predict the likelihood of regulatory approval or rejection based on historical data and trends.
• Regulatory submissions: AI can help automate, streamline and optimise the processes of preparing, reviewing and submitting regulatory documents such as clinical trial applications, marketing authorisation applications or post-marketing reports. For example, AI can help generate, format or validate regulatory documents according to predefined templates, standards or rules; check for consistency, completeness or accuracy of data and information across different documents or sources; or flag potential issues, errors or gaps that need further attention or clarification.
• Regulatory interactions: AI can help facilitate, enhance and document the communication and collaboration between the pharma company and the regulatory authorities, as well as other stakeholders such as patients, healthcare professionals or payers. For example, AI can help schedule, prepare or conduct regulatory meetings, consultations or negotiations; provide natural language processing or translation capabilities for cross-cultural or cross-lingual communication; or capture, store or retrieve relevant information or feedback from the regulatory interactions.
• Regulatory monitoring and reporting: AI can help track, measure and report the progress, performance and outcomes of regulatory affairs activities, as well as the safety, efficacy and quality of the products. For example, AI can help monitor the status, timelines or milestones of the regulatory submissions or approvals, analyse the impact or effectiveness of the regulatory strategy or actions or detect, assess or report adverse events, signals or risks associated with the products.

AI and market access/pricing

• Enhancing market research and intelligence: AI can help collect, analyse and synthesise large amounts of data from various sources such as payer databases, claims data, clinical trials, real-world evidence, competitor information and stakeholder feedback. This can help generate insights into unmet needs, value drivers, pricing strategies, market dynamics and customer preferences and behaviours.
• Optimising pricing and contracting: AI can help design, simulate and evaluate different pricing and contracting scenarios, taking into account factors such as market segmentation, price sensitivity, demand elasticity, budget impact, cost-effectiveness, risk-sharing and outcomes-based agreements. This can help optimise the price and value proposition for different products and markets and facilitate dynamic and adaptive pricing and contracting models.
• Supporting evidence generation and communication: AI can help generate and communicate evidence to support the value proposition and pricing strategy of a product, by leveraging advanced analytics, natural language processing and data visualisation techniques. For example, AI can help create and update value dossiers, health technology assessment submissions, budget impact models and value stories, based on the latest data and stakeholder feedback. AI can also help tailor the evidence and messages to different audiences and channels such as payers, providers, patients and policymakers.
• Enhancing negotiation and relationship management: AI can help in the preparation and conducting of negotiations and relationship management with payers and other stakeholders, by providing data-driven insights, recommendations and feedback. For example, AI can help identify and prioritise key decision-makers and influencers, assess their needs and expectations, anticipate their objections and responses and suggest optimal negotiation strategies and tactics. AI can also help monitor and evaluate the performance and outcomes of the negotiations and contracts and provide alerts and suggestions for improvement.

AI and pharmacovigilance

• Signal detection and evaluation: AI can help automate and improve the process of identifying, analysing and prioritising signals of potential safety issues from various data sources such as spontaneous reports, clinical trials, electronic health records, social media, literature and registries. AI can use techniques such as natural language processing, machine learning and semantic analysis to extract, classify, and interpret relevant information from unstructured and structured data and to generate hypotheses and evidence for causal relationships between drugs and adverse events. AI can also help reduce false positives, noise and bias in signal detection and enhance the quality and efficiency of signal evaluation and validation.
• Risk management and mitigation: AI can help design, implement and monitor risk management plans and risk minimisation measures for drugs with known or potential safety risks. AI can use techniques such as predictive modelling, simulation and optimisation to assess the impact and effectiveness of different risk management strategies and to identify and target high-risk populations, settings and behaviours. AI can also help communicate and disseminate risk information and education to various stakeholders such as prescribers, patients, regulators, and payers, through interactive and personalised channels such as chatbots, voice assistants, and mobile apps.
• Case processing and reporting: AI can help automate and streamline the process of collecting, reviewing, coding, and reporting individual case safety reports (ICSRs) from various sources and formats such as emails, faxes, web forms, and phone calls. AI can use techniques such as natural language processing, optical character recognition and rule-based systems to extract, standardise and validate relevant information from ICSRs and to generate and submit regulatory reports in compliance with different requirements and timelines. AI can also help enhance the quality and consistency of case processing and reporting and reduce the workload and errors of human reviewers.
• Literature screening and review: AI can help automate and improve the process of screening and reviewing scientific literature for relevant information on drug safety and efficacy. AI can use techniques such as natural language processing, machine learning and text mining to search, filter and prioritise literature sources such as journals, databases and websites and to extract, summarise and synthesise relevant information from literature articles such as drug names, adverse events, outcomes and study designs. AI can also help reduce the time and cost of literature screening and review, and increase the coverage and accuracy of literature search and analysis.

AI and medical affairs

• Scientific literature analysis: AI can help medical affairs professionals to scan, review and synthesise large volumes of scientific literature efficiently and comprehensively, identify relevant evidence and generate insights for various purposes such as medical education, publication planning, competitive intelligence or evidence generation. AI can also help to monitor and update the literature landscape, detect emerging trends and flag potential gaps or opportunities for further research.
• Medical information and communication: AI can help medical affairs professionals to provide accurate, consistent and timely responses to medical inquiries from various stakeholders such as healthcare professionals, patients or regulators, using natural language processing and generation techniques. AI can also help to create and deliver personalised and engaging medical content such as slide decks, podcasts or webinars, based on the preferences and needs of the target audience, using data-driven and interactive approaches.
• Stakeholder engagement and collaboration: AI can help medical affairs professionals to identify, segment and prioritise key stakeholders such as opinion leaders, researchers or patient advocates, based on their profiles, interests and behaviours, using data mining and analytics techniques. AI can also help to facilitate and optimise engagement and collaboration with these stakeholders, using digital platforms, chatbots or virtual assistants, which can provide tailored and relevant information, feedback or support and enhance the relationship and trust between the parties.
• Evidence generation and dissemination: AI can help medical affairs professionals to design, conduct and analyse real-world evidence studies, using advanced methods such as machine learning, natural language processing or computer vision, which can leverage various sources and types of data such as electronic health records, claims, registries, social media or images and generate robust and actionable insights. AI can also help to disseminate the evidence to various audiences, using innovative formats such as interactive dashboards, infographics or podcasts, which can enhance the accessibility and impact of the evidence.
• Medical strategy and planning: AI can help medical affairs professionals to develop and execute effective and agile medical strategies and plans using predictive and prescriptive analytics techniques, which can anticipate and respond to the changing needs and expectations of the stakeholders, the market dynamics and the scientific developments. AI can also help to monitor and evaluate the performance and outcomes of the medical activities using key performance indicators, metrics and feedback, and provide recommendations for improvement and optimisation.

AI and marketing

• Enhancing customer segmentation and targeting: AI can help pharma marketers analyse large and complex data sets from multiple sources such as electronic health records, claims data, social media, online behaviour and surveys, in order to identify and segment customers based on their needs, preferences, behaviours and outcomes. AI can also help optimise the timing, frequency, channel and content of marketing messages in order to deliver personalised and relevant communication to each customer segment.
• Improving customer engagement and retention: AI can help pharma marketers create and deliver engaging and interactive content such as chatbots, voice assistants, virtual reality and gamification, which can educate, inform and support customers throughout their journey. AI can also help monitor and measure customer feedback, satisfaction, loyalty and advocacy and provide insights and recommendations to improve customer retention and loyalty.
• Enhancing product comparison and differentiation: AI can help pharma marketers leverage data and insights from various sources such as clinical trials, real-world evidence, market research and competitive intelligence in order to identify unmet needs, gaps and opportunities in the market and to generate and test new ideas and concepts for product innovation and differentiation. 
• Optimising marketing performance and ROI: AI can help pharma marketers automate and streamline various marketing tasks and processes such as campaign planning, execution, management and evaluation by using data-driven algorithms and models that can learn and adapt to changing conditions and scenarios. AI can also help measure and optimise the impact and ROI of marketing activities by using advanced analytics and predictive modelling to assess and attribute the outcomes and value of each marketing channel, tactic and strategy.

AI and manufacturing & quality

• Enhancing process optimisation and control: AI can help companies to monitor, analyse and optimise their manufacturing processes in real time, using data from sensors, cameras or other sources. AI can also help to detect and prevent deviations, errors or failures and to adjust the process parameters accordingly. For example, AI can facilitate predictive maintenance, adaptive process control or automated root cause analysis.
• Improving product quality and compliance: AI can help pharma companies to ensure that their products meet the required quality standards and regulatory requirements by automating or augmenting the testing, inspection and validation activities. AI can also help to reduce the risk of human error, fraud or tampering and to enhance the traceability and transparency of the product lifecycle. For example, AI can enable image recognition, natural language processing or blockchain technologies to be used for quality and compliance purposes.

AI and supply chain

• Demand forecasting: AI can help pharma companies to analyse historical data, market trends, customer behaviour and external factors in order to generate more accurate and timely demand forecasts, which can improve inventory management, reduce waste and stockouts and optimise production and distribution planning.
  
• Quality control: AI can help pharma companies to monitor and inspect the quality of raw materials, intermediates and finished products using techniques such as computer vision, natural language processing and machine learning. AI can also help to detect and prevent defects, deviations and counterfeit products and to ensure compliance with regulatory standards and safety requirements.
• Logistics and transportation: AI can help pharma companies to optimise the routing, scheduling and tracking of shipments using methods such as geospatial analysis, route optimisation and predictive analytics. AI can also help to enhance the visibility and traceability of the supply chain and to respond to disruptions, risks and emergencies in real time.
• Supplier relationship management: AI can help pharma companies to evaluate and select the best suppliers based on criteria such as quality, reliability, cost and sustainability. AI can also help to monitor and manage the performance and compliance of suppliers using tools such as sentiment analysis, contract analysis and risk assessment. AI can also help to facilitate communication and collaboration with suppliers using platforms such as chatbots, voice assistants and smart contracts.
• Customer service and satisfaction: AI can help pharma companies to improve customer experience and loyalty by providing personalised and timely service, support and recommendations, using technologies such as natural language processing, speech recognition and recommender systems. AI can also help to collect and analyse customer feedback, preferences and behaviour and to generate insights and actions in order to enhance customer retention and satisfaction.