1、AI in Life ScienceFROM TRADITIONAL TO GENERATIVEAugust 20232IntroductionRecent developments in generative AI ignited a lot of interest towards artificial intelligence,traditional and generative.The new capabilities of responding to language prompts and creating different kinds of content not only ad

2、d new use cases but make the technology accessible to more users than ever,drastically lowering use barriers for non-tech professionals.At the same time,application of AI in Life Science has never been straightforward due to privacy and regulation concerns,data availability and challenges in stakeho

3、lder buy-in.The risks connected to patients well-being,healthcare professional trust and confidence,and impact on core product pipeline are always the top concerns in the industry.That is why in this overview we focus specifically on the application of AI,generative and traditional,in Life Science,i

4、ts potential impact,prospectives and challenges.This report covers 96 high-level AI use cases applicable in Life Science,grouped by topic and place in the value chain.There is often more than one single way to implement a use case,that is why we refrained from dividing them into generative and tradi

5、tional.We also looked at the potential impact and feasibility of the topics,barriers and trends that are present for AI in the industry.While for every organization implementation the impact of AI may look different,we hope that this overview will provide some inspiration and bring more clarity to h

6、ow AI can transform Life Science.AUTHORS SerikovaSenior Digital Strategy ConsultantTeun SchutteStrategy director and Healthcare&Life Sciences Practice leadMark de BlaauwSenior Data ScientistMarius BurgerDirector of EngineeringJorge Martnez BonillaApplication Security Engineer23Impact of the recent A

7、I developments456AI through Life Science value chain Research&development11Feasibility&impact by AI use case group12Evolution of AI use casesTable of contents13AI implementation14Barriers for adoption15How Mobiquity can support your AI journey?16Some of AI use-cases we are happy to define,build and

8、implement for you17DECODE AI workshop18Get in touch7 Manufacturing&logistics8 Sales&marketing9 Supporting functions10 Digital products34Recent releases of generative AI models from Open AI,Meta,DeepMind,and many more triggered a discussion about AI impact on various industries.Recent developments.AI

9、 and specifically generative AI is a hot topic.How can it be of use for Life Science?and their potential impact.Expected impact on economy from Generative AI uses cases is 2.6-4.4 trillion US dollars,estimated by McKinsey,mostly focused on topics of sales&marketing,IT operations and R&D.Generative A

10、I models demonstrated possibilities of models to generate content based on prompts,increasing the application space for AI and making it more accessible to a wider audience.New capabilities include content generation,conversation and coordination capacity,increasing individual productivity without t

11、he need for special knowledge.Releases sparked the interest towards established uses of AI,e.g.,clustering,recommendations,forecasting.Generative AI provides a value add towards existing models,simplifying customization and limiting the need for translation between people and model.raise a number of

12、 questions What has generative AI to offer in Life Science?What AI use cases have the most potential in Life Science?What are the main challenges in AI implementation?How Mobiquity can help?45AI opens opportunities across the whole value chain and adds to digital product offering.RESEARCH&DEVELOPMEN

13、T MANUFACTURING&LOGISTICS SALES&MARKETING SUPPORTING FUNCTIONS RESEARCH&DISCOVERY ACCELERATION CLINICAL TRIALS OPTIMIZATION PUBLICATIONS STREAMLINING REAL WORLD DATA ANALYSIS SUPPLIERS AND CONTRACTING LOGISTICS OPTIMIZATION QUALITY CONTROL MANUFACTURING OPTIMIZATION DEMAND FORECASTING MARKETING OPTI

14、MIZATION AND CONTENT CREATION SALES ENABLEMENT PATIENT ENGAGEMENT,SERVICES AND SUPPORT HR FINANCE IT LEGALDIGITAL PRODUCTSCORE OPERATIONS SUPPORT OPERATIONS OPTIMIZATION ADMINISTRATIVE SUPPORT SOLUTION MAINTENANCE AND SUPPORTcore business functions Solutions forsupporting functions Solutions forcust

15、omersSolutions forWe have identified 96 high-level use cases for AI in Life Science and grouped them in 20 topics across the value chain.AI can support core business functions,from product development to sales and patient support.It can also boost supporting functions leaving the company more space

16、to focus on what matters.Lastly,companies can build AI solutions for their customers,helping them to boost growth and optimize value.AI use case groups mapped onto Life Science value chain:6Research acceleration and insight generation are the main goals of AI application in research&development.Rese

17、arch&development RESEARCH&DISCOVERY ACCELERATION Scientific publication and competition analysis.Understanding disease mechanisms and disease modelling.Cohort and indications identifications,also for personalized medication.Preliminary molecule screening,drug discovery and structure prediction.Alter

18、native to testing in living organisms and digital twins.Initial future demand forecasting.Planning and staffing optimization.Virtual collaboratorCLINICAL TRIALS OPTIMIZATION Protocol development and instructions prep,study design and set up.Site performance prediction and selection.Documentation and

19、 training creation.Randomization management and ensuring diversity.Managing study logistics.Patient recruitment and enrollment.Monitoring of trial results.Medical coding,clinical data entry,review and analysis assistance.Reporting.PUBLICATIONS STREAMLINING Publication generation.Documentation genera

20、tion for the approval process,its customization for different authorities and countries.REAL WORLD DATA ANALYSIS Analysing patient and practitioner submitted data to identify and manage potential effects,including pharmacovigilance.Analysing open data such as social media to identify and manage pote

21、ntial effects,including pharmacovigilance.In research and development AI can significantly contribute to research acceleration.While discovery and clinical trial orchestration may bring the most impact,the barriers of data sensitivity,need for customized models and high complexity can limit feasibil

22、ity of the use-cases.Still,models relying on open data will have higher feasibility and can constitute pilot use-cases.At the same time,generative AI models can often be applied without heavy modifications to scientific research analysis,protocol,documentation,reporting,and publication generation,as

23、 well as to real world data analysis.7Operation optimization is often the goal of AI applications in manufacturing and logistics.Manufacturing&logistics LOGISTICS OPTIMIZATION Inventory management.Fleet and root optimisation for supply,manufacturing and distribution.Personalized medication productio

24、n and distribution.SUPPLIERS AND CONTRACTING Procurement analytics,risk assessment,vendor analysis,and selection.Contract analysis and contract building.QUALITY CONTROL Supplies quality control.Intermediate and final product quality control.Quality report generation.MANUFACTURING OPTIMIZATION Manufa

25、cturing process optimization.Predictive maintenance.Maintenance co-pilot.Documentation generation.DEMAND FORECASTING Initial demand forecasting and respective impact on supply chain.Manufacturing&logistics use cases can be leveraged not only in Life science,but in other industries as well.This makes

26、 the area attractive for external vendors who provide customizable AI enabled solutions.At the same time,cases like maintenance co-pilot,documentation,contract and report generations can be relatively easily realized in-house with generative AI.8Sales&marketing can leverage customization and co-pilo

27、t opportunities provided by AI.Sales&marketing MARKETING OPTIMIZATION AND CONTENT CREATION Competition analysis.Customer profile creation,followed by next best action type of tasks*.Marketing campaign planning and optimization.SEO and social media optimization.Content generation and customization.Pr

28、icing,including segment and time-dependent pricing*.Launch coordination.Social listening.Key opinion leader identification.Compliance checks and materials creation co-pilot.SALES ENABLEMENT Lead identification and scoring.Field-force optimization*.Incentives spend optimization*.Churn prevention*.Sal

29、es assistance,e.g.,scripts and conversation support.Payer and government research.Internal and external reporting.Virtual salesperson.PATIENT ENGAGEMENT,SERVICES AND SUPPORT Personalized health advice for patients.Behavior adjustment support.Patient reporting/generating input for HCP.Patient complia

30、nce predictions and compliance nudging.Patient engagement through wearables.Call-center routing/optimization.Assisted support dialogues.Customer(including HCP)self-service.*Local regulation may limit applications of algorithms in particular aspects of the topic.AI made its way to customer relationsh

31、ip management solutions,with leading players in this area offering churn prevention,next best action and other solutions.Sales&marketing can be a convenient domain to start AI exploration internally,due to quick feedback opportunities.The challenge will be how to keep the pace of these developments

32、with the compliance and legal teams.Social listening,material creation,reporting and research as well as assisted support are among the easier to implement generative AI use cases in this area.9Like other industries,Life Science can benefit from streamlining supporting functions with AI.Supporting f

33、unctions FINANCE Accounting support.Reporting.Conversation assistants(expert support).Analysis and forecasting.Fraud detection.Invoicing and invoice tracking.HR Analytics-driven hiring.Interview and recruitment assessment material generation.Employee retention.Performance management support.Reportin

34、g and internal communications assistance.Training material generation.Job description and other document generation.IT System design.Product design.Maintenance.Self-service and support.Data cleaning and mistakes correction.Code optimisation and writing.Testing.LEGAL Legal document drafting.Document

35、analysis:compliance and legal issues detection.Legal research.AI can significantly contribute to supporting functions effectiveness.There is potential for vendor-provided solutions,but some questions can remain too sensitive to use in solutions that share their data outside the company domain.Legal

36、use cases can be especially interesting for streamlining,as they are connected to the core functions.Generative AI has a lot of potential here as those functions are heavily text-dependent and data is often internal and pre-structured.10Life Science can offer AI solutions to its stakeholders,e.g.,he

37、althcare provider and payers.Digital productsOPERATIONS OPTIMIZATION Key parameter forecasting and optimization.Revenue sources analysis,support and enablement.Performance assessment and benchmarking.CORE OPERATIONS SUPPORT Diagnostics and treatment identification,decision support.Adherence and beha

38、vior support.Virtual HCPs.Solution/product customization.Research optimization.ADMINISTRATIVE SUPPORT Administrative support,including medical coding.Content creation.Reporting.SOLUTION MAINTENANCE AND SUPPORT Software maintenance.User support.As companies master AI applications,it can also offer AI

39、 solutions to its customers.Those solutions can offer support in core functions,e.g.,for healthcare providers.They can help to improve key parameters like mortality,days till dispatch or workload.Solutions can also help with administrative and support functions,helping healthcare providers to focus

40、on their core competences.In addition,maintenance and support of provided solutions can also be done with the help of AI.11R&D,sales and marketing optimization use cases will drive the impact of AI in the next five years.Impact of AI use cases will vary depending on the operations,maturity and value

41、 chain of the implementing company,as well as on market and client characteristics.Furthermore,the impact of individual use cases might shift with time.For example,while the impact of patient engagement use cases might be limited currently,development of personalized medicine will increase their imp

42、ortance in the next decade.IMPACT:indication of average potential cost savings or revenue growth effects from implementing the use-case group.Impact of individual use-cases varies.FEASIBILITY:indication of data availability,model and regulatory complexity connected with the use case group.Individual

43、 use cases can have higher or lower relative feasibility.FEASIBILITYIMPACTcore internal functionsinternal support functionsdigital productsHRdemand forecastpublication streamliningsales enablementFinanceadministrative supportsolution maintenance and supportLegalMarketing optimization and content cre

44、ation patient engagementITquality controlOperations optimization Logistics optimization Manufacturing optimization Suppliers and contracting Clinical trials optimization Research&discovery acceleration Impact and feasibility distribution of AI use case groups for Life Science companies in the next f

45、ive years:12Scale,complexity and autonomy of AI applications will further increase.The way they are addressed in companies will change in terms of:SCALEFrom specialized pilot applications to holistically enabling operations within selected journeys and offering new AI-supported capabilities to custo

46、mers.COMPLEXITYFrom isolated cases to solutions,addressing multiple business needs and supporting company through the value chain.AUTONOMYFrom co-pilot format supporting professionals to autonomous operations with limited yet present oversight.Regulatory and reputational concerns will limit autonomy

47、 growth in Life Science compared to other industries.PLURALISMFrom limited choice or need for in-house development to variety of vendors offering easily customizable easy-to-implement alternatives.EMERGING BEHAVIORSIntegration of AI into our daily work will raise new questions,cause new employee and

48、 customer behavior and trigger new use-cases.E.g.,one of the use cases coming up is how to tell a model-generated content from person-generated content.Individual use cases are going to evolve through time:Prompts for sales managers with correct and effective language.Generated letter customized to

49、sales person&recipient style and needs.Majorly automated communications with full context of interactions.Expansion to other means of communications.“Virtual salesperson”solutions.Virtual enterprise solutions.Solutions for other domains.SALES LETTER EXAMPLE13Models might be already available but org

50、anizations still need to provide right data and establish quality controls.Identifying the problem company wants to address,defining objectives,key stakeholders,initial architecture and requirements.Setting up data architecture,collecting,cleaning and pre-processing data,addressing possible bias.Ana

51、lyzing data for future and current decision making.Building,adjusting and/or training the selected model.Checking performance and optimizing it.Can focus on prompt engineering in case of generative model use.Testing the model and implementing output controls.This step is crucial for Life Science to

52、ensure proper quality control on model output,given risks.Deploying the model,maintaining and refining the model.Also includes any activities to change processes connected with the solution.Key stages for AI use case implementation.INITIATION DATA&PLATFORM PREPARATION MODEL TRAINING AND/OR TUNINGMOD

53、EL TESTING&CONTROLS IMPLEMEN-TATION DEPLOYMENT12345123451234514Barriers for AI adoption in Life Science are significant but can be mitigated.Life Science often works with highly-sensitive,specialized,fragmented and inconsistent data.Ensuring data sufficiency and quality,especially for projects relyi

54、ng on external data,is a challenge.Main barriers DATA AVAILABILITY Life Science is a heavily regulated industry,especially when it comes to operations facing patients and HCPs.Approval processes,privacy questions along with necessary studies can significantly slow down adoption.Local regulation diff

55、erences will lead to different adoption pace of adoption,with USA probably being in the general forefront.REGULATION&PRIVACY Introducing new practices is never easy,especially in such a complex environment as Life Science.With reputation being a crucial asset stakeholders will be demanding quality a

56、nd clarity from solutions.STAKEHOLDER BUY-IN Setting up AI requires skills that are in high demand at the moment.Life Science specificity may make the search for the right talent even trickier.NEED FOR TALENT It is important to start experimenting with AI now to understand how it can serve the busin

57、ess.Implementing smaller cases in selected 2-3 domains already helps to build better understanding of AI value and challenges.Mitigation strategiesPiloting operation optimization use-cases.Leveraging internal data or readily available models is often the most feasible for companies that start with A

58、I.This also helps to limit regulatory and reputation exposure in the beginning.Data strategy and practices are foundational for AI.Investment in them ensures quality and reasonable timelines for AI solutions in development.Current landscape already warrants make,partner or buy analysis before engagi

59、ng in a project,offering options to manage risks and source external talent.Leveraging AI to its full potential requires changes throughout the organization,not only in IT or digital.Legal,operations,sales and HR readiness can significantly contribute to the success or failure of AI programs.15Mobiq

60、uity supports organizations in AI journeys from outlining AI strategy to bringing new ideas to meaningful and operational solutions.We partner with our clients to create or reimagine AI strategies for organizations,domains and customer journeys.We identify the best place to start,ensure that AI serv

61、es the goals of organization and build the momentum to go beyond pilots to consistent research and implementation.AI STRATEGY We define a fitting approach,methods and tooling that serve your business goals,current and desired maturity level for your AI strategy or a particular use-case.We bring huma

62、n-centred approach to AI,making sure the solution addresses actual human needs,is fit for purpose and is financially beneficial.AI TOOLS,PROCESSES&TALENTS We build solutions based on a solid foundation of best practices,blueprints with repeatable processes,reusable assets,proven methodologies and pa

63、rtnerships.We provide the full set of capabilities required for AI from data engineering to change management and ensure proper maintenance or handover depending on your needs.AI SPEED&SCALABILITY 1516Some of AI use-cases we are happy to define,build and implement for you.RESEARCH&DEVELOPMENT MANUFA

64、CTURING&LOGISTICS SALES&MARKETING core business functions Solutions for Scientific publication and competition analysis.Planning and staffing optimization.Documentation and training creation,including regulatory approval,publication generation.Analysing patient and practitioner submitted data,open d

65、ata to identify and manage potential effects.Contract analysis and contract building.Manufacturing process optimization.Predictive maintenance.Maintenance co-pilot.Documentation generation.Content generation.Social listening,payer and government research.Sales assistance,e.g.,scripts and conversatio

66、n support.Internal,external reporting.Personalized health advice,behavior adjustment support.Call-centre routing/optimization and assisted support dialogues.HRsupporting functions Solutions for Job description,interview assessment and training materials generation.Performance management.Reporting.FI

67、NANCE Accounting support.Reporting,analysis and forecasting.Fraud detection.Invoicing and tracking.IT System and product design.Maintenance.Self-service and support.Data cleaning/correction.Code optimization,writing and testing.LEGAL Legal documents drafting.Document analysis:compliance and legal is

68、sues detection.Legal research.DIGITAL PRODUCTScustomersSolutions for Adherence and behavior support.Solution/product customization.Key parameter forecasting and optimization.Revenue sources analysis,and enablement.Performance assessment.Administrative support,including medical coding.Content creatio

69、n.Reporting.Software maintenance.User support.17LOOKING FOR A PLACE TO START?We look forward to helping you gain practical insights and strategies to decode the complexities,and harness the true potential of generative AI,driving real value for your business.Generative AI Ignition Workshop 1.RESEARC

70、H CHALLENGES Preceding the workshop our team will conduct market research and a couple of short interviews with your business stakeholders to already get a good understanding of its challenges and the target audience 2.EMPATHIZE&DEFINE Introduction to Gen AI and showcasing technological possibilitie

71、s.Inspiring examples of valuable use cases and scenarios.Break down challenge,create journeys and define needs,frictions and hypotheses.3.IDEATE&PRIORITIZE Ideate solutions based on our Gen AI value proposition canvas.Prioritize based on human desirability,technical feasibility and business viabilit

72、y.4.REPORT OUTAfter the workshop,our team creates a report out with:Portfolio of opportunities.Riskiest assumptions for validation.Execution plan with team,timeline and investment.Join the workshop17Get in touchLegal Disclaimer The material in this document has been prepared with the aim of providin

73、g information and is for illustrative purposes only and is not meant to be legally binding.Mobiquity accepts no liability whatsoever in contract,tort or otherwise for any loss or damage caused by or arising directly or indirectly in connection with any use or reliance on the contents of this documen

74、t.Rights and Permissions The material in this work is copyrighted.With the exception of fair use for journalistic or scientific purposes,no part of this report may be reprinted or reproduced in any form or by any means without the prior written permission of Mobiquity.In all journalistic or scientific purposes Mobiquity must be indicated as reference.Mobiquity encourages dissemination of its work and will normally grant permission promptly.Teun SchutteStrategy director and Healthcare&Life Sciences Practice lead+31 6 8275 7033 StolHealthcare Lead EMEA+31 6 83 33 9349