HEALTHCARE: WALKING THE AI TALK

HEALTHCARE:WALKINGTHE AI TALKACCENTURE SURVEY SHOWS HEALTHCARE ORGANIZATIONS BELIEVEIN ARTIFICIAL INTELLIGENCE, AND THEY ARE LEADING THE WAY INADOPTING MACHINE LEARNING-ENABLED PROCESSES.

Artificial intelligence (AI)is evolving at speed. Smartmachines are reinventing howwork is done across industries.Companies across sectors andregions are seeing an initialboost in process speed andperformance by implementingAI technologies. However, thereis significant untapped potentialin reimagining businessprocesses from the ground upas self-improving proceduresthat can sense, comprehend,act and learn—all in real time.MACHINELEARNING VS. AIMachine learning is a subset ofartificial intelligence that providessoftware, machines and robotsthe ability to learn without humanintervention or assistance, withoutstatic program instructions.Machine learning has been usedUntil now, many business leaders have taken toonarrow a view of AI. To maximize the potential of AIand be digital leaders, healthcare organizations mustreimagine and reinvent their processes from scratch—and create self-adapting, self-optimizing “livingprocesses” that use machine learning algorithms andreal-time data to continuously improve. Machinesthemselves will become agents of process change,unlocking new roles and new ways for humansand machines to work together.to generate personalized productThis is the era of “process reimagined.” It is alreadydelivering profound results across industries.Surprisingly, healthcare is at the forefront of thisburgeoning trend.Machine learning-enabledrecommendations to consumers,identify the root cause of qualityproblems and fix them, discover andrecommend treatment options tophysicians, and in numerous otherapplications.processes rely on software,systems, robots or other machineswhich use machine learningalgorithms.2

HEALTHCARE IS LEADING THE WAYIN AI AND SEEING THE BENEFITSAmong the 13 industries included inthe Accenture 2017 ProcessReimagined Survey, Healthcare istied with Mining and Minerals forNo. 1 overall (see Figure 1). Retailand Telecom follow close behind.Healthcare ranks No. 1 in Processand Data.About 15 percent of healthcarecompanies (US payers, providersand pharmacy benefit managers) areapplying machine learning acrossthree dimensions: Process, people anddata (see Figure 2). These healthcarerespondents in the process reimaginedphase are moving beyond automationto take full advantage of AI. Theprocess reimagined average across allindustries surveyed was 9 percent.Among healthcare organizationssurveyed, 42 percent are usingmachine learning to reimagineprocesses and process change.Furthermore, 42 percent of UShealthcare firms surveyed say theyare using machine learning in at leastone business process and 42 percentare using it to transform the humanmachine work relationship. More thanhalf (52) percent are harnessing datato create exponential improvementsin speed and KPIs.FIGURE 1. Healthcare ranks at the top among industry use of AI15%HEALTHCAREMINING & 0%PUBLIC SERVICE9%MANUFACTURING8%CPG8%IT/TECH/EHT8%OIL & GAS7%CHEMICALS7%5%METALSSource: Accenture 2017 Process Reimagined SurveyFIGURE 2. Leading organizations are applying machinelearning across three overlapping CEDATA &DATA MODELSSource: Accenture 2017 Process Reimagined Survey3

SEIZING VALUE THROUGH MACHINELEARNING-ENABLED PROCESSESSome AI initiatives are pilots andtherefore deliver benefits on a smallscale. But healthcare organizationsare thinking bigger. Many report thatmachine learning-enabled processesare reducing costs/improving revenueand also improving KPIs.More than half (51 percent) ofhealthcare organizations say thatmachine learning-enabled processeshave helped to reduce costs to “serviceproducts after sales” by at least 50percent. Many healthcare respondents(65 percent) say they have improvedrevenue by 10-20 percent by usingmachine learning-enabled processesto “understand markets, customers andcapabilities.”Almost all healthcare respondents(93 percent) “strongly agree” or“agree” that machine learning-enabledprocesses help achieve previouslyhidden or unobtainable value and86 percent believe these processesare finding solutions to previouslyunsolved business problems (seeFigure 3). A majority (91 percent) areseeing 200 percent improvement inKPIs in enterprise processes, and 77percent are doubling KPIs for salesand marketing in the front office.FIGURE 3. H ealthcare organizations agree that machine learning-enabledprocesses deliver benefitsMachine learning-enabled processeshelp achieve previouslyhidden or unobtainable valueMachine learning-enabled processesare finding solutions to previouslyunsolved business problems45%47%48%39% 93%STRONGLY AGREE86%AGREESource: Accenture 2017 Process Reimagined Survey4

BROADENING PERSPECTIVEON AI OPENS UP NEWPOTENTIAL IN HEALTHCAREHealthcare leaders can use AI to reimaginenew processes rather than simplyautomating existing processes. Whileautomation often brings a short-term jumpin productivity and speed, these benefitswill level off if the focus remains on processautomation rather than reinvention.Digital leaders will reimagine and reinventtheir processes, creating “living processes”that use machine learning algorithms andreal-time data to continuously improve.The leaders in the process reimagined phaseare focusing on three overlapping areas:PROCESSDATAWORKFORCE5

1. P ROCESS CHANGE:REIMAGININGPROCESSES FROMSCRATCHFORTY-TWO PERCENTof health organizations are reimagining processes by: Applying artificial intelligence to process change management Rethinking standardized processes as continuously adaptive Bringing AI-based change to multiple processes across the enterpriseHealth respondents in significant numbers are using or piloting machinelearning-enabled processes across several business process categories(see Figure 4).Multiple use cases show that artificial intelligence is being applied in healthspecific processes such as supporting clinicians in preliminary diagnosis andtreatment, and even generic enterprise processes such as sales, pricing andenterprise operations. Supported processes include: Personalized patient care Clinical connectivity Security and compliance Enterprise operationsAyasdi uses machine learning to analyze electronic medicalrecord data and perform clinical variation management forcommon surgical procedures. The platform automaticallysurfaces groups of similar patient procedures and generatesclinical pathways that could result in better patient outcomesat lower costs.6

FIGURE 4. Healthcare companies are pursuing machine learning across a variety of processes50%48%47%45%45%45%Manage informationtechnologyDeliver physicalproductsDevelop and managehuman capitalManage financialresourcesDevelop and managebusiness capabilitiesDevelop and manageproducts and servicesSource: Accenture 2017 Process Reimagined Survey7

2. DATA AND DATA MODELS:CAPTURING THEEXPONENTIAL POWEROF DATAFIFTY-TWO PERCENTof health organizations are harnessing data to create exponential improvementsin speed and KPIs, help solve previously unsolved problems and see patterns thatwere previously hidden. There are doing so by: Using data to train and sustain process change Making processes self-adapting and self-optimizing Discovering new patterns of opportunityHealthcare respondents reported experiencing a variety of benefits fromusing machine learning-enabled processes, including cost savings acrossseveral processes (see Figure 5).Tampere City Department of Social Services andHealth Care (Finland) built a way to capture datain its facilities to help monitor patients, enhancethe healthcare service provided, and improveonsite safety.8

FIGURE 5. Machine learning-enabled processes reduce costsfor healthcare organizations across several processesService products after sales6%13%6%13%13%Recruit, source and select employees11%5%11%HealthTap, aPerform revenue accounting9%9%9%World Economic9%Forum TechnologyPerform planning and management accounting21%Pioneer, launched13%Dr. A.I., a personalAI-poweredPerform general accounting and reporting10%“physician” that3% 3% 3% 3%translates a person’sManage enterprise quality11%7%4% 4%symptoms into73%personalized, doctorrecommendedManage compliance4%13%4%courses of action. Dr.9%A.I. uses advancedMaintain productive assets5%9%14%deep learning5%algorithms andEvaluate customer service operations and customer satisfaction8%17%17%4%Develop and implement security, privacy and data protection controls4%12%8%8%14%50% - 59% REDUCTION80% - 89% REDUCTION7%knowledge anddata to apply clinicalpatients to the rightlevel of doctor-3% 3%60% - 69% REDUCTION90% REDUCTION OR MORErepository of doctorexpertise and guideDeliver and support information technology services3%HealthTap’s vast70% - 79% REDUCTIONrecommended care.Source: Accenture 2017 Process Reimagined Survey9

3. WORKFORCE:UNLOCKING THE FULLPOTENTIAL FOR HUMAN/MACHINE INTERACTIONBY INVENTING NEW JOBSFORTY-TWO PERCENTof healthcare organizations are redesigning jobs to emphasize distinctive humanskills and human-machine augmentation. They are transforming the humanmachine relationship by focusing on four areas: Enlisting the C-Suite to remake the culture with AI Helping employees keep pace Emphasizing distinctively human capabilities when hiring Making sure that algorithmic decisions are ethical, fair, safe and auditableA majority (82 percent) is investing in machine learning by hiring experiencedmachine learning talent from other companies to develop capabilities. Many areinvesting because they believe that machine learning-enabled processes willpositively influence the future of work, for instance, allowing employees to domore and better work and enabling greater personalization (see Figure 6).Figure 6. Healthcare organizations believe machine learning will affect the future of workAugmenting employees though machine learning will notonly mean doing more and better work, but also managingfar more process variability and unpredictability.Just as products and services are becoming personalized tomeet individual preferences, in the future, MLEPs willlikewise personalize work.39%44%48%47%87%91%STRONGLY AGREEAGREESource: Accenture 2017 Process Reimagined Survey10

LEADING IN THE NEWERA OF AIHealthcare organizations have anopportunity to employ AI to make asystematic shift that affects every process,piece of data and worker. It is a newperspective encompassing strategy,technology and the future of work—onethat differs dramatically from the notionof creating value through the substitutionof human workers.This new perspective requires innovationat speed and scale. And it requiresunprecedented leadership vision withintentional shifts in process change, dataand workforce. Healthcare leaders mustfoster organizational cultures of creativity,collaboration and data competencyto capture the exponential power thatis possible through AI. Perhaps mostimportantly, they must strengthen alliancesbetween humans and machines to enablea future filled with growth.11