Transcription
Top Natural LanguageProcessing Applicationsin BusinessUNLOCKING VALUE FROMUNSTRUCTURED DATA
For years, enterprises have been making good use of theirstructured data (tables, spreadsheets, etc.). However,the larger part of enterprise data, nearly 80 percent, isunstructured and has been much less accessible.From emails, text documents, research and legal reportsto voice recordings, videos, social media posts andmore, unstructured data is a huge body of information.But, compared to structured data, it has been much morechallenging to leverage.Traditional search has done a great job helping usersdiscover and derive insights from some of this data. Butenterprises need to go beyond search to maximize the use ofunstructured data as a resource for enhanced analytics anddecision making.This is where natural language processing (NLP), a field ofartificial intelligence (AI) that’s used to handle the processingand analysis of large volumes of unstructured data, can be areal game changer.While AI describes a broad range of technologies, NLP isone of three AI-driven capabilities that enterprises canreadily harness to create business value and competitiveadvantage:1. Internet of Things (IoT): applying technologies, suchas real-time analytics, machine learning (ML), and smartsensors, to manage and analyze machine-generatedstructured data2. Computer Vision: using digital imaging technologies,ML, and pattern recognition to interpret image and videocontent3. Document Understanding: combining NLP and ML togain insights into human-generated, natural languageunstructured data.As it powers document understanding applications thatunlock value from unstructured data, NLP has become anessential enabler of the AI evolution in today’s enterprises.This white paper discusses the emergence of NLP as akey insight discovery technique, followed by examplesof impactful NLP applications that Accenture has helpedclients implement.NaturalLanguageProcessingMarket SizeEstimateSource: nguage-processing-nlp.aspTop Natural Language Processing Applications in Business1
Emergence of NLP applicationsin the enterpriseAs AI technologies like machine learning (ML), deep learning,and cognitive search have entered the mainstream, we’veseen a rapid evolution in NLP applications. This has beendriven by two key factors. First, having grown accustomed tovirtual assistants like Siri and Alexa, as well as sophisticatedsearch like Google, in their daily lives, users expect the sameexperience in the workplace. Second, today’s NLP no longerrelies on manual, rules-based approaches: integrating MLwith NLP allows for far greater automation, scalability, andaccuracy.Enterprises can leverage NLP applications in two ways– query understanding and content understanding – toimprove both user experience and insight discovery:Understanding the user’snatural language query inputs(text or speech) Provide better, more targeted responses byunderstanding the user’s questions and intentIdentify out-of-scope requests and present intelligentalternativesTop Natural Language Processing Applications in BusinessUnderstanding content(mining insights from vastamounts of unstructured data) Extract business entities from text documents to identifyemployees, customers, products, procedures, etc.Identify and understand the meaning of natural languagecontent – documents, reports, emails, etc. – to providenatural language answersNLP enables improved understanding of user queriesand enterprise content. This ensures that every user isconnected to the most relevant, helpful resources whichwould otherwise remain hidden within vast quantities ofdata.Today’s NLP applications – spanning industries and rangingfrom chatbots and virtual assistants to intelligent businesssolutions – are supported by a robust NLP technologyecosystem.2
NLP technology ecosystemRising demand for NLP-powered applications is fueledby a growing range of technologies, from open sourceframeworks and evolving vendors to cloud-based APIs.Commonly-used NLP technologies include:Evolving vendorsOpen source frameworksOther: Python NLTKPython SpacYPython GenSimUIMAGATEApache OpenNLPIndustry specific platforms Quid (market and competitive intelligence)Twiggle (e-commerce)Cloud API providers Google Natural Language APIAmazon ComprehendMicrosoft LUIS (Language UnderstandingIntelligence Service)Top Natural Language Processing Applications in Business Pool PartySmartLogicCotical.ioSAS NLP offeringIBM WatsonWhile NLP is a relatively new enterprise technology, it’sbeing enhanced every day. The ML algorithms supportingNLP are improving all the time, with industry giants likeGoogle, Microsoft, and Amazon all making strides toimprove accuracy.With this growing ecosystem of NLP solutions, enterpriseshave increasing flexibility to select appropriate approachesand toolsets for their specific use cases. Accentureleverages this robust ecosystem and our own technologyassets to orchestrate the different components of NLPapplications – making them easily maintainable andscalable.We’ve helped many clients implement NLP applications.Some of the most impactful are highlighted in the nextsection.3
EnterpriseNLP applicationsin actionTop Natural Language Processing Applications in Business4
Chatbots and virtual assistantsCHALLENGEChatbots are becoming ubiquitous. All around us, Siri,Alexa, Cortana, and Google Home are incorporating naturallanguage conversations between humans and AI intoeveryday interactions. As consumers get accustomed tothese virtual assistants, they expect the same experience inthe workplace. The emphasis is on how well computers talkto humans, interact with them, understand text and so on.11The challenge for enterprises? Providing their people withsophisticated natural language understanding applicationsthat meets their expectations – and improving informationdiscovery and collaboration.1Sources:1. Accenture Research: Chatbots are Here to Stay2. IDC FutureScape: Worldwide IT Industry 2018 Predictions3. Gartner: Conversational AI to Shake Up Your Technical and Business Worlds4. Gartner: Market Guide for Conversational PlatformsTop Natural Language Processing Applications in Business3245
Chatbots and virtual assistants (contd.)SOLUTIONBENEFITSPowered by NLP, enterprise chatbots and virtual assistantscan enable conversations between humans and computersin everyday business processes. Bringing deeper naturallanguage understanding, they enhance search as well asproviding an entirely new way for employees, customersand partners to interact with enterprise data and work moreproductively. For example, an enterprise chatbot that canhold a dialog with the user to acquire targeted information(dates, times, order details, etc.) and/or integrate withbusiness systems to complete common office tasks quickly(reservations, orders, etc.). Improve business processes and reduce support costsby enabling AI-driven self-service virtual assistants forcustomers and partnersEnhance search and knowledge-seeking experiences forcustomers, partners and employeesImprove brand reputation through more efficientprocesses and more engaging customer experiencesBy combining content processing, ML, NLP, and voicerecognition, enterprises can transform simplesearch-based intranets or support portals into AI-poweredvirtual assistants. Whether it’s a question about technicalsupport on a consumer electronics website, pricing andproduct details on a partner or e-commerce website,investment products on an investor portal, or managementadvice on a hotel franchisee site, AI-powered chatbots andvirtual assistants can play key roles in process automationand user experience improvement.Top Natural Language Processing Applications in Business6
Intelligent document analysisCHALLENGEWith overwhelming volumes of incoming documentsand correspondence every day, large organizations oftenstruggle to properly analyze and derive useful insights fromthis content. Whether it’s legal documents, medical records,policies, or contracts, without a centralized and automatedanalysis approach, it may be difficult for organizationsto effectively understand and make use of the documentcontent to support decision making and operationalefficiencies.Core functionalities of this solution can include: SOLUTIONOptical character recognition (OCR) – converts differenttypes of documents, such as scanned paper documents,PDF files, or images into editable and searchable dataText analysis – analyzes documents to identify specificlanguage or terms and extract linguistic meaningDeterministic classification – uses a pattern-basedclassifier to look for sequences of terms that indicate aspecific sort of documentMachine learning – trains a ML model with exampledatasets to predict document type or extract and classifytext (e.g. learning aircraft component names)Intelligent document analysis uses AI techniques includingNLP, entity extraction, semantic understanding, and ML toanalyze content, extract meaning, and reliably aid processautomation and decision making. These applications canidentify specific items of information in documents – likedate, order number, and policy number – so they can becategorized and analyzed.Top Natural Language Processing Applications in Business7
Intelligent document analysis (contd.)BENEFITS Improve compliance and risk managementDrive internal operational efficienciesEnhance business processesTop Natural Language Processing Applications in Business8
Document search and matchCHALLENGESOLUTIONEnterprises store vast quantities of documents, so it’s criticalto be able to identify and match documents of relatedthemes – not just keywords – quickly and accurately.The ideal solution is an application that can indicate people’sfit for jobs by accurately and automatically comparing jobdescriptions to people’s resumes, CVs or other documents.In the recruiting and staffing industry, for example, recruitersare in a race to find the right candidate before competitorsidentify the same person in their candidate databases. Giventhe large amounts of candidates that most recruiters dealwith, it’s easy to lose track of the best matches.The solution we’ve developed is bidirectional and works likethis: a job description (the whole document) is submitted asa search request and the comparison system automaticallyreturns a short list of the best-qualified candidates from adatabase of CVs. Alternatively, a job seeker (or professionalrecruiter) submits a CV and the system returns a list of themost appropriate, currently-available vacancies.There are many candidate search and match solutionsdelivered as standalone products or embedded intoapplicant tracking systems which can provide convenience,but they’re limited to basic resume parsing and matchingcorresponding metadata to a job posting. As a result, mostof these embedded search systems fail to significantlyimprove fill rates or deliver a better experience to recruiters.Top Natural Language Processing Applications in BusinessIt’s an approach that combines advanced search techniques,analytics, NLP, and ML to provide statistical and linguisticcapabilities for understanding applicant profiles andidentifying the best candidates. Predictive analytics usespreviously successful candidates’ placement histories tofind additional candidates ahead of a new job posting –giving recruiters a head start against the competition.9
Document search and match (contd.)ML leverages feedback from recruiters to improve resultsover time.BENEFITS This solution also adds value to human resources teams inglobal enterprises, enabling them to identify the right peoplefrom hundreds of thousands of geographically-scatteredexperts. For example, a project manager can submit aproject requirement document as a search request or ask aquestion like “Who are our NLP experts in North America?”The system will return a list of relevant experts for the projectby automatically analyzing and identifying matches fromemployee databases.Reduce the time to fill a job position or to identify expertsfor specific skillsets within an organizationScale to millions of jobs postings and resumes (inrecruiting) or multiple employee and project databases(in an enterprise “Find the Expert” application)Increase revenue and reduce costs – candidates’matches can be found within the recruiter’s databaserather than via paid job boardsDelivering results across multiple use casesDocument search and match can be applied to varioustypes of enterprise documents. Examples includeanalyzing legal documents for variations of riskycontractual languages or, in the financial servicesindustry, cross-analyzing loan or mortgage documentswith borrowers’ profiles.Top Natural Language Processing Applications in Business10
Data storage optimization analyticsCHALLENGESOLUTIONCorporations routinely maintain petabytes of content inexpensive, on-premise storage, so it’s no surprise thatCIOs want to reduce costs by deleting or migrating staleor obsolete content. Standard storage managementsoftware applications can count the bytes, but they can’tcompute the business value of content or the risk of losingdata. Organizations need an approach for automating themigration of content to lower-cost storage, while providingfile-level traceability and real-time tracking of ROI.With a data storage optimization analytics solution, ITdepartments, CIOs, and content owners get the visibilitythey need to identify content eligible for deletion ormigration to lower-cost storage.To reduce storage, organizations must answer these FAQs: When was the content created? Last accessed?Who created the content? Are they still employed?Which business unit owns the content?Which products are associated with it?How much does it cost to store?Is it covered by a data retention policy?Can we move to low-cost cloud storage?Top Natural Language Processing Applications in BusinessFirst, organizations should use enterprise data connectorsand search engines to help them understand the enterprisecontext for their data. Connectors help join content sourcemetadata with other corporate sources, with sourcemetadata and context indexed into a search engine andmade accessible for ad hoc search and visualization.Business rules should also be developed by stakeholdersto help drive data movement to optimize storage utilizationand achieve program ROI. Once the business rules arelaid out, organizations can identify files eligible for lowercost storage, archiving, or deletion, if obsolete. Wherepossible, this should be an automated process that providestraceability.With NLP and ML, documents and their content can be moreaccurately read and identified. This drives storage costsavings through rapid detection of duplicate ornear-duplicate content, as well as providing a 360-degreeview of enterprise data.11
Data storage optimization analytics (contd.)Example business rules Retain active files on primary storageArchive content inactive for two or more yearsArchive content authored by contractors and terminatedBENEFITS Provide actionable visibility into content in high-coststorageIdentify content provenance, ownership, and frequencyof useApply domain-specific business rules to identify eligiblecontent and migrate it to lower-priced storageEnable cloud-readinessTrack savings (potentially millions per year) and computeROI in real-timeTop Natural Language Processing Applications in Business12
Sentiment analysisCHALLENGESOLUTIONSentiment analysis, also known as opinion mining, uses AIto automate the process of identifying opinions about aspecific subject from a piece of content.As mentioned earlier, it’s estimated that 80 percent of anorganization’s data is unstructured in the form of emails,chats, articles, documents, web content, and social media.Manual analysis of this information for customer, product, oremployee sentiment would be practically impossible. WithAI-driven sentiment analysis, an organization can detectpeople’s opinions or feelings about a topic and uncoveractionable insights that would otherwise be unobtainable.In its simplest form, sentiment analysis categorizes asentiment as positive or negative. It could also quantify thesentiment (e.g. -1 to 1) or categorize it at a more granularlevel (e.g. very negative, negative, neutral, positive, verypositive). By combining NLP technologies, text analytics,linguistic analysis, and ML, sentiment analysis applicationscan cope with the complexities of language.Top Natural Language Processing Applications in Business13
Sentiment analysis (contd.)Example use case:Voice of the customer – sentiment analysis is widely appliedto voice of the customer materials like reviews and surveyresponses, online and social media, and healthcare materialsfor applications that range from marketing to customerservice to clinical medicine. By using sentiment analysisacross this data, organizations can better understandcustomers (and possibly change how they engage withthem), predict demand, and improve overall companyperformance.Top Natural Language Processing Applications in BusinessBENEFITS Provide marketing and competitive intelligenceEnhance product developmentImprove customer retentionAnalyze the impact of an event (e.g. a product launch orredesign)14
Insider threat detectionCHALLENGEON THE RISE – Insider threats account for nearly 75percent of security breach incidents.According to the Ponemon Institute’s report, “2018 Cost ofInsider Threats: Global Organizations,” the average cost ofan insider threat annually is 8.76 million.Costly data breaches may originate from external culpritsor employees. But in many cases, an insider (employee,former employee, contractor, business associate, etc.) whohas authorized access to valuable data can intentionallyor inadvertently cause even greater damage. And as datavolumes and accessibility grow, organizations – governmentand commercial, large and small – need to guard their dataagainst these incidents.SOLUTIONInsider threats could involve fraud, the theft of confidentialor commercially valuable information, the theft ofintellectual property, the sabotage of computer systems,or the disclosure of information damaging to a company’sbrand or reputation.With an effective insider threat detection solution in place,organizations can avoid the lengthy legal battles – as wellas monetary and reputational losses – resulting from insidertrading, non-compliance, leaks of trade secrets, databreaches, and government intelligence leaks.However, when organizations have large volumes of bothstructured data (documents, spreadsheets, transactionrecords) and unstructured data (social media, emails, voicerecordings, notes), it becomes very difficult to conductaccurate data classification, monitoring, and analysis.Integrating search, analytics, and NLP can help to solvethis challenge.This is where integrating NLP-based insider threatapplications can help determine if there is any illegal ornefarious intent within communications and detect threatpatterns for rapid risk mitigation.Source:1. h-incidents/2. nsider-threats/Top Natural Language Processing Applications in Business15
Insider threat detection (contd.)Solution features:BENEFITS Search engines can scale to billions of recordsThreat investigation using data available via the searchengineIntegration with third-party archiving solutions andanalytics dashboardsGreater flexibility and customization for in-depth analysisand reportingA 360-degree view across organizational data frommultiple sourcesTop Natural Language Processing Applications in Business Mitigate risk by detecting patterns and identifying redflagsKeep a better pulse on the organization by customizing a“risk” score for any employeeFaster time to incident and response with comprehensiveenterprise visibilityDeeper insights with a full complement of analyticsframeworks for all threat detection workloads16
NLP in your enterpriseAnalyzing structured data alone is no longer enough.Sophisticated business analyses, predictions, and decisionmaking all need more. The use cases we’ve providedshow how unstructured content can be used to unlocktremendous new insights. With a well-implemented NLPsolution in place, organizations can enable a deeperunderstanding of unstructured content, providing enhancedBI and analytics.As your organization starts designing and building your NLPapplications, it’s essential to ensure that your IT staff and/orimplementation partners have the bandwidth and expertiserequired to conduct a thorough assessment for aligning NLPtechnologies with your business objectives. Whether yourorganization is only in the initial phase of evaluating variousNLP technologies or has already identified the preferredsolution for implementation, we can help – from evaluatingand selecting the best-suited solution to implementing,tuning, and managing the application.Looking for support and want to leverage Accenture’sNLP expertise? Contact us to discuss your goals and startdefining a strategy and implementation roadmap for yourNLP applications.Top Natural Language Processing Applications in Business17
About Search & Content AnalyticsSearch & Content Analytics, formerly Search Technologies, is part of AccentureApplied Intelligence. We live in a data-driven world. But not everyone ismaking the most of their data. 80 percent of all data is unstructured – imaginethe hidden insights trapped within unstructured enterprise content such asvoice, images and emails. At Search & Content Analytics, our mission is to helpenterprises unlock the full value within their unstructured and structured data.We combine innovative technologies such as machine learning and naturallanguage processing with search and big data analytics to transform the waypeople work. Whether it’s improving intranet and website search, monitoringinternal communications to detect insider threats, helping recruiters matchjobs to résumés, analyzing oil wellhead reports, or exploring molecular data,we bring comprehensive search and analytics services to clients acrossindustries. Clients include organizations in e-commerce, media, healthcare,financial services, recruiting, manufacturing, and the government sector.What knowledge and insights are trapped in your data? Let us help you findbetter answers.Visit our webpage to learn more about our capabilities and client work.About AccentureAccenture is a leading global professional services company, providing a broadrange of services and solutions in strategy, consulting, digital, technologyand operations. Combining unmatched experience and specialized skillsacross more than 40 industries and all business functions – underpinned bythe world’s largest delivery network – Accenture works at the intersection ofbusiness and technology to help clients improve their performance and createsustainable value for their stakeholders. With more than 469,000 peopleserving clients in more than 120 countries, Accenture drives innovation toimprove the way the world works and lives.Visit us at www.accenture.com.Copyright 2019 Accenture.All rights reserved.Accenture, its logo, and High PerformanceDelivered are trademarks of AccentureTop Natural Language Processing Applications in Business18
unstructured data as a resource for enhanced analytics and decision making. This is where natural language processing (NLP), a field of . (text or speech) Understanding content (mining insights from vast . for NLP-powered applications is fueled by a growing range of technologies, from open source fram
