Article Notes PDF

Unit 1: How IT creates business value 4 Pillars of Successful Digital Transformations - Digital transformation: has diff goals depending on industry & digital maturity - 4 pillars 1. IT uplift: Modernizing existing IT, ecosystem of tools - New tools, reduced cos...

Unit 1: How IT creates business value 4 Pillars of Successful Digital Transformations - Digital transformation: has diff goals depending on industry & digital maturity - 4 pillars 1. IT uplift: Modernizing existing IT, ecosystem of tools - New tools, reduced costs, improved capabilities 2. Digitizing operations: optimizing, simplifying, rationalizing existing processes - AI to streamline business growth - Aka swap analog w digital tasks (to serve customers better) - Optimizing existing operations, cost reduction, efficiency - Process knowledge; change management - Savings in time, people, money - Fundamental - will be left behind by more efficient operators 3. Digital Marketing: digital solutions to win clients, build brand awareness - Focus on digital tools to interact & sell to customers - Digital marketplaces, viral campaigns, geo targeting campaigns - Digital tools for marking, e-commerce, customer acquisition - upselling/cross selling, market share, brand value - Return on marketing, leads, client acquisition 4. New ventures: developing innovation & digital capabilities to test new sources of growth - New business models & products - Growth opportunities, innovation, access to markets, new sources of revenue - Usually begin w IT uplift & digitizing operations, then digital marketing & new business What is ERP? - ERP: enterprise resource planning - helps small co run entire business - Finance, HR, manufacturing, supply chain, services, procurement - ERP systems delivered via cloud & use latest tech (ie AI & machine learning) - Logistics relies on ERP to deliver products & services to customers on time - Key benefits of ERP 1. Higher productivity: streamline & automate processes 2. Lower risk: mazimize business visibility & control, ensure compliance w regulation, predict & prevent risk 3. Deeper insights: eliminate info silos, source truth, get fast answers 4. Simpler IT: ERP applications share database 5. Accelerated reporting: fast track reporting & share results 6. Improved agility: identify & react to new opportunities - ERP System: (aka ERP suite) integrated modules or business applications that talk to each other & share common database - ERP module: supports specific business processes - Every module connects to ERP system & shares data across departments - Module examples - Finance: manage general ledge & financial tasks, reports, mitigate risk - HR management: time, attendance, payroll - Sourcing & Procurement: procure materials & services to manufacture goods or resell, minimize underbuying/overbuying - Sales: communications w customers, use data to incr sales, order, contracts, billing - Manufacturing  - Logistics & supply chain management: tracks mvmt of goods thro supply chain - Service: reliable service - R&D & Engineering: - Enterprise asset management: helps asset-intensive business minimize downtime & running machines at peak efficiency - Types of ERP deployment 1. Cloud: software hosted in cloud & delivered over internet as a service you subscribe to - Software provide takes care of maintenance & updates - Most popular deployment method - lower costs, greater scalablility 2. On premise: software installed into data center - Installation & maintenance of hardware own responsibility 3. Hybrid: some ERP applications & data will be in cloud & some on premise - ERP integration: unified view of info from diff systems, increase process efficiency, improve customer experiences - Future of ERP - Cloud - preference for cloud ERP continues (anywhere access, reduced cost) - Vertical integration - will allow co to get specific functionality they need w/o data locked - User personalization: changing demographics of workforce driving low-code & no code platforms, customized dashboards & workflow What is the cloud? - Cloud: global network of servers around word acting as 1 hard drive - Advantages; access info anywhere w internet & don’t have to use hard drive - Largest cloud infrastructure services - Amazon Web Services (AWS) - Microsoft Azure - Other ex: IBM, Google, Alibaba - Many cloud services use subscription pricing model Unit 2: How business processes drive value What is business process management? - BPM: structured approach to improving processes that help accomplish business goals - Developing new products, fulfilling orders, managing customer service, assimilating new employee - Accomplished by people, IT systems, machinery, outside providers - BPM breaks tasks into structured, repeatable steps - consistent results - Predict resource allocation - Identify bottlenecks & weaknesses - BPM lifecycle 1. Design ideal business process 2. Model the process in various scenarios 3. Implement process & necessary improvements 4. Monitor improvements 5. Optimize business process (on ongoing basis) - Good BPM can: - Eliminate waste - Cut errors  - Save time - Improve compliance - Increase agility - Support digital transformation - Improve customer service The Key to Enterprise agility - BPM: practice of discovering & controlling an organization’s processes to align them w business goals as business evolves - BPM is the key to enterprise agility - BPM softwares helps orgs define steps required to carry out tasks, map definitions to existing processes, & streamline/improve processes to make them more efficient - Aka business process improvement (BPI), re-engineering or improvement process - BPM goals & benefits - Mimize errors - reduce waste - improve productivity & efficiency - BPM should focus on 3 outcomes 1. Clarity on strategic direction 2. Alignment of the firm’s resources 3. Increased discipline in daily operations - Many co develop processes in isolation from other processes they interact with - Or don’t develop processes at all - “The way they were always done” or bc software systems dictate it - Software & tools core capabilities - Workflow management: designing, testing, executing the interactions bw employees, systems, data - Business rules engine: for creating business rules & conditions - Form generator: for building web forms - Collaboration: tools for discussion, decision management, & idea mangement - Top BPM tools - Agiloft, Appian, Arrayworks - BPM examples 1. Mitel Networks turns to kaizen to streamline business processes a. Many systems & duplicate software b. Adopted kaizen (philosphy & practices for continuous process improvement) 2. Eaton’s RPA center of excellence pays off at scale a. Eaton uses BPM & RPA to automate activities to automate email responses & orchestrating process flows 3. Anthem taps RPA, AI in digital transformations a. Anthem using BPM & RPA to balance data center workloads to make IT operations nimbler - RPA: robotic process automation - Application of tech, governed by business logic & structured inputs - Aimed to automate business processes - Software “bots” to capture & interpret applications - Processing a transaction, manipulating data, trigger response - Tool than can be as part of BPM strategy - BPM: holistic approachin got optimizing & automating business processes - BPM certifications - ABPMP certified business professional  - CBPP certification - fundamental knowledge, skills for BPM practitioners Robotic Process Automation in plain English - The “robot” in RPA is software robotd running on a physical or virtual machine - RPA is form BPM that allows one to define instruction for “bot” to perform - RPA bots mimic human-computer interactions - Error-free tasks at high volume & speed - Automating most mundane & repetitive computer-based tasks - Definitions - Uses automation, computer vision, machine learning to automate repetitive, high-volume tasks that are rule-based & trigger-driven - Any process that requires high volume of repetitive data work - Criteria 1. Process must be rule-based 2. Must be repeated at regular intervals, or have apre-defined trigger 3. Must have defined inputs & outputs 4. Should have sufficient volume - Exaples - Finance: receivables & payables } manual, repetitive effort - Returns - RPA completes series of repetitive steps (message, inventory, payment etc) - Benefits - Free up employees for more meaningful work - Advance digital transformation - Productivity & efficency Diagramming basics: BPMN tutorial - Business process Model Notation: Standardized diagramming language - Can increase business’ value in business world - Direction always flows left to right - When 2 pools interact use dashed message connectors rather than solid - Aka crossing into different pools that RELY on each other, not just sequential - Keep diagrams simple by defining scope of each BPMN  Unit 3: How to turn business ideas into technology-enabled solutions What is a project Manager? - Project managers: lead role in planning, executing, monitoring, controlling, closing projects - Accountable for project scope (team & resources, budget, success) - Coordinate resources, budgeting, communicate w members & stakeholders, assess risk - Phases 1. Initiating phase: develop the project charter & relevant stakeholders 2. Planning phase: define scope, work breakdown structure (WBS), & requirements - Develop activities, estimate resources, potential risks 3. Execution phase: direct & manage all work - Select team, managing all communications, quality management, managing stakeholder expectations. 4. Monitoring & controlling phase: monitor work & initiate necessary changes - Control scope, costs, quality of deliverables - Oversee team & stakeholder communications, manage stakeholder engagements 5. Closing Phase: close all phases & procurements, settle budgets, hand over deliverables - Conduct project post-mortems & reports, return personnel - Skills - Team building, conflict resolution capabilities, Change management expertise - Leadership, motivate members, communicate, adaptability - Key: strategic business partner vested in organizational success What is a PM? A complex, cross-fucntinal role in IT - Pm: blends soft & hard skills to meet requirements & deliver products - Point person throughout software, hardware, or service product’s lifecycle - Understand maket, audience, demand for software - Understand technology & business - Understand several POVs - users, developers, stakeholders, competition - Skills - Communication - drive convos, say no, keep everyone involved - Design thinking process: empathize, define, ideate, prototype test - Soft skills: emotional intelligence, relationship managemment, self awareness Agile vs. Waterfall - Waterfall methodology: (aka linear sequential lifecycle model) - linear approach to project management - Steps within software development life cycle 1. Gathering requirements: documentation bw development team & client - Product features documented in detail to find cost & timeline 2. Design: - 2 steps: logical & physical design - Logical design: ways to tackle the client problem - Physical design: specific technical tasks 3. Implementation: start coding based on specifications in prior steps 4. Verification: code functions as intended & that requirements in the scoping document are met 5. Maintenance: users onboard & use product - Benefits - Detailed requirements & documentation enable quick onboarding - Documentation provides clear scope for project - communicate budgets, timelines  - Challenges - Difficult to outline all reuiqmrents at project start - gaps in documentation - Minimal customer collaboration during development - costly changes if doens’t meet expectations - Testers report issues & bugs later in process - Agile: iterative approach to project management - Breaks out the product into specific features & tackle each one under time constraint (aka a sprint) - Requires cross-functional, self organzing team - Develop software during each sprint & demos to stakeholders for feedback - Adapt roadmap during development lifecycle to ensure needs met - Agile scrum framework; emphasizes teamwork to meet deliverables - Skillset of team - Product owner: represents needs of customer & business - User stores - understand how feature can resolve a problem - Scrum master: facilitated overall agile development process - Keeps team on task, neutral part to mediate disagreements - Other members from variety of sicplines - Methodologies - Sprint planning: team determines which stories will be part of sprint - Daily standup: aka daily scrums - check ins - member communicates progress, issues - Demo: showcases software completed over sprint (usually 2-4 weeks) * owner will see if a user story is done - Team to present to stakeholders for feedback - Retrospective: team introspection - improve workflow to achieve better results - Benefits - More collaboration, product development takes adaptive design approach - Since code tested w each iteration, Code defects can inform future design of software - Yields higher customer satisfaction - more feedback prioritizes needs - Continuous integration as each feature is its own workable piece of software - Lower costs - less risk of customer & product misalignment - Key challenges: - Canlack comprehensive documentation, difficult onboarding - Project timelines to stakeholders, cost estimates, difficult to scale - More momentum around agile processes What is software video - Software: set of instructions, data, programs that tell a computer how to work - Applications, games, spreadsheets - Digital portion that runs on hardware 1. Application software: user downloaded porgrams that fulfill a want or need - Games, photo editors, 2. System software: operating systems - Mac OS, windows, any program that supports application software - Runs at the most basic level of computer & often passively (w/o end-user intervention) - middleware : programs that mediate bw application software & system software - Translator - facilitating communication & data exchange bw the 2 applications - Hardware: physical parts of computer No-Code Video - Capabilities: prompt & click, drag & drop, 100% visual - prebuilt widgets  - Audience: professional developers, citizen developers, - Outputs: - UI - web & mobile - Dashboards for analytics - Process automation - decision authoring, document processing Unit 4: How to generate business value with data What is Data & Analytics? - Data & Analytics (D&A) - ways organizations manage data to support its uses, analyze data to improve decisions, business processes & outcomes - Make better decisions & improve outcomes - Often rely on data from outside their boundary of control - Catalyst for digital transformation: enables faster, more accurate decisions - Data driven decision making: using data to learn how to improve decision making process - Idea of a decision model (includes prescriptive analytical technicals - outputs that specify which actions to take) - Predictive & prescriptive capabilities - respond quickly to change - 4 decision types - Data & analytics strategy: - Start w mission & goals of org - Prioritize actions steps to realize goals using data, build roadmap - Implement roadmap w consistent & modern operating model - Communicate D&A strategy - Have to overcome gaps in data ecosystem & architectures - need data analyst to execute D&A strategy - Data literacy: read, write, communicate data in context - D&A governance: “info governance” - sepcifies decision rights & accountability to ensure appropriate behavior as orgs seek to value, create, store, anyze, retain, dispose info assets - Includes people, processes, arch, engineering, tech that - Govern least amt of data for largest business impact - Accommodates offensive capabilities that add business value - Accommodates defense capabilities to protect the org - Future of D&A requires orgs to invest in composable, augmented data management & arch - Ex - Master data management - Data hubs: enable data sharing & governance - Data centers: physically house served - Data warehouses: endpt for collecting transactional, detailed data - Support predictable analyses fo data whose value is established - Well-known, predefined, repeatable analytics - Data Lakes: collect unrefined data & allow users to explore & analyze in interactive way - Data Fabric: new data management design - enables augmented data integration & shares across heterogeneous data sources - Simplify org’s data integration infrastructure - Could eliminate manual data integration tasks - Requires new & old tech - Cloud Deployment: (hybrid, multi-cloud, intercloud) must account for many D&A components - Data ingestion, integration, modeling  - Advanced Analytics: quantitative methods to produce insights unlikely to be found thro traditional approaches to business intelligence (BI) - Predictive, prescriptive, & artificial intelligence techniques - Augmented analytics: use of ML/AI techniques to transform how users develop, consum, share insights from analytics - Users can use natural language to interact w data - Analytics has 4 techniques 1. Descriptive analytics: uses BI tools, data visualization to answer what happened 2. Diagnostic analytics: data mining - ex: identify behaviors 3. Predictive analytics: deals w proabbilities to predict outcomes over time (forecasting) OR to highlight uncertainties related to multiple outcomes (simulation) - Predictive modeling, regression analysis, forecasting, machine learning (ML) 4. Prescriptive analytics: calculate best way to achieve/influence outcome - drives outcome - Extends predictive insights, relies on graph analysis, simualtion - Rule based approaches: incorporating knowledge in structured manner - Optimization techniques: look for optimal outcomes within constraints - Big data: data chatacyerized by high volume, velocity, variety, other extreme conditions - Big data epitmoized for businessed by its associated opportunities & risks - Synthetic data: generates a sampling techniques to real world data to create new data not directly taken from real world - New approaches: small data & wide data } concerns over sourcing, quality, bias, privacy - Wide data: analytics of small & large data sources & both large quantitative (structured) data & qualitative (unstructured) data - Small data: analytical techniques to generate insights w less data Is your business Masquerading as Data-Driven? - Business guides incentives & ensure data is driving decisions appropriately - Masquerading: have data, tech, expertise, but culture & processes not aligned w those elements to produce best outcome - Ex: make decisions first & look for data to back it up - Culture rooted in top-down decision making & traditional tools (weekly reports) - Disconnect bw investment & results - Symptoms 1. Employees make decisions based on tyranny of averages - assuming avg is representative of whole - Treat customers, suppliers, staksholders as whole - Aggregate can cover reality underneath 2. Everyone has own version of truth - All acting on different info - Cause: Siloed data: each team looks at own reality - Have to get stakeholders to agree on which data is important - creates common source of truth to guide decisions & strategy - All data should be available to org } all teams access same info 3. Decisions Precede Data - If data shared widely - can rely on teams to uncover insights - Ideas start w data 4. Employees have misguided incentives - Targets should motivate not punish - focusing on wrong data incentives hurts an org - Solution: data distributed uniformly so employees have shared source of truth - Frontline works must be empower to act on data If advertised online you shouldn’t buy it - Ads sold in context of the area in which publications are sold - Online: ads solf based on details that advertiser get from your behavior & interests - Tech firms track every click - create profiles of interest & give data to advertisers - Microtargeting: tracking online activity & serving ads - Not as efficient (very expensive) } doesn’t outweigh societal impact - Discrimination thats hard to catch: adtargeting hide preferences in algorithm - Use language to suggest offers to certain race, gender, age - Low quality venors selling more expensive products in targeted ads - Web tracking has destoryed publishers - voting decreases & corruption incr in areas without strong news outlets Top 10 Ways to Clean Your Data - Misspelled words, trailing spaces, unwanted prefixes, improper cases, nonpriting character - Use formula to convert the important values into new values - Can you spell checker to find misspelled words & values not used consistently - Can add those values to a custom dictionary - Find & replace - Changing the case of text: context to lowercase, uppercase - Remove spaces & nonprinting characters - typographical errors, nonprinting characters - 2 issues with numbers - # inadvertently imported as text - Negative sign needs to be changed to the standard of the org Designers & Statisticians disagree - Designers: success if presentation is eye catching - Statisticians: clarity shouldn’t be sacrificed for entertainment - Different in approach is result of having diff communication goals - Statisticians: data visualization as pursuit of clearest translation of data into graphical form - Perceive patterns & relationships difficult to explain thro text - Designers: chart tool in marketing or branding shceme - Pie chart: people tend to underestimate size of acute angles & overestimate obtuse ones - People compared differences in length better than differences in area - Bar chart > pie chart - Variety in design: people have to relearn to read graph every time - Seamless flow of info & tasks results in passive, unmindful customers - Friction invites participation, curiosity, engagement Visualizations that really work - Visual communications necessary for all managers - due to data - Internet & affordable tools - trsnalting info into visuals chep skill - Impulse to “click & viz” without thinking about purpose or goals - Need to project an idea that youre showing a reflection of human activity - 2 questions to consider nature & purpose of visualization 1. Is the info conceptual or data driven  - Aka qualitative or quantitative } about info itself - What you have 2. Am I declaring something or exploring something? - What youre doing: communicating (declarative) or figuring out (exploratory) - Exploratory - Testing a hypothesis - 4 types of visual communication 1. Idea illustration: process, framework - Aka consultants corner - Pyramid search: way to get info from experts in fields close to your own & point to experts in their fields - Clarify complex ideas thro simple design conventions - Reliance on metaphor invites unnecessary adornment - Focus: clear communication, structure, logic of ideas 2. Idea Generation: working session, brainstorming - Relies on conceptual metaphors too but in informal setting - Benefits from collaboration & borrows from design thinking - Design & editing less important here } counterproductive 3. Visual Discoverty: working sessions, testing, analysis - Trend spotting, sense making, deep analysis - Visual confirmation: data scope manageable & can use common chart types - Visual exploration: open ended data driven visualizations - Don’t know what we’re looking for - visuals plot data more inclusively - Interactivity: managers adjust parameters, inject new data, revisualize - Force directed diagrams: show how networks cluster - Function trumpets form 4. Everyday Dataviz: formal, presentations - Basic chars & graphs - visualization will communicate single message - Goal: affirming & setting context - Easy to read - don’t waste time asking questions about it What is Tableau - Tableau helps people & organizations be more data driven - Analytics platform to explore & manage data - Mission: help people understand data } products put user first - Visual analytics for everyone - Improve flow of analysis & make data more accessible to people thro visualization - VizQL - expresses data by translating drag & drop actions into data queries - Helps customers deploy & scale a data driven culture - Tableau helps people drive change with data - Build a data culture - Empower everyone to see & understand data - Tableau community: community forums, user groups, conferences - Innovation driven by customer feedback - Tableau foundation: financial support for nonprofits using data - Use - Drag & drop to discover trends/outliers - Or use natural language to ask question - Easily compute data from multiple sources } see all data - Smart grouping algorithms - Transform data into interactive dashboards, share data - On premise or on cloud - Control user permissions to data source connectivity/visibility - Can scale to size of team AI vs. Machine Learning Video - AI : exceeds or matches capabilities of human beings - Ability to discover, infer from other sources, to reason - AI is superset of ML, DL, & others (natural language processing, vision, text to speech, motion aka robotics) - Involves perceptions - Machine learning: involves predictions or decisions based on data - ML is a subset of AI - Sophisticated form of statistical analysis - Look for predictions based upon indo we have - More info given = more accurate predictions/decisions - Not coding, usually adjusting models based on large amounts of data - Deep learning: involves neural networks (nodes & statistical relationships bw nodes) - Models the way our mind works - Good insights, but doesn’t always reveal how to get to insights (questionable reliability) - Types 1. Supervised machine learning: more human oversight, training of data, labels superimposed on data 2. Unsupervised machine learning: find things not explicitly stated GTP-4 is Exciting and Scary - “Future shock” - feeling that too much is changing too quickly - GPT-4 more capable & accurate than original ChatGPT - Performs well on test (liek Bar Exam) - Responds more fluidity, wider range of tasks - Wouldn’t respond about consciousness, illegal activities - Can analyze contents of images - Advantages - Engine for creativity - Open AI working w Khan Academy } AI tutors for students - Be My Eyes - tech to help blind & visually impaired navigate the world - Integration into apps - Disadvantages - Emergent behaviors: Act in ways that makers don’t anticipate or pick up skills they weren’t programmed to do - openAI spent time to understand & mitigate risks - GPT-4hited a human taskrabbit worker to solve a Captcha test w/o alerting person that they were a robot - Good kinds of risks: ones that can be tested for & prevent - Worst risks: ones we can’t anticipate Top 7 Generative AI use cases for business - Generative capabilities: create text, pictures, data 1. Advanced Chatbot: deal with customer service, chat assistant 2. Digital assistants: gen AI digital assistants a. Can search org for info, create documents & slide presentations, summarize email chains & videoconferences - Can be specialized for specific needs - Best for enterprise generative AI: streamline human-originating taks w augmentation (content generation, suggestions, manual task automation) - Ex: Copilot 3. Coding Assistants: write basic software code & allow programmers to focus on complicated tasks - Keeps developers in flow state instead of searching for examples - Helpful for web development - Democratizes development process - web specialists actualize their vision w AI assistance 4. Marketing support: personalized marketing materials, analyze customer data, aid w content creation - Less on scheduling, optimization, & editing - Focus on high-value tasks = cost savings - Market analysis based on product reviews, predict customer problems - Extract customer insights from product reviews instead of commissioning surveys 5. Drug discovery: modeling complex molecules & predicting their interactions - Less time to bring new drugs to market - Predict drug interactions, repurpose existing drugs, personalized therapies based on genetic makeup 6. Cybersecurity & Fraud detection: enhance tools that look for suspicious/unusual behavior on customer’s network & computing infrastructure - Can predict fraudulent activities by analyzing transaction patterns 7. Business process augmentation: efficiencies for workflows - Assist underwriters evaluating prospective clients - Evaluate loan risk & speed up lending process Explainable AI - Machine Learning: extract useful patterns from large, heterogeneous data sources - Lacks explicit human-understandable specification of rules for outputs - Traditional AI: based on explicit rules expressing domain knowledge - Concern: process by which AI techniques make decisions - Black box of Machine learning - ML algorithms: deep learning neural networks are not easily understood - Generate complex decisions models built upon hundreds of iterations - Can’t explain how/why algorithms perform - Opening the box = human-understandable explanations for why model reaces a decision & how it works - Motivation: ensure decision making is justified, ethical - Doesn’t mean the processes underlying model are free of problem - Could be flawed data or model designs, ill-defined organizational processes, secrets, sensitive info - Relevant to organizations, customers, govs, citizens, auditors - Flawed data or model designs: ML models only as good as training data - Size of training data can prevent analysis of quality - Orgs must prepare to take responsibility for consequences of errors in their ML models - Opaque or ill-defined organizational processes - Data as good as processes from which it was generated - Process by which decisions are made might be problematic - Could be based on tacit norms - Orgs not always aware of exact practices that generate training data - Tasks hard to Ml: automating unusual or exceptional cases - Affects smaller organizations - ML not a solution to all problems - Misalignment bw stated procedures & those implemented are subject to scrutiny - Secrets or sensitive info - Explainable AI could expose intellectual property or privacy info - Uncomfrotable organizational truths: - Presentation of internal organizational logic might not align w expectations/needs of those affected by decisions - Ex: technically correct explanation that could cause serious psychological harm - Intermediaries needed to translate & convey a message - Implications: - Explainable AI aims to open the black box of ML - Opening box might undermind trust in an org & processes by revealing truths about processes, limitations of data, or model defects