Lecture Notes on Business Processes and IT PDF

- Aim prompts new forms/uses of energy 1. Automate processes - offload low-value work 2. Democratize data - make info accessible 3. Reduce user friction - making doing things easier - IT enables Network Effects: the more people involved, the more valuable the tech - IT enables...

- Aim prompts new forms/uses of energy 1. Automate processes - offload low-value work 2. Democratize data - make info accessible 3. Reduce user friction - making doing things easier - IT enables Network Effects: the more people involved, the more valuable the tech - IT enables global commerce: - Ebay reduces user friction, automates processes - IT lowers cost of production & services - Ex: robots } automate processes - IT increases personal & organizational productivity How IT creates Business Value - Digital transformation: rewiring of how an organization operates - build competitive advantage by deploying tech - Improve customer experience & lower costs - Information systems - Functional information Systems - IS that supports personal productivity - Facilitates everyday personal or professional tasks - Ex: microsoft office, Grammarly - Network Information Systems - IT that supports interaction & collaboration - Facilitates communication & info sharing among employees - Ex: Gmail, Zoom, WhatsApp - Enterprise Information Systems - IT that supports business processes - Facilitates workflows & processes of an organizaion - Ex: SAP, Microsoft, Workday - 4 Pillars of Successful Digital Transformations - How are pillars enabled by - Functional IS - Crowdsourcing - Ideas reside outside companies - Working collaboratively & competively - Rely on wisdom of crowd - wikipedia, Waze - Cirque du Soleil How Business Processes Drive Value - Business process: an activity or set of activities that accomplish a specific organizationl goal - Ex: payroll, hiring, procurement, sales, order fulfillment - Logically related activities or tasks performed together to produce a defined set of results - Information Systems: carry out business processes by capturing their data & executing their activities - Ex: flight check in 1. Passenger check in (phone, kiosk, counter) } workflow same  2. Issue the boarding pass 3. Tag luggage & send to cargo 4. Security screening 5. Board the passenger - Robotic Process Automation (RPA) - Process by which a software bot uses a combo of automation, computer vision, machine learning to automate repetitive, high volume tasks that are rule based & trigger driven - AI can enhance RPA by handling variation & complexity (ex: missing data, ambiguity - Read & sort documents - Interpret text - Analyze images & generate reports - Classify complex info - Business process modeling - Visual representation of a business process - Manage exiting processes - Understand - Communicate - Diagnose - Manage process change - Investigate alternatives - Purpose changes - BPM are organizational roadmaps - Business process modeling notation (BPMN) - 1 standard notation for modeling business processes - The objective is to describe a process in a common way to all users, regardless of the tool used to describe the process - Common Symbols in BPMN - Start & end nodes - Indicates where the process starts & ends - Can have multiple ends pts depending on outcomes - Depicted using a circle - Thin circle = start - Thick cicle = end - Activities - A named activity, process, or task that occurs over time - Transforms inputs into output - Performed by participant (role or system) - Depicted using a rectangle - Naming convention: active verb + noun - Flows - Sequence and/or mvmt of data or materials - Depicted w arrow - Arrow in direction of flow - If name, use a noun phrase (invoice, packing list) - May include passive verb (validated record, audited tax form) - Input: flow that’s transformed or consumed by a process/system - Output: flow that is produced by a process/system  - Gateways - Model decision points (ex: diverging flows) - Depicted using a diamond - Label w a question - Label exit arros close to the gateway - Pools & Swim lanes - Pool: defines perimeter of the process - its name identifies the modeled process - Pool can be subdivided into swim lanes that represent actors - Allows us to associate activities - Roles are NOT specific people (ex: scientist not einstein) Donut ordering process - Roels: customer, cashier, kitchen - Activities: - Cashier: cashier takes order, sends order to kitchen, if in stock then collects money, gets donuts from kitchen gives order - Customer: drives up to customer, orders food, - Step 3: Complete a business process organizer - Roles become actors in swim lanes - Activities are associated w roles - Decisions become gateways - flows connect activities Unit 3: How to turn business ideas into technology solutions - Application software: perform a specific task for user - Word, imessage, SAP - System software: make the hardware accessible to application software user; provides a way to read & write files - Windows, mac OS, android - Create software using programming languages or low-code/no-code platforms - Programming: python, R, javascript - Convenience, do thinks you can’t otherwise do - Low-code/no-code: Thunkable, Altair AI studio - Assembling statements together - less flexibility - Options for implementing IT solutions 1. Buy (use) existing solutions 2. Create new IT solutions - In either case, IT solutions are products delivered via 1 or more projects - Projects & products are fundamental to business - Projects: means of getting things done in an organization - Products: means of delivering value to customers - The information technology project - “All projects are temporary effort to create value thro unique product, service, result” - Projects have: - Defined start & end - Specific objectives - Scope (boundaries) - Budget  - But not a routine function - routine functions are never ending - The IT product - Hardware, software, mobile app, website, user interface, software service, product feature - Product manager: strategic role, CEO of product - Researchings, communicating vision to stakeholders, strategic plan - Project manager: operational role - Breaking down initiatives into tasks, timelines - 2 project management approaches - Waterfall methodology: well defined, sequential steps - Efficient, reliable delivery based on initial specifications - Agile methodology: sequential steps of limited scope that repeat (iterations) - Quick & continuous delivery - Make course corrections - Hybrid approach - Waterfall & agile used together - Waterfall used at beginning & agile used in “develop” phase Managing projects: waterfall vs. agile - Neither is good or bad - Project success depends on the use of approach & details of project - Waterfall: clearly defined project - Agile: know less about whats wanted, need to adapt constantly - Hybrid: part of it well defined How to generate business value with data - Issue 1: Data must be organized - Data are observtions/symbols/representations that are recorded - Info is data placed in a meaningful context - Knowledge is application of info to achieve a goal - Issue 2: We are drowning in data - 2.5 quintillion bytes of data is generated each day - Business analytics Process 1. Prepare: collect/gather data - clean & transform data for analysis 2. Perform: analyze data - ex: sentiment & trend analysis, machine learning 3. Use: make recommendations to aid decision making & generate business value - Types of data analytics - The past - Descriptive: what has happened? - Summarize & aggregate raw data - often need basic math - Diagnostic: why did it happen? - Find relationships in the data - can require statistical techniques (ex: regression) - The future - Predictive: what is going to happen? - Use past data to make predictions about future - use AI to learn patterns - Prescriptive: what should we do? - Uses descriptive & predictive analytics - uses AI but w a focus on recommendation - Pitfalls of analytics initiatives  - Tyranny of averages - avg skewed by outliers - Multiple versions of the truth - only want 1 version of data being used by employees - Decisions precede data - reinforcing a wrong conclusion/bias - Misguided data-driven incentives - looking at wrong data to get certain outcome - Analytics & digital transformation - IT uplift: get better analytics tools, find bottlenecks - Digitizing operations: increase efficiency - Digital Marketing: better campaigns, targeting maketing - New ventures: predict sales of new veneers - predictive analytics How to Communicate well with data Why are effective visualizations important? 1. Improves speed & efficiency of data processing 2. Reduce time to insight 3. Some data should be visualization to make sense Best practices - State your key point - Be complete yet concise - Avoid necessary clutter - Acknowledge data source - Use color wisely - Sometimes text is best - Start axes from 0 - large deficit when doesn’t start from 0 - Avoid chartjunk How AI and Machine Learning Create New Business opportunities - “Explainable” AI - processes behind AI, gives reasoning - Dangers - Emergent behaviors - Privacy - Bias - people hwo made it & data its trained on - 2 business problems suited for predictive & prescriptive analysis 1. Churn analysis (predictive): customers most likely to go to a competitor - Issue: high churn (customers leave) - Solution: give incentives to those most likely to leave - Ignore the ones that are not predicted to churn - Insight: not all customers have the same value 2. Cross selling (prescriptive): What products customers are likely to purchase - Issue: identifying target customers - Insight: likelihood of purchase depends on items already purchased - Incentivize customers who have bought associated items w a coupon - Place product next to associate products - Predictive: what’s going to happen - Prescriptive: what you should do - ML uses unstructured data to make predictions about next work  - Supervised ML: give data to build data to make predictions - Decision trees (Supervised ML): used classify data according to a pre-defined outcome - Based on characteristics of that data - Supervised = “train” the model w a data set of correct answers to predict outcomes - Association mining (unsupervised ML): find which events predict occurrence of other events - Often used to see which products are bought together - Not training a model, you are looking at relationships in data - Higher lift = less likely to be random chance - Large language models (deep learning): predicts the next word in its responded based on training data (wikipedia, books, articles, journals, web scraping) - Both unsupervised (finds associations in data) & supervised (gets feedback about answers) - Has rules to guide its responses - Challenges of machine learning models - Accuracy of decision making: - naive/incomplete models - bad or incomplete data - ChatGPT hallucinations - Transparency & explainability: black box problem - Fairness & algorithmic bias: “ability for people to feasibly modify the outcome of algorithmic decision