Automated Decision-Making and Artificial Intelligence PDF
Document Details
Uploaded by PlayfulTragedy
Rome Business School
2024
Tags
Summary
These lecture notes from Rome Business School cover automated decision-making and artificial intelligence, including a case study on a national retailer. The material discusses topics like data-driven strategy, the importance of data analytics for business success, and challenges in implementing such strategies in different business contexts. The provided information gives an understanding of real-world applications of artificial intelligence in improving business operations.
Full Transcript
Automated decision-making and Better Managers for a Better World Artificial Intelligence Module: Data Analytics Master: Blended Law, Media, Food, Energy OL Date of lecture: October 2024...
Automated decision-making and Better Managers for a Better World Artificial Intelligence Module: Data Analytics Master: Blended Law, Media, Food, Energy OL Date of lecture: October 2024 romebusinessschool.com Agenda 1. Business case (Class discussion) 2. Automated decision-making and Artificial Intelligence 3. Data-driven company success cases 4. Q&A romebusinessschool.com Better Managers for a Better World Disclaimer All product and company names are trademarks or registered® trademarks of their respective holders. The use of registered names, logos, brands, etc. does not imply any affiliation with or endorsement by them. romebusinessschool.com Better Managers for a Better World Business case (Class discussion) romebusinessschool.com Better Managers for a Better World Data-driven Strategy for Retail Home ware retailer 30 stores all over EU Annual turnover of 200 millions of € Available data on customers, products, inventories, sales, operations, etc. romebusinessschool.com Better Managers for a Better World 5 Data-driven Strategy for Retail Challenge Implement an effective decision-making process, based on customers’ behaviors and product performance data, to develop a data-driven strategy Working in teams, please define: 1. at least 5 questions (Business needs) the DDDM process is expected to answer 2. the main steps to set-up and implement an effective DDDM process (from data acquisition to data elaboration and visualization) 3. at least 3 outputs (KPIs, Reports, Graphs, etc.), related to the Business needs, needed to be evaluated for making effective decisions romebusinessschool.com Better Managers for a Better World 6 The real Case study that has inspired our Business case https://medium.com/@datasmiles/how-data-analytics-revolutionized-a-national-retailer-a-case-study-b9dfadc5398d romebusinessschool.com Better Managers for a Better World 7 How Data Analytics Revolutionized a National Retailer Executive Summary (1/2) GGV: national retailer with approx. 40 stores across the U.S. and 8-digit annual revenue Data Smiles, partner involved to address their fragmented data infrastructure GGV struggled with disparate reporting systems from multiple software vendors, leading to inefficiencies and decision-making bottlenecks The proposed comprehensive solution involved building a centralized data warehouse, complete with ETL data pipelines, and deploying Power BI reports for unified analytics https://medium.com/@datasmiles/how-data-analytics-revolutionized-a-national-retailer-a-case-study-b9dfadc5398d romebusinessschool.com Better Managers for a Better World 8 How Data Analytics Revolutionized a National Retailer Executive Summary (2/2) The new ecosystem covers a wide range of operational metrics, including sales, customer behavior, products, inventory, and store operations As a result, GGV now has a single, centralized hub for all their reporting needs This streamlined approach has not only expedited decision-making but also unlocked analytical capabilities that were previously unattainable The project has effectively transformed GGV into a data-driven organization, well-equipped for both current challenges and future growth https://medium.com/@datasmiles/how-data-analytics-revolutionized-a-national-retailer-a-case-study-b9dfadc5398d romebusinessschool.com Better Managers for a Better World 9 How Data Analytics Revolutionized a National Retailer Background The company’s founder and CEO is very tech-savvy and data-driven He wants his company to make decisions based on data The Head of Operations, a skillful and obsessed operation leader, observes their stores on a daily basis and makes all types of adjustments in order to improve their efficiency and effectiveness, trying new tactics or twisting their processes just to see if they work Their mind was full of questions, from understanding their customers’ behaviors to analyzing their product performance They also want to optimize their operations through data analysis. https://medium.com/@datasmiles/how-data-analytics-revolutionized-a-national-retailer-a-case-study-b9dfadc5398d romebusinessschool.com Better Managers for a Better World 10 How Data Analytics Revolutionized a National Retailer Some questions deserving an answer… What are the optimal opening and closing times for each store? How does severe weather, like blizzards or storms, impact store performance? How effective are vendor promotions in driving new product sales? Which products, brands, and categories are top performers? What’s the best-selling brand within each category and product within each brand? Are there popular products that underperform in specific stores? Where should the next store location be? Who are the top customers, and what criteria should be used to define them? How can lapsed customers be re-engaged through promotions and gifts? https://medium.com/@datasmiles/how-data-analytics-revolutionized-a-national-retailer-a-case-study-b9dfadc5398d romebusinessschool.com Better Managers for a Better World 11 How Data Analytics Revolutionized a National Retailer Initial Problem Statement The company had no centralized data infrastructure As a consequence, they couldn’t effectively track important metrics like sales, inventory, or customer behavior Reports were isolated and couldn’t be shared across departments, making it difficult for the team to act on any insights Definitely, GGV was operating without a cohesive data system The challenge was to build a comprehensive data ecosystem from scratch, enabling GGV to monitor key operational metrics and make informed decisions https://medium.com/@datasmiles/how-data-analytics-revolutionized-a-national-retailer-a-case-study-b9dfadc5398d romebusinessschool.com Better Managers for a Better World 12 How Data Analytics Revolutionized a National Retailer Data Strategy Framework The goal: empowering both executive leadership and operational teams to make decisions informed by high-quality, actionable data Strategic roadmap segmented into four pivotal stages: 1. Data Collection: cornerstone of any data strategy, it involves the gathering of raw data from diverse sources, including customer interactions, transactional records, and supply chain metrics 2. Data Centralization: consolidate the collected data into a centralized data warehouse, not only to enhance data accessibility but also to ensure data integrity and security, leveraging on cloud-based solutions like AWS or Azure to provide scalable and secure data storage options 3. Data Processing: data transformations and pre-aggregations to prepare the data for analysis, such as data cleansing, normalization, and the application of machine learning algorithms for predictive analytics, with data models tailored to align with GGV’s specific business questions and KPIs 4. Reporting and Analysis: deployment of advanced analytics tools (like Power BI or Tableau) for data visualization and reporting, with custom dashboards designed to provide real-time insights, enabling agile decision-making across all levels of the organization https://medium.com/@datasmiles/how-data-analytics-revolutionized-a-national-retailer-a-case-study-b9dfadc5398d romebusinessschool.com Better Managers for a Better World 13 How Data Analytics Revolutionized a National Retailer Solution Data collection and data quality Capture customer data: Loyalty program encourage shoppers to identify during in-store transactions Identification process integrated into the checkout workflow: each transaction is uniquely associated with an individual customer Build the technological architecture https://medium.com/@datasmiles/how-data-analytics-revolutionized-a-national-retailer-a-case-study-b9dfadc5398d romebusinessschool.com Better Managers for a Better World 14 How Data Analytics Revolutionized a National Retailer Solution Customer Segmentation using RFM (Recency, Frequency, Monetary) Customer360 RFM score + Last Transaction + Preferred store + Secondary store + Average per transact Product360 Product Name, Variant Name, Store Name, Inventory Level, Sold in 1mo/3mo/6mo, Sell thru Rate 1mo/3mo/6mo, Last sale date https://medium.com/@datasmiles/how-data-analytics-revolutionized-a-national-retailer-a-case-study-b9dfadc5398d romebusinessschool.com Better Managers for a Better World 15 How Data Analytics Revolutionized a National Retailer Solution Location heat map for new stores https://medium.com/@datasmiles/how-data-analytics-revolutionized-a-national-retailer-a-case-study-b9dfadc5398d romebusinessschool.com Better Managers for a Better World 16 How Data Analytics Revolutionized a National Retailer Results Operational Efficiency: significant reduction in the time spent generating reports Centralizing all reporting into a single hub has streamlined the decision-making processes, for quicker and more informed choices. Customer Engagement: loyalty program to enhance customer satisfaction The rewards system has not only incentivized customer loyalty but also created a positive feedback loop that encourages ongoing engagement. Cost Savings and ROI: tangible efficiencies in operations (i.e. inventory management) The data-driven approach has optimized the product and brand selections, maximizing profitability. Risk Mitigation: unexpected benefit, ability to assess the impact of external factors New data insights allowed confident decisions to close stores when necessary, such as when severe weather conditions were occurring, without adversely affecting revenue. https://medium.com/@datasmiles/how-data-analytics-revolutionized-a-national-retailer-a-case-study-b9dfadc5398d romebusinessschool.com Better Managers for a Better World 17 How Data Analytics Revolutionized a National Retailer In summary The data strategy has empowered GGV to become a more agile, efficient, and customer-centric organization. The project has not only solved immediate operational challenges but also laid the groundwork for sustainable growth and competitiveness. https://medium.com/@datasmiles/how-data-analytics-revolutionized-a-national-retailer-a-case-study-b9dfadc5398d romebusinessschool.com Better Managers for a Better World 18 Automated decision-making and Artificial Intelligence romebusinessschool.com Better Managers for a Better World https://www.gartner.com/en/newsroom/press-releases/2014-10-21-gartner-says-advanced-analytics-is-a-top-business-priority romebusinessschool.com Better Managers for a Better World 20 2D analysis romebusinessschool.com Better Managers for a Better World 21 3D analysis romebusinessschool.com Better Managers for a Better World 22 4D analysis romebusinessschool.com Better Managers for a Better World 23 5D analysis + time romebusinessschool.com Better Managers for a Better World 24 6D analysis 7D analysis 8D analysis... n-D analysis Questa foto di Autore sconosciuto è concesso in licenza da CC BY-ND romebusinessschool.com Better Managers for a Better World 25 Can we trust in Data Algorithms Automated Decision-Making (ADM) systems Artificial Intelligence ??? romebusinessschool.com Better Managers for a Better World 26 Can we trust Automated Decision-Making (ADM) systems? European Parliament resolution on AI & ADM February, 12 2020 - Automated decision-making processes: ensuring consumer protection, and free movement of goods and services Consumers should “be properly informed about how the system functions, about how to reach a human with decision-making powers, and about how the system’s decisions can be checked and corrected” Humans must always be ultimately responsible for, and able to overrule, decisions that are taken in the context of professional services”, e.g., legal professions https://www.europarl.europa.eu/doceo/document/TA-9-2020-0032_EN.html https://eucrim.eu/news/ep-resolution-artificial-intelligence-and-automated-decision-making romebusinessschool.com Better Managers for a Better World 27 Example of Automated Decision-Making (ADM) process An online bank is offering loans Clients insert their data of the bank website An algorithm produces results on whether they should be offered a loan or not, and the suggested interest rate Someone of the bank company/organization needs to review the algorithm output before communicating the decision to the prospective client and inform him that he may express his opinion and eventually contest the decision, keeping in mind that the individual has the right not to be subject to a decision based solely on automated means https://ec.europa.eu/info/law/law-topic/data-protection/reform/rules-business-and-organisations/dealing-citizens/are-there-restrictions-use-automated-decision-making_en romebusinessschool.com Better Managers for a Better World 28 Can we trust Automated Decision-Making (ADM) systems? European Parliament resolution on AI & ADM February, 12 2020 - Automated decision-making processes: ensuring consumer protection, and free movement of goods and services ADM systems should use: high-quality and unbiased data sets explainable and unbiased algorithms https://www.europarl.europa.eu/doceo/document/TA-9-2020-0032_EN.html https://eucrim.eu/news/ep-resolution-artificial-intelligence-and-automated-decision-making romebusinessschool.com Better Managers for a Better World 29 Unbiased data sets? https://www.statice.ai/post/data-bias-types https://towardsdatascience.com/types-of-biases-in-data-cafc4f2634fb https://www.visualcapitalist.com/wp-content/uploads/2018/03/18-cognitive-bias-examples.html romebusinessschool.com Better Managers for a Better World 31 Explainable and unbiased algorithms? https://www.nature.com/articles/d41586-018-05707-8 https://www.geeksforgeeks.org/5-algorithms-that-demonstrate-artificial-intelligence-bias https://www.technologyreview.com/2020/10/20/1009452/ai-has-exacerbated-racial-bias-in-housing-could-it-help-eliminate-it-instead romebusinessschool.com Better Managers for a Better World 32 https://translate.google.com romebusinessschool.com Better Managers for a Better World 33 vs. romebusinessschool.com Better Managers for a Better World 34 https://cloud.google.com/vision romebusinessschool.com Better Managers for a Better World 35 Poor quality datasets and biases lead to Unreliability Inaccuracy Discrimination Privacy violation that affect fundamental rights https://fra.europa.eu/sites/default/files/fra_uploads/fra-2022-bias-in-algorithms_en.pdf https://www.vice.com/en/article/jgq35d/how-a-discriminatory-algorithm-wrongly-accused-thousands-of-families-of-fraud romebusinessschool.com Better Managers for a Better World 36 What is AI (Artificial Intelligence)? romebusinessschool.com Better Managers for a Better World 37 What is AI (Artificial Intelligence)? IBM: a field, which combines computer science and robust datasets, to enable problem-solving Encyclopedia Britannica: the ability of a digital computer to perform tasks commonly associated with intelligent beings ChatGPT: a multidisciplinary field focused on developing intelligent machines capable of performing tasks that typically require human intelligence, such as perception, reasoning, learning, and decision-making https://chat.openai.com https://www.ibm.com/topics/artificial-intelligence https://www.britannica.com/technology/artificial-intelligence romebusinessschool.com Better Managers for a Better World 39 What is AI (Artificial Intelligence)? OECD (Nov 2023): an AI system is a machine- based system that can, for explicit or implicit objectives, infers, from the input it receives, how to generate outputs such as predictions, content, recommendations, or decisions that can influence physical or virtual environments. Different AI systems vary in their levels of autonomy and adaptiveness after deployment. https://oecd.ai/en/wonk/ai-system-definition-update romebusinessschool.com Better Managers for a Better World 40 Applications of AI https://www.aicra.org/aicrapost/two-major-activities-of-ai-to-process-human-intelligence romebusinessschool.com Better Managers for a Better World 41 Classifications of AI https://smartcompliance.co/blog/artificial-intelligence-improve-compliance romebusinessschool.com Better Managers for a Better World 42 Classifications of AI https://smartcompliance.co/blog/artificial-intelligence-improve-compliance romebusinessschool.com Better Managers for a Better World 43 Classifications of AI e AI rativ G en e https://smartcompliance.co/blog/artificial-intelligence-improve-compliance romebusinessschool.com Better Managers for a Better World 44 Artificial Intelligence (AI) romebusinessschool.com Better Managers for a Better World 45 Artificial Intelligence (AI) ChatGPT https://www.howtheygrow.co/p/how-openai-grows romebusinessschool.com Better Managers for a Better World 46 Artificial Intelligence (AI) https://www.howtheygrow.co/p/how-openai-grows romebusinessschool.com Better Managers for a Better World 47 romebusinessschool.com Better Managers for a Better World 48 romebusinessschool.com Better Managers for a Better World 49 Unclear legal implications of using generative AI copyright infringement ownership of AI-generated works unlicensed content in training data Mitigate potential risks when using generative AI ensure that the training data are free from unlicensed content develop ways to show provenance of generated content https://hbr.org/2023/04/generative-ai-has-an-intellectual-property-problem romebusinessschool.com Better Managers for a Better World 51 romebusinessschool.com Better Managers for a Better World 52 romebusinessschool.com Better Managers for a Better World 53 romebusinessschool.com Better Managers for a Better World 54 AI hallucinations AI outputs not justified by its training data when insufficient, biased or too specialised A lawyer submitted a brief citing six judicial decisions He produced the brief using ChatGPT The court found out that all the six judicial decisions were non-existent https://www.reuters.com/legal/transactional/lawyer-used-chatgpt-cite-bogus-cases-what-are-ethics-2023-05-30 romebusinessschool.com Better Managers for a Better World 55 XAI - Explainable Artificial Intelligence AI systems that offer explanations on how a certain decision or conclusion is reached Understand AI output to build trust Explainable Artificial Intelligence needs Human Intelligence https://www.datanami.com/2018/05/30/opening-up-black-boxes-with-explainable-ai https://edps.europa.eu/press-publications/press-news/blog/explainable-artificial-intelligence-needs-human-intelligence_en romebusinessschool.com Better Managers for a Better World 56 The jobs of the future… …haven't been invented yet The jobs of the future… …are soft skills based https://www.weforum.org/agenda/2020/10/top-10-work-skills-of-tomorrow-how-long-it-takes-to-learn-them romebusinessschool.com Better Managers for a Better World 57 AI whisperers deeply understand ML algorithms data structures programming languages possess strong problem-solving skills ability to identify opportunities for improving AI models and algorithms effectively communicate with machines understand how AI systems process and analyze data interpret output and results https://www.linkedin.com/feed/update/urn:li:activity:7070419815860989952 https://d1lzrgdbvkolkd.cloudfront.net/The_Anatomy_of_the_A_I_Whisperer_d2d1cfc3e2.pdf romebusinessschool.com Better Managers for a Better World 58 March 13, 2024 European Parliament adopts the Artificial Intelligence Act (AI Act) Unacceptable risk Prohibited: AI-enabled manipulative techniques, social scoring, biometric categorization, facial recognition, remote biometric identification High risk Compliance required: AI systems with significant harmful impact on the health, safety and Limited risk fundamental rights (i.e. employment, admission, etc.) Transparency: Chatbots, Deepfake, Minimal risk Biometric for (cyber)security https://www.europarl.europa.eu/doceo/document/TA-9-2024-0138_EN.html romebusinessschool.com Better Managers for a Better World 59 Data-driven company success cases romebusinessschool.com Better Managers for a Better World Cherry Sprite Analysis of data from Social Channels Analysis of data from Beverage Machines https://www.forbes.com/sites/bernardmarr/2017/09/18/the-amazing-ways-coca-cola- uses-artificial-intelligence-ai-and-big-data-to-drive-success romebusinessschool.com Better Managers for a Better World 61 Fast food, slow service Acquisition of Dynamic Yield (AI/ML engine) for $300 Million Drive-through: suggestions based on time of day, weather, most popular products, but also purchase history Kiosks and online (App) Big Data acquisition of customers behaviors Enhanced and personalized customer experience https://www.forbes.com/sites/forbestechcouncil/2019/04/26/what-you-can-learn-from-mcdonalds-acquisition-of-dynamic-yield romebusinessschool.com Better Managers for a Better World 62 Analysis Users' data: history, purchases, reviews Market data: trends, prices, competition Operational data: warehouses, logistics, etc. Actions Customized offers Dynamic prices Fraud prevention Supply Chain optimization https://www.businesstechweekly.com/operational-efficiency/data-management/big-data-use-case romebusinessschool.com Better Managers for a Better World 63 Supply Chain 4.0 Company Suppliers Customers Purchasing Production Distribution Internal Supply Chain Product or Service Flow Information Flow romebusinessschool.com Better Managers for a Better World 64 Analysis Users' data: history and reviews but also pauses, skipped scenes, unfinished movies RT operational data from streaming infrastructure 1. Data should be accessible, easy to discover, and easy to Actions process for everyone Personalized recommendations 2. Whether your dataset is (based on the user's habits and the behavior of similar users) large or small, being able to New productions visualize it makes it easier Proactive approach to explain UX/UI optimization (A/B Testing) $1 million prize for an algorithm to increase the 3. The longer you take to find accuracy of the recommendation engine by 10% the data, the less valuable it https://research.netflix.com/research-area/analytics becomes romebusinessschool.com Better Managers for a Better World 65 U.S. bank: improving Call Center productivity and employee retention Problem Different call centers varied greatly in productivity and employee retention Why specific locations were more successful than others? Data Analysis Communication patterns, employee satisfaction, work Solutions habits, and tenure A new break schedule Highest levels of productivity and engagement: employees with the most cohesive networks, who 23% increase in productivity regularly socialized and communicated with others The majority of this communication happened during 28% increase in retention scheduled break times https://humanyze.com/case-studies-major-us-bank romebusinessschool.com Better Managers for a Better World 66 HR Analytics Analysis Examples Employees’ personal data: Experian: savings +$10M/year recruitment data, tenure, promotion history, performance, role, salary, Credit Suisse: savings +$70M/year location, job role, and more Best Buy: +0,1% of engagement Sentiment Analysis → +$100,000 of revenues Clarks: +1% of engagement Actions → +0.4% of performance Ability to predict who might be most likely to resign Shell: +1% of engagement → -4% of absenteeism Ability to predict also why someone might resign https://www.aihr.com/blog/hr-analytics-case-studies romebusinessschool.com Better Managers for a Better World 67 Google People Analytics to make people happier, healthier, and more productive Actions Project Oxygen: how to make great managers PiLab: how maintain a productive working environment, also acting on diets Unbiasing, to support and encourage diversity and inclusion Proactive retentions Better hiring decisions https://rework.withgoogle.com/subjects/people-analytics romebusinessschool.com Better Managers for a Better World 68 Questions & Answers romebusinessschool.com Better Managers for a Better World Conclusions romebusinessschool.com Better Managers for a Better World Conclusions 1. The world is VUCA, or BANI, or both 2. Data is the key to make informed decisions 3. Data-driven companies gain a real competitive advantage 4. Being a data-driven company is mainly a cultural rather than a technological aspect 5. A clear vision and a clear data strategy are fundamental 6. Data Security and Data Governance are key elements to consider 7. AI: handle with care 8. You can’t improve what you don’t manage 9. You can’t manage what you don’t measure romebusinessschool.com Better Managers for a Better World 71 One last hint Always have a View challenges as opportunities Growth Mindset Persist in facing obstacles Efforts as the path to mastery as intelligence can be developed Learn from feedback as brain can be trained Get inspired by success of others Focus on process, not on result romebusinessschool.com https://fs.blog/carol-dweck-mindset Better Managers for a Better World 72 No part of this video or any of its contents may be reproduced, copied, modified of adapted, without the prior written consent of the author, unless otherwise indicated for stand-alone materials. Copyright Rome Business School All rights reserved romebusinessschool.com Better Managers for a Better World Better Managers for a Better World Thank you Via Giuseppe Montanelli, 5 00195, Roma RM romebusinessschool.com