Index

• 20 percent rule (Google), 24
• 100 percent test coverage, 124

• A
• Acceptance, 125f
  • tests, 124
• Access
  • control systems, 87
  • methods, 75–76
• Accountability, 181
• Ackerson, Chris, 169–172
• Advanced Encryption Standard (AES), 91
• Agile, 101
  • framework, 103
    • application, 164
  • process, 68
  • term, usage, 48
  • user stories, generation, 59
  • weakening, 136–137
• Akolkar, Rahul, 183–186
• Algorithms, public appearance, 159
• AlphaSense, 170
• Amazon Web Services (AWS), 111, 167
• Analytics software, 82
• Apache Hadoop, usage, 8, 85
• Apache JMeter, 132
• Application code, promotion, 121f
• Application programming interface (API), 75
  • internal APIs, 112
  • programmatic calls, 124
  • usage, 110–111
• Artificial Intelligence (AI)
  • adoption, 6–10, 98, 172
    • challenge, 170, 174, 176, 178, 181, 184, 188
    • roadmap, 7f
  • advances, 171, 174, 176, 178, 181, 185, 188
  • AI-focused meetups, 144
  • algorithm, usage, 6
  • assistance, job function target, 172, 174, 176, 178, 182, 186, 188
  • benefits, 175
  • Bin-Picking product, model reliance, 3
  • capabilities, 64
    • overestimation, problem, 45, 196
  • chatbots, avoidance, 43
  • Cloud Services catalog, sample (IBM), 111f
  • data consumption, 58–59
  • deep learning (relationship), Venn diagram (usage), 21f
  • development, 157
  • expertise, 184, 187
  • experts, 169
  • focus, narrowness, 44
  • implementation, process, 164–165
  • lifecycle, 10, 51, 139, 167–168
    • pitfalls, 159–161, 206–207
    • roadmap action checklist, 193
  • limitations, 41–44
  • pitfalls, 44–45
  • potential, 163
  • power/responsibility, 158–159
  • project plan, building, 64–66
  • security, 177–178
  • solution, outsourcing, 100–101
  • strong/weak artificial intelligence, 42
  • technology, data scientist application, 82
  • training AI models, data availability, 73f
• Artificial Intelligence (AI) models
  • comments, 154
  • data availability, 73f
  • description, 151, 153
  • files, 152, 153
  • hosted demo, 153
  • library
    • building, 150–155
    • components, 151–153
    • entry, example, 153–154
    • indexing mechanism, 151
    • solutions, 154–155
  • licensing, 152, 154
  • metrics, 152, 153
  • parameters, 152, 154
  • performance, quantification, 145–147
  • tagging, 151
  • tags, 153, 154
  • technology, 152, 154
  • testing, example, 125–127
  • training data, 152, 154
  • validation data, 152
• Artificial Intelligence (AI) system
  • creation, 76
  • human intervention, 129–131
  • improvement, 144–145
  • information flow, 30f
  • knowledge, 71
  • learning process, 142–144
  • output generation, 82
  • robust AI system, ensuring, 128–129
• Asset, Liability, Debt and Derivative Investment Network (Aladdin) software, deployment, 5–6
• Automated chat support system, 128–129
• Automated testing, 124–128
• Awesome Public Datasets, 78
• Azure (Microsoft), 167

• B
• Babbage, Charles, 13–14
• Back propagation (backpropagation), 18, 42
  • neural network, 22
• Backups, usage, 93–94, 202–203
• Bad data (AI limitation), 43
• Balanced ground truth, 72–73
• Balanced scorecard, perspectives, 63
• BERT model, 170, 171
• Binary files, 150
• Black box, 42
• BlackRock, Inc., case study, 5–6
• Bootstrapping, 79
• Boundaries, 58–69
• Bradford, Jeff, 173–174
• Brainstorming, usage, 38–41
• Business
  • business-critical code, release, 156
  • process mapping, 27–28
  • stakeholders, categories, 32

• C
• Cambridge Analytica, practices, 92–93, 202
• Capital allocation, group (filtering example), 36
• Categorizing, 34–37, 79
• Cause and effect (AI limitation), 42
• Chance encounters, value, 38–41
• Change
  • Agile process, 68
  • proactive policy, 25
  • request procedures, neglect, 198–199
  • request process, 52
• Chaos Monkey code (Netflix), 128
• Charges/budgets, 52
• Chatbots, 23, 130
  • architecture, sample, 131f
  • login interface, 50
  • product comparisons/reviews, 98
  • support chatbot, logical architecture, 107f
  • technologies, sample, 108t
  • usage, hybrid approach, 130–131
• Chief Technology Officer (CTO), process map creation, 33
• Chunking, 17
• Classifier
  • confusion matrix, ample, 146f
  • image classifier, building, 79–80
  • machine model, 72
• Classifying, 35–37
• Client-dependent variables, requirement, 4
• Cloud
  • APIs, 110–111
  • APIs, SLA, 135
  • cloud-based API model, 111
  • deployment paradigms, 133–135
  • provider, 135
  • scalability, relationship, 132–133
  • solutions, 202–203
• Cloud Engine Learning Engine (Google), 132–133
• Cloud Machine Learning Engine (Google), 111
• Cloud Platform (Google), 167
• Code base, parts (identification), 155–156
• Code defect, appearance, 127
• Code repository, creation, 118
• Comma-separated value (CSV)
  • files, 78
  • format, 75
• Computer security, 92, 201
• Conclusion, forming, 38
• Confusion matrix, sample, 146f
• Constructive criticism, destructive criticism (contrast), 39
• Constructive feedback, 83–84
• Containers, 134
• Continuous integration, 119–123
  • benefit, 123
  • pipeline, 119–121
  • principles, 123
  • true continuous integration, 121–123
• Convolutional neural network (CNN), 23
• Core competency, 99
• Co-relations, AI usage, 158
• Core technology, replacement, 114
• Cost-benefit analysis, usage, 38
• Creativity, fostering, 24
• Cross-departmental exchanges, 40–41
• Cross-disciplinary knowledge base, 148
• Crowdsourcing, platform/usage, 79–80
• Culture, focus, 44–45, 196
• Customer
  • balanced scorecard perspective, 63
  • interaction, reduction, 65
  • satisfaction
    • increase, 65
    • KPI example, 62
  • user stories, 65
• Customer relationship management (CRM) system, 84

• D
• Daily stand-ups (team role), 102–103
• Dark designs, 92
• Data, 173–174
  • access, 84–85
  • analysis, third-party vendors usage, 92–93
  • anonymization, 87
  • availability, 27
  • bad data (AI limitation), 43
  • collation, 29
  • consumption, 58–59
  • curation (technology adoption phase), 8, 71, 157, 166
    • pitfalls, 90–94, 199–203
    • roadmap action checklist, 192
    • tasks, 81
  • data-powered capabilities, 110–111
  • data-restore operation, 202
  • exploration, 74–75
  • governance, 166
    • roadmap action checklist, 192
  • improvement, 157–158
  • licensing
    • disadvantage, 79
    • insufficiency, 90, 199
  • non-free data, licensing (disadvantages), 78–79
  • readiness, 89–90
  • recovery, untested backup failure, 93
  • responsibility, 89
  • science flow, 82f
  • scientist
    • AI technology application, 82
    • role, 81–82
  • security, insufficiency, 91–92, 200–201
  • storage technology, 82
  • temporary data, usage, 90
  • text-based data, 154–155
  • training data, 152, 154
  • transformation, tasks, 81
  • tweaking, 9
  • validation data, 152
• Database structure, reverse engineering, 75–76
• Data collection, 73–80, 180–181
  • crowdsourcing, usage, 79–80
  • licensing, usage, 77–79
  • opt-in/opt-out, 86
  • policies, 86
• Dataflows, tracking, 29–30
• Data.gov, 78
• Data governance (technology adoption phase), 8, 71, 85–89, 157
  • board, creation, 87–88
  • completion, goal, 86
  • initiation, 88
  • pitfalls, 90–94
• Datasets, labeling, 80
• Decision tree process, 130
• Deduction complexity, compounding, 4
• Deep learning, 21, 82, 152
  • AI relationship, Venn diagram (usage), 21f
  • models, usage, 3
• Defect tracking, digitization, 57
• Deliverables
  • defining, 50
  • list, 51
• Delphi method, 7, 60, 166
• Delphi technique, usage, 104
• Demo, production, 103–104
• Departmental cross-talks, implementation, 40–41
• Design thinking, 7, 53–58, 61, 166
  • process, 54f, 64–65
  • session, sample, 53–54
• Developers, code conflict (absence), 119
• Development
  • environment, 119
    • promotion, 120
  • feedback loop, 105–106
  • team
    • micromanagement, 136–137, 205
    • role, 102
• Digital internal data collection, 74–76
• Digital neuron, firing, 20–21
• Direct database connection, 75–76
• Docker (container technology), 134
• Documentation, quality, 151
• “Do Not Track,” browser requests, 85
• Duesterwald, Evelyn, 177–179

• E
• Electromagnetic interference (EMI), external stressor, 59
• ELIZA, usage, 16
• Embedded code, ubiquity, 1
• EMNIST dataset, 72
• Empathy, AI limitation, 43–44
• Empathy maps
  • creation, 55–56
  • generation, 65
  • sample, 56
• Employee
  • impact, group (filtering example), 36
  • performance (KPI example), 62
  • time, saving (KPI example), 62
  • turnover/attrition, 147
• Encryption, 86
• Epochs, number (deep learning), 152
• Error cost escalation, 107–108
• Explainability, 42, 181
• External stressors, mitigation, 59

• F
• F1 score, 147
• Failover protocol, 129–130
• Fallback capability, 130
• False positive/negative, 145
• FANUC Corporation
  • case study, 2–3
  • robot, example, 4f
• FANUC Intelligent Edge Link & Drive (FIELD), usage, 3
• Feedback
  • constructive feedback, 83–84
  • gathering, 103
  • ignoring, 160
  • loops, 82–84, 105–106
    • application, 121–122
    • continuation, 135
  • mechanisms, 141
  • qualitative feedback, 65
  • receiving, 142
  • stages/roles, 105f
  • user feedback, incorporation, 140–142
• Fibonacci Sequence, 104
• File export, 75
• Filtering, 34–35
• Filters, usage, 26
• Financial balanced scorecard perspective, 63
• Financial success, measurability criteria, 51
• Firm
  • discovery, 99–100
  • references, request, 100
• First-mover advantage, achievement, 23–24
• Flores, Steven, 187–189
• Flowcharts
  • sample, 34f
  • usage, 28–29
• Formal change request procedures, defining (neglect), 68
• Freelancer.com, 101
• Friedl, Jeffrey, 16
• Fully connected neural network, multiple layers, 21f

• G
• General Data Protection Regulation (GDPR), 88–89, 166
• Generalizations (AI limitation), 41–42
• GitHub, 128
• Goals
  • defining, 56–57
  • setting, 102
  • transformation, 57
• Goldstein, Rob, 6
• Google search results, usage, 81
• Governance
  • data governance, 8, 71, 85–89
  • term, application, 85
• Graphical processing units (GPUs), usage, 134–135
• Grinder, The, 132
• Ground truth, 71
  • absence, 90–91, 200
  • balanced ground truth, 72–73
  • distribution, building methods, 72
  • proportional ground truths, 73
• Guide rails, presence, 129

• H
• Hashing, 86–87
• Health Information Technology for Economic and Clinical Health Act (HITECH Act), 88
• Health Insurance Portability and Accountability Act (HIPAA), 88
• Hidden Markov models, 19–20
• H&R Block, case study, 4–5
• Hybridization, 118
• Hype (AI limitation), 43
• Hypercare, 139
• Hypothesis/predictions, forming/testing, 38

• I
• IBM
  • AI Cloud Services catalog, sample, 111f
  • Watson, capabilities, 4
  • Watson Services, 111
• Idea
  • creation/discovery, 31
  • expected returns, group (filtering example), 36
  • grouping, example, 37f
  • independent requests (number), group (filtering example), 36
• Idea bank
  • control, 148
  • maintenance, 25–27
  • organization, 35–37
  • review, 37–38, 147–148
  • sample, 26t–27t
  • updating, 147–148
• Ideation, 165
  • focus, narrowness, 195
  • pitfalls, 44–45, 195–196
  • process, problems, 44, 195–196
  • roadmap action checklist, 191
  • technology adoption phase, 6–7, 13
• If-this, then-that decision, 130
• Image classifier, building, 79–80
• Image recognition/classification, 23
• Imitation game, 14
• Indexing mechanism, 151
• Information
  • flows, 29–31, 30f
  • personal information, 77
  • tracking, 29–30
• Infrastructure-as-a-service (IaaS), 133
• “Infrastructure as code” paradigm, 134
• Infrastructure testing, 127–128
• Innovation
  • culture, impact, 44–45
  • innovation-focused organization, 23–25
  • method, 38–39
  • priority, 25
• Integration, 125f
  • tests, 124
• Intelligent business model, 164
• Interactive Voice Response (IVR) system, 43
  • usage, 64
• Internal APIs, 112
• Internal data collection
  • digital component, 74–76
  • physical component, 76–77
• Internal process, balanced scorecard perspective, 63
• Internal systems, monthly review meetings (incorporation), 141
• Internet of Things (IoT), 1
  • devices, 74
  • usage, 157
• Inverse document frequency, 18

• J
• Java-ML library, 111
• JavaScript, usage, 110
• Jenkins (CI tool), 123
• Just-in-time (JIT) ordering, 2

• K
• Kaggle, 78
• Key drivers, selection/modeling, 60
  • examples, 62
• Key performance indicators (KPIs), 48
  • measurability criteria, 51
• Keys, 17
• Knowledge base, 148–150
  • cross-disciplinary knowledge base, 148
  • expansion, 150
  • materials, categories, 149
  • online knowledge base, features, 149
• Knowledge sharing, 141
• Kubernetes, 134

• L
• Language-independent method, usage, 110
• Learning chatbot, project creation, 65
• Learning/growth, balanced scorecard perspective, 63
• “Lessons learned” document, usage, 145
• Licensing, usage, 77–79
• LoadRunner, 132
• Load testing, usage, 132
• Load tests, 131
• Logical architecture diagram, usage, 106
• Logic, explainability, 42
• Long short-term memory (LSTM) neural networks, usage, 22–23
• Lovelace, Ada, 13–14

• M
• Machine learning, 18–19, 82
  • models, 121, 132–133
  • precision, definition, 146
  • system, building, 205–206
• Machine Learning on Azure (Microsoft), 111
• Machine model, 72
• Mailing lists, feedback mechanisms, 141
• Market research, usage, 54–55
• Markov chains, 19
• Markov models, 19
  • hidden Markov models, 19–20
• Markov property, 19
• Mastering Regular Expressions (Friedl), 16
• Measurability criteria, 51
• Metadata, curation, 5
• Microservices, 110–111
• Models. See Artificial Intelligence models
• Monthly report, replacement, 57
• Multiserve deployment, 128
• Mutual trust, value, 136–137

• N
• Naive Bayes approach, 204
• Narrow AI, definition, 171
• Natural language processing (NLP), 15, 82
  • application, 5
  • BERT model, 170, 171
  • machine learning, 17
  • programmatic NLP, 15–17
  • statistical NLP, 17–18
• Need-based approach, 40
• Nephew, Jill, 42, 179–183
• Networked devices, 74
• Neural networks, 20–23
  • convolutional neural network (CNN), 23
  • impact, 114
  • single neuron example, 20f
• Nightly build, running, 120
• NodeJS application, 111
• Non-free data, licensing (disadvantages), 78–79
• Numerical analysis, 82

• O
• Observation, making, 38
• OneSignal, profitability, 77
• Online knowledge base, features, 149
• Open source, 168
  • community, leveraging, 156
  • contribution, 155–156
  • projects, impact, 156
  • technologies, usage, 98
• Operating system (OS), copy, 134
• Organizational chart, enhancement, 28f
• Organizational flowcharts, usage, 28–29
• Out-of-the-box IoT equipment, usage, 2–3
• Overfitting, 84

• P
• Palantir, founding, 158
• Parameter tweaking, 9
• Performance
  • AI model performance, quantification, 145–147
  • benchmark level, defining, 51–52
  • testing, 120
  • tests, 131
• Periodic updating, facilitation, 143
• Permission structures, 10
• Personal information, 77
• Personas
  • determination, 54–55
  • development, 64–65
• Physical architecture, 108f
  • diagram, 107
• Physical data, value, 77
• Physical internal data collection, 76–77
• Precision, term (usage), 146
• Present circumstances, assessment, 60
• Probability, 17
• Problems, invention/misrepresentation, 67
  • pitfalls, 197
• Process
  • flowchart, 34f
  • map, CTO creation, 33
• Production, 117, 167
  • code defect, appearance, 127
  • code repository, creation, 118
  • environment, 119
  • model, promotion, 122f
  • pitfalls, 135–137, 204–206
  • roadmap action checklist, 193
  • skills, absence, 137
  • team/schedule, 30
  • technology adoption phase, 9–10
• Product owner (team role), 101
• Products, tagging/categorization, 34
• Programmatic NLP, 15–17
• Programming language, selection, 109–110
• Programming techniques, usage, 16
• Project
  • AI pieces, process, 118
  • assumptions, 51
  • breakdown, approaches, 53–61
  • completion/assumption, 160, 206
  • completion criteria, 51–52
  • defining (technology adoption phase), 7–8, 47, 165–166
    • pitfalls, 196–199
    • roadmap action checklist, 191–192
  • definition, pitfalls, 66–68
  • deliverables, list, 51
  • focus, problem, 44–45, 196
  • governance, 49–50
  • kickoff, 50
  • measurability, 62–63
  • oversight activities, 49–50
  • plan
    • building, 64–66
    • components, 48–52
    • metrics, 48
  • roadmap, clarity, 103
  • scope, 49
  • stakeholders, feedback, 97
  • success criteria, 65
  • work schedules/locations, 49
• Proportional ground truths, 73
• Prototype
  • code, leveraging, 117–118
  • design, 106–107
  • reuse, 117–118
  • technology scales, ensuring, 131–133
• Prototyping (technology adoption phase), 8–9, 97, 166–167
  • excess, problem, 113, 203–204
  • pitfalls, 112–114, 203–204
  • planning, problem, 113
  • roadmap action checklist, 192–193
  • solutions, 97–99
  • tool, problem, 113–114
• Python
  • HTTP library, 111
  • selection, 109–110
• PyTorch, 151

• Q
• Qualitative feedback, 65
• Quality assurance (QA)
  • specialists, 56, 57
  • team, monitoring activity, 55

• R
• Ranking, 17, 35–37
• Recall
  • equation, 146
  • metric, 146–147
• Recurrent neural network (RNN), 22–23
• Red Hat, 156
• Reports accuracy (KPI example), 62
• Representational state transfer (REST) API, 110–111
• Research, performing, 38
• Results, iteration/sharing, 38
• Risk, group (filtering example), 36
• Rituals, importance, 44
• Rivest-Shamir-Adleman (RSA), 91
• Robinson, Nathan S., 175–177
• Robust AI system, ensuring, 128–129
• Row-level security, 92, 201

• S
• Safari Books Online, 149
• Salesforce (CRM system), 84
• Scalability
  • cloud, relationship, 132–133
  • handling, 122
  • testing, 120
• Scenario planning/analysis, 60, 61
• Schedule, 30, 49
  • estimation, 52
• Schwaber, Ken, 101
• Scientific method, usage/steps, 38
• Scrum
  • framework, 103
  • master (team role), 102
  • overview, 101–103
• Service level agreement (SLA), 135
• Shortcuts, coding, 123
• Single sign-on (SSO), 10
• Smoothing, inclusion, 19
• Social media, usage, 67
• Sociological, technological, economic, environmental, and political inputs (STEEP) model, 60
• Software as a service (SaaS) architecture, usage, 123
• Software checks, conducting, 92–93, 201–202
• Solutions, 97–99
  • premature construction, 67–68, 198
• Sorting, 34–35
• Source control, code conflict (absence), 119
• Spear phishing, 91
• Specification document, design (problem), 49
• Sprint
  • planning (team role), 102
  • review, 103
• Stage environment (test environment), 119
  • testing, 121
• Staging environment, 122–123
• Stakeholders
  • buy-in, absence, 66, 196–197
  • categories, 32
  • feedback, 97
    • gathering, 103
• Standard operating procedures (SOPs), 28–29
• Statistical NLP, 17–18
• Step size (deep learning), 152
• Story
  • points, usage, 104
  • user stories, 57–58, 65
• Strong artificial intelligence, 42
• Subject matter experts (SMEs), 51–52, 72, 136, 205
  • feedback, 83
  • usage, 76
• Subsystems, usage, 59–60
• Success, criteria, 50, 65
  • absence, 68, 199
• Support chatbot
  • logical architecture, 107f
  • physical architecture, 108f
• Support query, resolution (KPI example), 62
• Sutherland, Jeff, 101
• Switches, 163
• System
  • bootstrapping, 79
  • building, decision point, 109
  • confidence, erosion, 129
  • power, leveraging, 81
• Systems planning, 7, 61
• Systems thinking, 58–60

• T
• Tagging, 34
  • format, 151
• Tags, 153
  • selection, 37
• Talent, employing/contracting (contrast), 99–101
• Technological evaluation, 9
• Technology, 144–145
  • adoption
    • end user resistance, 136
    • fear, mitigation, 205
    • phases, 6–10
  • selection, 107–110
    • chatbot technologies, sample, 108t
• Temporary data, usage, 90
• TensorFlow, usage, 109–110, 151, 155
• Term frequency, 18
• Term frequency-inverse document frequency (tf-idf), 17
  • reliance, 19
• Test-driven development (TDD), 10
• Test environment, 119
• Testing frameworks, 10
• Test types, 124–125
• Text-based data, 154–155
• Text generators, 23
• Text-to-speech capability, provision, 67
• Thiel, Peter, 158
• Think Fridays (IBM concept), 24
• Third-party vendors, usage, 92–93
• Time, group (filtering example), 35–36
• Tokenization, 17, 92, 201
• Tone analysis technology, 130
• Training
  • AI models, data availability, 73f
  • data, 152, 154
• Transparency, value, 136–137
• Travis CI, 123
• True continuous integration, 121–123
• True positive/negative, 145
• Turing, Alan (Turing test), 13–15, 42
  • standard interpretation, 14f

• U
• Unit testing, 125f
• Unit tests, 124
  • creation, 125
• Upwork, 101
• User
  • password hashing, 86–87
  • privacy, ignoring, 92–93, 201–202
  • rights/privacy, respect, 93
  • training, inadequacy (provision), 160–161, 207
  • type, defining, 55
• User experience (UX) skills, 100–101
• User feedback
  • ignoring, 160, 206–207
  • incorporation, 140–142
• User/security model, 9–10
• User stories
  • accomplishment, 102
  • defining/creating, 57–58
  • development team construction, 117
  • establishment, 65
  • prioritization, 97, 103–104

• V
• Validation, defining, 120
• Value
  • analysis, 31–34, 104
    • example, 33–34
  • concept, 31–32
  • defining, 104
  • identification process, 32–33
  • monetary value, relationship, 32
• Vendor selection, risk, 67–68
• Venn diagram, usage, 21f
• Verification, defining, 120
• Virtual machines (VMs), 133
• Visibility gaps, 157
• Visual defect identification system, building, 56
• Visualization, popularization, 133

• W
• Walls, separation ability, 40–41
• Weak artificial intelligence, 42
• Web-based APIs, usage, 110–111
• WebLOAD, 132
• Weizenbaum, Joseph, 16
• Word2vec, development, 20
• Word embedding, 22
• WordPress (blog), 149, 154
• Work
  • schedules/locations, 49
  • stream activities, 50–52
• Workload, example (spike exhibit), 133f