The AI-Driven Data Team
Kaup valmöguleikar
Rapidly transform your analytics teams to deliver AI-driven insights.The AI-Driven Data Team by Nicholas Kelly is a proven-to-work playbook for data and analytics leaders who want to align analytics capabilities with the demands of an AI-powered business environment. Written for leaders accountable for data, analytics and business intelligence, this book provides tools for diagnosing capability gaps among data analytics teams, modernizing legacy stacks and delivering AI-driven insights to help organizations make better decisions.
You'll learn how to:- Diagnose skills gaps and map AI-augmented career paths- Integrate modern AI tools with existing analytics stacks- Launch six revenue-driving and cost-focused pilots in 90 days- Embed governance without slowing innovation- Apply an ROI framework, governance checklist and 90-Day CharterDrawing on expert insights and real-world applications, this book helps you upskill your analysts, strengthen AI governance and provide AI-driven insights that drive real results.Themes include: AI strategy, data governance, analytics leadership, ROI from AI, organizational transformation, executive decision-making
Nánar um bókina
- Kogan Page
- 9781398627581
- 9781398627574
- ePub
- 1
- Nicholas Kelly
- English
- 2026-07-03
- 100
- 10
- 10
Kaflar
- Cover
- Endorsements
- Titlepage
- Dedication
- Contents
- About the author
- Acknowledgments
- PART ONE Setting the Groundwork (Days -14 to 0)
- 1 Building the AI-Driven Data Team
- What “AI-Driven” Actually Means
- Roles, Reimagined
- The Common Thread
- What This Means for You
- Key Takeaways
- 2 AI as a Force-Multiplier
- Why 20–35 Percent Capacity Reclaimed Is Just the Starting Point
- The Data Team, Reimagined
- Building the Business Case: From Hype to Evidence
- Key Takeaways
- 3 Mindset Reset
- The Evolution from Dashboards to Decisions
- Why Dashboards Failed (and Why It’s Not Your Fault)
- The Business Value Question: Your North Star
- Reduce Interfaces, Raise Adoption
- From Insight to Impact: Closing the Last Mile
- Key Takeaways
- 4 Skills and Tools Matrix
- Introducing Decision Velocity
- The Skills Matrix: Mapping Your Teams’ Capability Landscape
- Role-by-Role Stretch Zones
- The Tools Matrix: Matching Capability to Acceleration
- Key Takeaways
- 5 Building Your AI-Ready Team
- Data Analyst
- BI Developer
- Data Engineer
- Data Scientist
- Product Owner/Project Manager
- Key Takeaways
- 6 Upskill vs. Hire vs. Outsource
- What Should We Do About Our Skill Gaps?
- Why Hiring Isn’t the Silver Bullet Leaders Imagine
- Outsourcing: Fast Lift or Hidden Drag?
- Upskilling: The Highest-Leverage First Move
- The Five-Minute Cost Calculator
- Key Takeaways
- 7 Readiness Audit
- The Decision Loop as a Reality Check
- ASK: Can the Right Questions Be Raised Safely and Clearly?
- ANALYZE: Can the Work Be Trusted, Reused, and Explained?
- ACT: Can Outputs Be Used Without Heroics?
- ASSESS: Can Outcomes Be Reviewed Without Blame or Amnesia?
- Reading the Flags, Not Chasing Perfection
- Key Takeaways
- 8 The 90-Day Charter
- What a 90-Day Charter Is (and What It Is Not)
- Inputs: What You Bring into the Room Before You Write a Word
- Choosing the Three Objectives That Matter Enough to Win This Quarter
- Key Performance Indicators (KPIs)
- Budget
- Scope and Boundaries
- Operating Rhythm
- 90-Day Charter Template
- Sign-Off: The Moment It Becomes Real
- Key Takeaways
- PART TWO Sprint 1: Foundations and Upskills (Days 1 to 30)
- 9 Kill the Bottlenecks
- Why Bottlenecks Are the Real Starting Line
- Find the Bottlenecks That Are Actually Stealing the Week
- Choose the Week 1 Targets (Don’t Boil the Ocean)
- The “Week 1 Automation Ladder”
- Turn Manual Prep Into a Reusable Asset
- Reporting Without Theater
- Define Success in Week 1 (Evidence, Not Vibes)
- The Week 1 Plan (a Simple Five-Day Sequence)
- Common Failure Patterns (and How to Avoid Them)
- Key Takeaways
- 10 Coding Copilots
- What Copilots Change (and What They Don’t)
- The Citizen Data Scientist Boundary
- The BlockDuster Example: The Predictive Question
- Choose the Right “First Models” in Ops and Finance (No Heroics)
- AutoML Track: A Responsible Prototype in 90 Minutes
- Vibe Coding Track: Copilot-Assisted Python
- Minimum Validation Standard (Holdout and Required Metrics)
- Explainability Minimums: “Can You Defend This in a Meeting?”
- Where the Output Lands: From Prediction to Decision
- Deliverable: The Predictive Prototype Packet
- Key Takeaways
- 11 Analytics Engineering Reimagined
- The Minimum Operating Standard
- Version Control Without the Drama
- Sprint 1 Checklist
- Tests Without the Pain
- The Lightweight Metrics Layer
- BlockDuster in Action
- Two-Week Implementation
- Key Takeaways
- 12 Redesigned Workflows
- Why Tickets Fail in AI-Enhanced Work
- The Decision-Linked Backlog
- Four Work Types, Four Lanes
- Intake and Triage: Soft Gate, Hard Honesty
- Learning Sprints: Two Weeks, Two Visible Wins
- The Weekly Operating Rhythm
- Measuring Flow: From Output to Decision Velocity
- Templates: One Board and One Sprint Plan
- Key Takeaways
- PART THREE Sprint 2: Six Quick-Win Pilots (Days 31 to 60)
- 13 Selecting Lighthouse Use Cases
- What a “Lighthouse” Means in This Context
- The Three-Lane Funnel for Sorting Candidates
- The Non-Negotiable Gate
- Landing Patterns That Force Behavior Change
- Running the Selection Meeting in 60–90 Minutes
- Key Takeaways
- 14 Minimum-Viable AI Stack
- The Two Worlds Rule: Explore vs. Publish
- The Minimum Stack as Capabilities, Not Tools
- The Default Pattern for Conversational BI
- Cloud-First Minimum Stack
- On-Prem-First Minimum Stack
- Hybrid Minimum Stack
- BlockDuster Walkthrough
- The Promotion Gate
- The One Anti-Pattern: No Ownership
- Key Takeaways
- 15 Build, Buy, or Extend
- Definitions That Prevent Arguments
- The Principle Stack
- Tools and Model Families
- The Red/Amber/Green Scorecard
- Rule of Thumb
- BlockDuster Walkthrough
- Key Takeaways
- 16 Pilot Execution Playbook
- The Sprint 2 Pod: Five Hats, One Goal
- The Pilot Contract: One Decision, One Weekly Moment
- Security and Governance Pre-Wire
- Week 1: Baselines, Boundaries, and the First Memo
- Week 2: Retrieval-First, Bounded, Cited
- Week 3: Wire It into a Rhythm
- Week 4: Operate, Learn, Decide
- The Pilot Packet
- Key Takeaways
- PART FOUR Sprint 3: Scale and Govern (Days 61 to 90)
- 17 Lightweight Machine Learning
- The Minimum Operational Standard
- The Deployment Pattern
- Observability That a Small Team Can Actually Run
- Governance Questions and Edge Cases
- The Runbook Plus Rollback
- Retraining Without Debate
- Key Takeaways
- 18 Plain-English Governance
- When These Questions Get Asked (Without Slowing Everything Down)
- The 10 Must-Ask Questions
- The “Good Enough” Governance Packet
- Key Takeaways
- 19 Bias, Privacy, and Security
- A Plain-English Threat Model for Data Leaders
- The Default Safety Architecture: Retrieval-First, Bounded, Cited, Permissioned
- Concrete RAG Examples: From Question to Safe Answer
- Security Checklist: Sprint 3 Minimum
- Privacy and PII Redaction: Three Layers of Defense
- PII Redaction Walkthrough: Before and After
- Bias: Publish-Safe Checks Without Turning into a Research Lab
- The 90-Day Safety Operating Rule: Human-in-Loop by Design
- Key Takeaways
- 20 Storytelling That Lands: Explaining Models Everyone Can Trust
- Data Storytelling vs. Trust Storytelling
- The Trust Narrative Arc
- BlockDuster Trust Story
- The Three Trust-Killers (and the Antidotes)
- Silent Changes (Yesterday vs. Today)
- Pocket Failures (Right on Average, Wrong Where It Matters)
- The Simple Model Trust One-Pager
- How to Present It Without Triggering Resistance
- Key Takeaways
- PART FIVE The AI Flywheel (Day 90+)
- 21 Measuring Impact: From Hours Saved to Decision Velocity
- The Universal Scoreboard
- Finance-Grade Rules of the Road
- The Attribution Ladder
- The ROI Framework
- Converting Hours Saved into CFO Language
- Detailed Calculation Example: Analyst Time Savings
- The CFO Value Pack
- The BlockDuster Example
- Key Takeaways
- 22 Learning Loops and Communities
- The Upgrade Fund
- How to Set the Percentage Without Starting a Budget Fight
- What the Upgrade Fund Buys
- Learning Program Structure
- Skill Diffusion
- Community-Building Strategies
- Case Study Showcase
- The Compounding Mechanism
- Key Takeaways
- 23 The Human in the Loop
- The Three Anxieties That Kill Adoption
- The Journey of the Team
- What Stayed Human
- Long-Term Sustainability: Building Systems That Last
- Forward-Looking: The Next Wave
- Your Next Seven Days
- A Closing Story: Three Years Later
- The Real Ending
- Index
- Copyright