Vin Vashishta, one of the world's best experts in data strategy and leadership, shared in his talk about the key pitfalls of organizations that fail to unlock the value of data teams. According to him the common traits and mindset of successful leaders should focus on the following:
⚡️Adopt a holistic strategy encompassing digitalization, data, analytics, and the use of AI and machine learning
⚡️ Integrate AI in user-centric product discovery to create an alignment in dataset curation and the user-centric values
⚡️Consider the user experience of machine learning (ML UX) to deal with the unreliability of the data-driven outcome
⚡️Educate literacy around AI and data-driven use cases in cross-functional teams
⚡️Launch AI in production along with MLOps for the long-term and continuous learning of data-driven use cases
⚡️Track why-KPIs and transformation metrics for an organizational transformation
⚡️Build coalition around AI use cases and leverage coalitions to get a seat in the strategy planning process
⚡️Enable leadership roles and c-level champions who can foster the adoption of AI
2. DataScience.vin
Who Am I? Vin Vashishta
Product Lines With $100M+ In ARR
Clients: Fortune 100, SMEs, and Startups
V Squared Is One Of The Oldest Data Science &
AI Strategy Consulting Companies
25+ years in tech
10 in data science
Decision support systems
Customer behavioral models
Pricing models
Technical
Started in product management
in 2010
First published in 2013
Last 8 years as a C-level data &
AI strategy Advisor
Strategy
15 years in leadership roles
2X Founder
Launched V Squared in 2012
Build data organizations from 5
to 37 data professionals
Leadership
3. AI Last Mile Problem
DataScience.vin
Why Is Value Trapped?
Failure To Launch
Miss The Mark For User
Or Customer Needs
Unusable And Incapable Of
Integrating
Unreliable
4. DataScience.vin
Behaviors & Attitudes
Leaders Laggards
Centralized Organization & Resources
Continuous Transformation
Product and Process Managers
Tracking Transformation Metrics
C-level Champion
Articulates Reliability Requirements
Continuous Improvement Paradigm
Disparate Teams & Initiatives
1-Off Transformation
Treat Data Initiatives Like Software
Tracking Vanity Metrics
A Few Frontline Advocates
Assumes All Data & Models Are The Same
No Monitoring Process
Laggards vs Leaders
5. DataScience.vin
Leaders: Centralized Organization & Resources
Laggards: Disparate Teams & Initiatives
More Easily Managed
Better Innovation Mix
Lower Cloud & Infrastructure Costs
Higher Project Velocity
Lower Headcount Requirements
Single Source For Data
Higher Security & Privacy
Advantages
7. DataScience.vin
Leaders: Product and Process Managers
Laggards: Treats Data Initiatives Like Software
Data Research
Model
Development
Maintenance
& MLOps
Low Reliability &
Human-In-The-Loop
Products
Artifacts: Unique
Datasets &
High Reliability Models
Product-Grade
Implementations That
Scale
Continuous
Improvement & Early
Issue Detection
8. DataScience.vin
Leaders: Articulates Reliability Requirements
Laggards: All Data & Models Are The Same
Raw Data & Simple Reporting
Descriptive & Analytical Models
Diagnostic, Predictive &
Prescriptive Models
High Reliability or
Autonomy Use Cases
Low Reliability or
Human-In-The-Loop
Use Cases
9. DataScience.vin
Leaders: Continuous Model Improvement
Laggards: Fire & Forget With No Monitoring Process
Leaders: C-level Champion
Laggards: A Few Frontline Advocates
Leaders: Tracking Transformation Metrics
Laggards: Tracking Vanity Metrics
10. DataScience.vin
Early Maturity Phase 3 Maturity
Phase 2 Maturity
Initial Assessment Data Strategy AI Strategy
Technology
Model
Transformation
Strategy
Operations
Platform
Roadmap
Transformation
Roadmap
Data Product
Roadmap
AI Product
Roadmap
D&A
Organization
Build Out
Innovation
Strategy
Business Model
Transformation
Strategy
Data & AI Maturity Model
11. 5 Steps That Launch A
Data Strategy
5 Steps That Derail A Data
Strategy
Assess The Business's Data Maturity
Establish Data As An Asset And Define Its
Value For The Business
Build Coalitions Around Use Cases
Leverage Coalitions To Get A Seat In The
Strategy Planning Process
Define Why The Business Uses Data To
Support Existing Strategic Goals
1.
2.
3.
4.
5.
Start With Technology
Ignore Users And Customers
Confuse Strategy With Tactics Like Hiring
Or Buying Infrastructure
Keep Initiatives Siloed In The Data Team
And Ask For Buy In Alone
Position Data And AI As The Solution To
Every Problem
1.
2.
3.
4.
5.
Starting A Maturity Journey
12. DataScience.vin
I appreciate you all for coming and giving
me your valuable time.
Questions? Comments?
Learn more with my Substack
vinvashishta.substack.com/
Join one of my cohorts in November.
Learn Data & AI Strategy or Strategic Leadership from me.
Discover more at DataScience.vin