Driving Data Science Adoption in Enterprise BI
Unifying Data Science & ML Inside WebFOCUS This started as a concept in the annual roadmap. I took it from idea to shipped architecture — bringing mockups to the table first, with tickets following after. I owned all three workflows and eventually handed off to two designers.
I brought mockups first — tickets followed.
— From roadmap concept to shipped architecture, owning all three DSML workflows
At a Glance
Three powerful data science features existed but were scattered across the platform. Different entry points, different patterns, low adoption. Users didn't know they were there.
Unify them into one discoverable experience. Make data science accessible without hiding it.
Defined the architecture and base screens for all three. Established consistent interaction patterns and made everything responsive.
Integrated three features into the Hub — no extra space, everything inside. NLQ and Insights live in production. DSML Hub shipping 2027.
The Transformation
Narrated walkthrough of the fragmented legacy workflows vs. the unified modern experience.
Current Workflows (Public)
NLQ (Natural Language Query)
Standalone NLQ workflow in Explore Data — separate entry point, different mental model.
Automated Insights
Standalone Automated Insights workflow — another separate entry point with its own interface.
ML Functions (Predict Data)
ML Functions from Reporting Server — redesigned with unified patterns.
Unified DSML Hub (Password Protected)
The new unified DSML Hub: all three DSML capabilities consolidated into one cohesive entry point.
Password Required
Click to unlock unified workflow
Key Differences
- •Three separate entry points
- •Different patterns for each feature
- •Users had to hunt for features
- •Intimidating for business users
- •Single unified entry point
- •Consistent patterns everywhere
- •All DSML in one place
- •Approachable for all users
D.E.S.I.G.N. Framework Applied
Discover Deeply
Three powerful DSML tools existed — but nobody could find them. The problem wasn't quality; it was visibility.
Empathize with the Ecosystem
Data scientists needed depth. Business users needed simplicity. One entry point had to serve both.
Simplify the Chaos
Three scattered entry points → integrated into the Hub. No extra space — smart integration inside existing patterns.
Iterate with Inclusion
Mockups first — tickets followed. Architecture defined collaboratively with PM.
Grow Through Constraints
Built within the Hub ecosystem. No new infrastructure — smart integration with existing patterns.
Navigate Forward
NLQ and Insights are live in production. The DSML Hub ships in 2027.
Discover Deeply: From Roadmap Concept to Strategic Definition
Three powerful features existed — but nobody could find them. Customer feedback confirmed: visibility was the problem.
- •NLQ — Natural Language Query: ask questions against your data in plain English ("Which product sold most last quarter?")
- •Insights — Auto-generates visualizations and patterns from your dataset
- •ML — Train predictive models: classification, regression, time-series forecasting, anomaly detection
Empathize with the Ecosystem: Bridging the Gap Between Personas
Data scientists needed depth. Business users needed simplicity. One entry point had to serve both.
- •Data scientists wanted depth, control, and advanced configuration options
- •Business users wanted simplicity, guidance, and quick answers without technical complexity
Simplify the Chaos: Mapping and Unifying the System
Integrated three scattered entry points into the Hub. Already owned all three — integration felt natural.
- •Smart integration into the Hub — no extra space, everything inside
- •One DSML Hub as the single entry point
- •Consistent interaction patterns across NLQ, Insights, and ML
- •Fully responsive — works on any screen size
- •No new infrastructure; built within existing Hub ecosystem
The Idea Lab
3 products in 1 plugin. From napkin sketches to high-fidelity wireframes, mapping out how Natural Language, Insights, and Analysis would coexist.
Rapid Board Ideation
System Architecture Flow
Low-Fidelity Wireframes
Initial Concept Mockup
Hub Integration Spec
Architecture Blueprint
Detailed artifacts and sensitive diagrams are available for authorized reviewers.
Design Principles Applied
Iterate with Inclusion: The Evolution of a Unified Hub
Same features, transformed experience. Three separate tools became one cohesive hub.
Design Evolution
Detailed artifacts and sensitive diagrams are available for authorized reviewers.
Empty States
Detailed artifacts and sensitive diagrams are available for authorized reviewers.
Protected Content
Detailed artifacts and sensitive diagrams are available for authorized reviewers.
Grow Through Constraints: Strategic Ownership at Scale
Fought for features, lost some battles. Built within constraints, enabled two designers to continue.
Architecture Blueprint
Detailed artifacts and sensitive diagrams are available for authorized reviewers.
Navigate Forward: Adoption, Impact, and Future
NLQ adoption +25%. Insights live. DSML Hub shipping 2027.
- •NLQ with +25% adoption increase
- •Insights with auto-generated visualizations
- •ML Functions redesign → 2026
- •Unified DSML Hub → 2027
Workflow Comparison
Detailed artifacts and sensitive diagrams are available for authorized reviewers.
Protected Content
Detailed artifacts and sensitive diagrams are available for authorized reviewers.
Explore Other Architectures
Interested in working together?