Visit on Desktop
Not Available on Phone
CONFIDENCE UX
Risky drilling ops,
now a beacon of
confidence
Safely drill for oil. Orchestrates globally dispersed teams making hundreds of high-stakes environmental and billion dollar economic decisions. I made risk legible at the moment of the decision.
Confidence UX for Halliburton
Modern Design Drilling Ops
Enterprise UX
Oil & Gas
system Design
Bolt on AI
0 to 1
40+
risks & hesitation
signals reduced or eliminated
Fragmented tools ⟶ single source of truth
DOS report cycle ⟶ months to hours
Projected $22 M / yr cost reduction
Lead Product Designer
2025 - Present
My Contribution
Owned design from research to delivery
Established object model, system and interactive design
Architected UI across 6 surfaces
Research and strategic sprint facilitation to create insights and clarity
Owned design decisions and partnered with product and delivery on roadmap
Led handoff and interaction with engineering & PM
Millions $
per run
Millions $
per run
at stake
risk of environmental devistation
risk of injury and loss off life
strong industry competition
Background Story
Why it mattered
A bottom-hole assembly is a string of tools run miles into the earth to strike oil reservoirs. Meer minutes of non-production time, hesitant decision making and human error were million dollar threats to the ops pipeline, however project ops was being performed by manually stitching together spreadsheets, PDFs and tribal memory.
I was brought in to collaborate with a project team to design a unified system and workflow that to reduce the workflow time to hours.
North star
Reduce workflow from month to hours
Problem • What I heard
“
Today I estimate we need 50% additional headcount just because of inefficiency in software and process.
— Halliburton Drilling Engineer
My take
These were seasoned engineers, fear of the tool was less a hesitation point than the weight of being responsible for errors and disaster and the labor of crude collaboration.
5
Separate Apps
No data sync. Redundant user tasks. Heavy context switch.
Market Pressure
Competitors were automating Halliburton needed a leap forward to remain competitive.
Limited Scale
Legacy platform siloed from other applications, regions or internal orgs outside of drilling.
Low
Productivity
Work lost due to system crashes. Manual reports creation cost 80% production time/wk.
100+
Complex Tasks
The workflow involved 15-20 individual contributors, distributed globally.
High
Risk & Regulatory
Mistakes in planning can cost time, damage equipment, create legal threat.
Approach
Ambiguity between product team and design stalled the project.
Harsh industry domain knowledge
I required learning deep of industry positioning and contextual details about tasks, team roles and common challenges
Solution bias
Stakeholder resistance from legacy UX mistaken for requirements; Users also built expectations around legacy tool pattern
Evolving direction
The ambiguous part of the north star was to make a modern experience; Design workflows also changed as understanding evolved
My strategic leadership added clarity that got the team unstuck
Role
Interview/Testing Facilitator
Discovery lead
UX storyteller
UX roadmap Owner
Designed & build measurement plan
Deliverable stack
1.UX Deliverables
2.UI Deliverables
3.Handoff Artifacts
4.System Priorities
5.Design Feature
6.Product Feature
Discovery Outcomes
50+
Touchpoints
4
User Roles
12
Key Screens
100+
Required tasks
7 apps; each step involves switched between apps
Serves 20 Countries/ 5 global regions
13 collaborators
8 Regulated checkpoints
Legacy System Profile
Complex non-linear app suite
Future System Profile
Unified collaborative platform
Need modern feel and forward-thinking interactions
Need reduce time to complete
Design Implementation
Three moves that I made to build system
Confidence
Role
UI Designer
Design Director
UX & Interaction Strategy
AI-Generated erodes design authority
the Push
A design lead proposed polished AI-generated screens days before sprint close. Accepting the work as is would have been faster and avoided debate.
My Decision
I challenged the design in ways the team hadn't accounted for: wholistic system impacts, tradeoffs and how frequency changes user expectations. I led the team toward a workflow grounded in system logic.
Demonstrated
✓ Design Judgment
✓ Systems Thinking
✓ Cross-functional Influence
✓ AI Generated Design Review
Balancing competing goals with taste
the Push
Development lead optimized performance. Product pushed for a modern experience. The initial UX direction favored a single interaction pattern.
My Decision
I introduced context. I challenged the assumption that consistency always improves usability and I weighed interaction patterns to task frequency and consequence.
Demonstrated
✓ Exception by design
✓ Trade-off resolution
✓ Risk-calibrated interaction design
Define ambiguous requests
the Push
Product asked for a "confident yet modern" feel with no measurable direction. The tension was design toward a mood, not a spec.
My Decision
i turned a vague aesthetic ask into a structural system. i cut friction by reduce data-entry effort across the product. I made progress personal and rewarding, not just a status bar.
Demonstrated
✓ Motivation-driven interaction design
✓ Translate ambiguity to system
✓ Design to unify system
✓ Apply Confidence UX
System thinking transforms decision patterns
Page Banner
Semantic:
Page Banner
Tab Bar
Semantic:
Tab
DETAIL
System Status
Page Banner Function: User can identify or change project from single screen
Well Section Tab Bar: Manage varying project complexity
2➞1
App consolidation
↑usage
Stay on page
DETAIL
Version Control
DETAIL
User Decision
DETAIL
Reduce Error
Challenging system design - BHA Tool
$M's per run at stake but risk + complexity stalls
Building a BHA tool was a technical and non-linear workflow
- 6 steps completed by up to 7 uses in tandem.
Building a BHA tool was a very
technical and non-linear workflow
- 6 steps workflow completed by up to 7 uses in tandem.
Drilling Tool
BHA Tool
Add Tool
components
Logistics
Customize
Preview Tool
Analyze
Measurements
Load Project
Cognitive Load
Workarounds
Fragmentation
Complex
I designed learnable a entry experience
User-friendly micro-copy, interactive motion on triggers and intuitive decision points enabled new users to easily navigate the screen.
Eliminated hesitation
17➞5
input fields
Cognitive Load
Cognitive Load
Workarounds
Fragmentation
Complex
I designed learnable a entry experience
User-friendly micro-copy, interactive motion on triggers and intuitive decision points enabled new users to easily navigate the screen.
Eliminated hesitation
17➞5
input fields
Cognitive Load
AI Ready is modern design
AI ready means a designer has adopted AI into their design workflow through giving direction and contributing, but also a designer can seamless design AI patterns into systems.
01 Approach to Design Workflow with AI
3 strategies
Design Workflow
intent-based Design stack
Agentic
Setup
5 steps from idea to handoff
hybrid design stack 3 day to prototype
Design Thinking & expertise
🏞️
Use SCAMPER to explore designs for an AI copilot screen.
Precise prompts reduce tokens use
design-led prompts reduced token waste
Led Design with Taste & Rigor
⟶
Module
a
Module
b
Module
C
Scalable interaction patterns
critique-led decisions
I didn’t just bolt on AI.
I designed for the hesitation around using it.
2 patterns
AI CoPilot
02 Building AI Design patterns
02 Building AI Design patterns
Objective
User needs to identify risks & challenges to the drilling job
Reduce risk of error
Usage ↑ - no need to leave screen
Update legacy to modern feel
Risk Visibility
Inline AI in the Knowledge panel
Risk Score
Engineers are rightly to be skeptical of AI suggestions. I made the risk visible on the screen - severity grade, frequency counted, rationale expanded. No need to open spreadsheets to calculate.
Engineers are rightly to be skeptical of AI suggestions. I made the risk visible on the screen - severity grade, frequency counted, rationale expanded. No need to open spreadsheets to calculate.
High
Lost circulation — depleted fields
82% Confidence
Found in 7/7 Wells
LEM-2543
|
EQWR-WX-003
+5 Wells
Rationale
Hover over Rationale button to preview AI reasoning
Severity tag - color coded red [High], yellow[Medium], green[Low] by risk level
Severity tag - color coded by risk level
red [High], yellow[Medium], green[Low]
Frequency - count how often this risk occurs across other wells
AI Confidence Score - the AI model reports its own certainty
Expand Rationale - reveals the rationale behind each risk severity tag
Restratint
Inline AI in the Knowledge panel
Deliberate control of AI
Because constant screen refresh creates distraction and stacks up token costs, AI runs only when a button is clicked. Engineer stays in control.
Review
Initiate AI suggestions
More Results
Adjust filter get more results
Refresh
3
Add AI new suggestions
Reversible Decision
Inline AI in the Knowledge panel
Frictionless control
Because fear of irreversible actions stalls use, every AI action is undoable and the view stays user-editable
Edit Risk content
Undo Inserted Risk
03 The Result
Confidence Signals
~ 80%
projected number of DOS tasks
to reach confident assisted decision
users reach confident assisted
decision on DOS tasks
Pending Measurement
Legacy App - No AI
Modern - With AI
User Confidence
Workarounds;
low trust in tool
The risk grade
builds trust
Decision Quality
High Variance by
user experience
Score standardizes
decision making
Tools Required
Abandon app;
switching tools required
Single app;
reduced switching
Approach to Design Workflow with AI
3 strategies for Design workflow with AI
intent-based Design
5 steps from idea to handoff
Agentic Setup
Design Thinking
Prompts reduce tokens use
🏞️
Use SCAMPER to explore designs for an AI copilot screen.
Lead with Taste & Rigor
Patterns speed up shipping
Module
a
Module
b
Module
C
Summary
Designed Confidence with end-to-end leadership
Confidence UX domonstrated by end-to-end design leadership
before
After
Single modern cloud platform + global shared intelligence + live collaboration
Scalable confident decisions.
Alignment Weeks → Hours
Report published in 3 clicks
Modern
Look & Feel
Look & Feel
Modern
Personalized
Workflow
Workflow
Personalized
Expert Approach
Design Thinking
Platform
Single System
40+
risks & hesitation
signals reduced or eliminated
Outcomes
Established user trust → reduced resistance to change
Established user trust →
reduced resistance to change
Sigals
Decision quality - established standard for understanding AI results
Eliminated spreadsheet workaround - reduced context switch
Risk Visibility pattern
high
Lost circulation
82% Confidence
Found in 7/7 wells • [Rationale]
Made risk legible with 3 layer pattern
restraint
AI Copiolt
Autosave
Review
refresh
AI
EDit
Ready
Completed
user • Date • Time
Key decision about friction across patterns
Key decision about friction across patterns
Human controls → reduced errors, task time, costs
Human controls →
reduced errors, completion time, costs
Sigals
Conserve costs from AI token usage
Design consulted on tradeoff decisions - more clicks vs frictionless workflow
Reduced error - user clarity on versions, AI decisions reversible, fewer rework
Scalable design → boost team confidence and alignment
Scalable design →
boosted team confidence and alignment
Sigals
Turned unmeasured request into a principled north star
Tasteful and adaptive decision making when team had got stuck, errored or was uncertain
Owned design system - repeatable patterns, wholistic system
Navigated Ambiguity in key moments
Owned design direction
Owned design direction
Discovery alignment agent
Discovery alignment agent
Discovery
Prompts using Design Thinking
Prompts using Design Thinking
Design
DS imagery & motion components
DS imagery & motion components
Design
Extend DS with 2 AI Patterns
Extend DS with 2 AI Patterns
Design
Adapt to AI product team workflow
Adapt to AI product team workflow
Deploy
ImplementedTesting & metrics
Implemented testing & metrics
Deploy
Systems thinking
1
5 apps →
1 screen
2
Muliti-user
Collaboration
3
Role-based/
Personalization
4
Flow: Weeks → Hours
Efficient system → focus quality and collaboration increased
Efficient system →
focused screens, collaboration, single source
Sigals
Cognitive load: Multiple application + multiple screens -> All-in-one system
Reduce complexity: Multiple application + multiple screens -> All-in-one system
Standard output and consistent knowledge share in under
Clearer screens for non-linear workflow 17 -> 5 steps
Globally multiple users use same screen live simultaneous; cloud file managment