Unified provisioning to GCP

---

overview

Role

Lead Product Designer

Tools

Figma & Figma AI

Miro & Miro AI

Confluence

Jira

Copilot

Company

CVS Health

Duration

2 months,
Jun 25 - Aug 25

---

problem

Provisioning workflows across CVS Health were fragmented, inconsistent, and riddled with manual exceptions. Teams relied on multiple disconnected flows, undocumented steps, and shadow IT workarounds.

Divisional architects—responsible for spinning up 18–20 projects per week—faced repetitive data entry, approval bottlenecks, and unclear prerequisites. Each project required duplicative translation of business requirements into technical specs, often re-justified in bi-weekly VP calls.

The result was wasted time, inconsistent governance, and infrastructure friction that slowed innovation and undermined confidence in the platform.

Lean UX Canvas

---

goal

Streamline the provisioning flow within our frontend interface - Data Portal - by:

• Automating work and reducing approval bottlenecks

• Clarifying prerequisites and eliminating redundant steps

• Ensuring compliance and governance through embedded workflows

• Delivering a consistent, transparent, and scalable provisioning experience for architects

---

OUTCOME

This initiative, delivered in only two months, resulted in an effective and quick unified provisioning flow that accelerated approvals, improved compliance, and boosted user satisfaction. By combining high-fidelity mockups, service blueprinting, and iterative testing, the project:

• Reduced provisioning cycle time by 33%

• Increased governance audit coverage by 35 points

• Improved user satisfaction by 7 points

• Established a scalable foundation for future onboarding and application flows

---

user archetype

The original user archetype, derived from the Enterprise Data User Archetypes study was based on the Technical Data Owner archetype. And required multiple user interviews with highly technical Data Architects, responsible for building and approving the pipelines of all our databases at CVS Health.

technical data owners / DATA ARCHITECTS

Data Literacy: High (on a scale Low-Medium-High)

Goals

• Deliver infrastructure quickly and consistently.

• Minimize manual overhead while ensuring compliance.

• Maintain visibility and traceability across workflows.

Behaviors

• Translate business needs into scalable, compliant infrastructure.

• Coordinate with governance and engineering teams.

• Manage multiple concurrent projects using automation and custom flows.

Common Roles

• Data Architects

• Distinguished Engineers

Pain Points

• Repetitive data entry across fragmented tools.

• Approval bottlenecks and unclear prerequisites.

• Difficulty maintaining consistency across dozens of projects.

---

PROCESS

Discovery & ALIGNMENT

Activities

1. Partnered with the VP of Architecture, Governance Director, Product Head, and Technical Writing Lead to define goals and constraints

2. Reviewed six months of prior UX research on onboarding and provisioning

3. Conducted four iterative rounds of interviews and group testing with divisional architects

Learning

1. In-house jargon and unclear prerequisites caused confusion

2. Lack of process transparency led to delays and rework

3. Architects needed a single source of truth to manage dozens of concurrent projects

OUTCOMES

1. Clear articulation of user pain points and mental models

2. Alignment with the need for a unified, transparent provisioning flow

IDENTIFIED POINTS FOR IMPROVEMENT

1. Needed to simplify language and clarify expectations

2. Required embedded documentation and Q&A to reduce reliance on external calls

Form Prerequisites Wireframe

Data Portal form step titled Form Pre-Requisites listing required approvals and checkboxes before starting GCP project provisioning

Application Info Page Wireframe

Data Portal form step titled Application Info for provisioning a GCP project, with fields for app ID, cost center, and owner details

Network Requirements Wireframe

Data Portal form step titled Network Requirements showing configuration options and checkboxes for disaster recovery, sources, and sinks

Tools & Permission

Data Portal form step titled Tools & Permissions with input fields for setting up Active Directory groups for engineers, admins, and readers

Overview Page Wireframes

Data Portal form step titled Overview summarizing entered details for application info, project environment, permissions, and network

design & validation

Activities

1. Delivered polished mockups using the Data Portal design system

2. Co-led a service-blueprint workshop mapping frontstage, backstage, and support flows

3. Validated mockups through rapid user interviews and stakeholder reviews

4. Leveraged AI to decode terminology, refine form copy, and recommend new Documentation Center pages

Learning

1. Automated pipelines could eliminate manual handoffs and accelerate approvals

2. Live dashboards surfaced bottlenecks in real time

3. Embedded Q&A hubs and proactive notifications reduced abandonment and improved satisfaction

OUTCOMES

1. Automated pipelines cut approval time by 2 days

2. Live dashboards reduced total cycle time by 5 days

3. Proactive notifications boosted satisfaction by 7 points

4. Audit trails and structured rejection comments improved governance coverage

IDENTIFIED POINTS FOR IMPROVEMENT

1. Needed to expand automation to the application and user onboarding

2. Required ongoing refinement of dashboards and SME routing to reduce SLA breaches further

Form Pre-Requisites High-fidelity

Data Portal self-service page titled Form Pre-Requisites listing requirements, approval documentation, and confirmation checkboxes

Application Info High-fidelity

Data Portal form step titled Application Info with fields for CVS App ID and cost center, part of GCP project provisioning workflow

Project & Environment High-fidelity

Data Portal form step titled Application Info with fields for CVS App ID and cost center, part of GCP project provisioning workflow

Tools & Permission High-fidelity

Data Portal form step titled Tools & Permissions with input fields for assigning Active Directory groups to manage project access.

Network Requirements High-Fidelity

Data Portal form step titled Network Requirements showing configuration options and checkboxes for disaster recovery, sources, and sinks

Overview High-Fidelity

Data Portal form step titled Overview summarizing application, environment, permissions, and cloud presence details in a six-step workflow

---

key Metrics

Objectives

• Cut total provisioning cycle time

• Improve user satisfaction and reduce form abandonment

• Increase governance coverage and auditability

• Reduce SLA breaches and approval delays

Key Metrics

• Total cycle time reduced from 15 → 10 business days.

• Architecture approval time reduced from 7 → 5 business days.

• SLA breach rate reduced from 18% → 14%.

• User satisfaction score increased from 68% → 75%.

• Form abandonment rate reduced from 22% → 19%.

• Audit trail coverage increased from 45% → 80%.

---

outputs

• High-fidelity mockups of unified provisioning flow

• Service blueprint mapping frontstage, backstage, and support processes

• Embedded Q&A hub and proactive notification system

• Automated pipelines, live dashboards, and lineage-based SME routing

• Governance artifacts: audit trails and structured rejection comments

---

Thank you for your time.