EQUIPMENT DATA INFRASTRUCTURE

EQUIPMENT DATA INFRASTRUCTURE

Redesigning XOi’s equipment data foundation to prevent unreliable records from affecting downstream products

Overview

I led the end-to-end design strategy across technician capture, admin operations, and downstream planning workflows

Overview

I led the end-to-end design strategy across technician capture, admin operations, and downstream planning workflows

My role and team

Design strategy, UI, UX and product research. Team included Data scientist team, product manager and design manager

My role and team

Design strategy, UI, UX and product research. Team included Data scientist team, product manager and design manager

Timeline

3 months, 2025

Timeline

3 months, 2025

I created a shared data-quality system that prevented bad records from entering the platform and helped administrators repair existing data

  • 9:41

    Site Survey

    Finish workflow

    Provide a data plate Image

    Capture with your camera or upload an image

    Check for a match before adding new equipment.

    A duplicate can impact lifecycle history and

    maintenance recommendations

    Existing equipment at this site

    RTU-01

    View Details

    Serial number

    CX35-30/36B-6F-20

    Model number

    1519K09200

    Age

    6yrs

    Make

    trane

    Inactive equipment (1)

    RTU-05

    Replaced by RTU-03

    CX35-30/36B-6F-20

    Add new equipment

    My work

    Knowledge

    Search

    Settings

  • 9:41

    RTU-05

    ⚠️

    Replaced unit

    This unit was replaced on Jan 20, 2025 and is no longer active. Continue work on RTU-03

    Replace unit with RTU-03

    RTU-05

    Replaced by RTU-03

    Serial number

    CX35-30/36B-6F-20

    Model number

    1519K09200

    Age

    6yrs

    Make

    trane

    Last activity

    Last scanned Jan 20, 2025 · 11:43

    Scanned by D. Okonkwo

    Something is wrong with the equipment?

    Report a problem to admin

    My work

    Knowledge

    Search

    Settings

  • 9:41

    Duplicate equipment found

    We found an existing equipment record. Creating another record may result in duplicate lifecycle history and recommendations

    Compare with existing equipment at this site

    Review the match below and choose what to do next

    Equipment captured

    RTU-01

    Serial number

    CX35-30/36B-6F-20

    Model number

    1519K09200

    Age

    6yrs

    Make

    trane

    vs

    Existing equipment

    RTU-01

    Serial number

    CX35-30/36B-6F-20

    Model number

    1519K09200

    Age

    6yrs

    Make

    trane

    Use RTU-01

    Add as new equipment

  • 9:41

    Add Equipment

    Add unit names to continue

    Unit Name

    *

    e.g RTU1

    Serial number

    CX35-30/36B-6F-20

    Model number

    1519K09200

    Age

    6yrs

    Make

    trane

    Location

    Rooftop

    Replaced By

    RTU-03

    Save

    My work

    Knowledge

    Search

    Settings

Why we started redesigning equipment
data foundation?

Nearly half of the equipment data entering

the platform bypassed meaningful validation

40%

Technician OCR capture

40%

Technician OCR capture

50%

Invalidated CSV imports

50%

Invalidated CSV imports

Manual entry

Inconsistent data entry

Manual entry

Inconsistent data entry

One unreliable equipment record could distort maintenance
recommendations and capital plans

Duplicate equipment captured could lead to an inaccurate downstream recommendation loosing admins trust

Duplicate equipment added into the maintenance quote

Incorrect quote being generated for the maintenance with duplicate equipment

Before: technicians could not reliably distinguish active, replaced, and duplicate equipment

Lack of visibility of existing equipment

Equipment added into the system without validation

The problem extended across three connected workflows, not just OCR

I led cross functional meetings with customer success, product manager and data scientists to join me in evaluating the platform

The problem was not limited to OCR

Bad data entered through three connected workflows

  1. Lifecycle planning workflow

Failure

Downstream reliance on an unverified asset database

Problem

Flawed calculations lead to unreliable replacement plans

Impact

Drastically reduces user trust and feature adoption, as users cannot confidently rely on the replacement plans

  1. Equipment capture workflow

Failure

Lack of real-time matching

mechanisms during field data entry

Problem

Field technicians unknowingly log existing assets as new installations, driving the duplicate equipment problem

Impact

Creates immediate data pollution at the point of origin, wasting technician field time and forcing redundant asset creation

  1. Admin operations console

Failure

There's no filter at the front door to stop bad data from getting in

Problem

Bulk-uploaded asset sheets bypass strict a CSV validation

Impact

Operations teams bear the administrative burden of manual data cleanup

Admin operations console

Failure

There's no filter at the front door to stop bad data from getting in

Problem

Bulk-uploaded asset sheets bypass strict a CSV validation

Impact

Operations teams bear the administrative burden of manual data cleanup

Lifecycle planning workflow

Failure

Downstream reliance on an unverified asset database

Problem

Flawed calculations lead to unreliable replacement plans

Impact

Drastically reduces user trust and feature adoption, as users cannot confidently rely on the replacement plans

Equipment capture workflow

Failure

Lack of real-time matching

mechanisms during field data entry

Problem

Field technicians unknowingly log existing assets as new installations, driving the duplicate equipment problem

Impact

Creates immediate data pollution at the point of origin, wasting technician field time and forcing redundant asset creation

Mapping the ecosystem

It revealed that isolated feature fixes would not protect the platform

Because we could not redesign the entire platform at once, I proposed a phased data-quality strategy

I organized the work into 3 phases balancing urgency, technical complexity, existing customer workflows, engineering effort, and downstream risks

Phase 1

Reduce immediate duplicate creation

Phase 1

Reduce immediate duplicate creation

Phase 2

Improve capture quality and administrative review

Phase 2

Improve capture quality and administrative review

Phase 3

Establish a durable data-quality foundation

Phase 3

Establish a durable data-quality foundation

Instead of blocking possible duplicates, I gave technicians enough context to make an informed decision

Automatically blocking all suspected duplicates could create false positives and prevent technicians from adding legitimate new equipment

9:41

RTU-05

⚠️

Replaced unit

This unit was replaced on Jan 20, 2025 and is no longer active. Continue work on RTU-03

Replace unit with RTU-03

RTU-05

Replaced by RTU-03

Serial number

CX35-30/36B-6F-20

Model number

1519K09200

Age

6yrs

Make

trane

Location

Rooftop

Replaced By

RTU-03

Last activity

Last scanned Jan 20, 2025 · 11:43

Scanned by D. Okonkwo

Something is wrong with the equipment?

Report a problem to admin

My work

Knowledge

Search

Settings

9:41

Duplicate equipment found

We found an existing equipment record. Creating another record may result in duplicate lifecycle history and recommendations

Compare with existing equipment at this site

Review the match below and choose what to do next

From dataplate

RTU-01

Serial number

CX35-30/36B-6F-20

Model number

1519K09200

Age

6yrs

Make

trane

vs

Existing equipment

RTU-01

Serial number

CX35-30/36B-6F-20

Model number

1519K09200

Age

6yrs

Make

trane

Use RTU-01

Add as new equipment

Automation identified risky records; administrators
retained control over uncertain changes

Automation identified risky records; administrators
retained control over uncertain changes

I designed admin review workflows to catch corrupted data before it propagated the system

Testing revealed that preventing duplicate records also saved technicians time in the field

"I had 15 units to capture. I opened the camera, and seeing the existing equipment right there helped me know what was already in the system and what still needed to be added."

Field technician

Equipment data went from a liability to a
foundation the product could build on

24%

User impact

Expected outcome

Technicians could see what already existed, avoid unnecessary recapture, and make more informed decisions

24%

User impact

Expected outcome

Technicians could see what already existed, avoid unnecessary recapture, and make more informed decisions

40%

Operational impact

Expected outcome

Administrators received safer tools for identifying, reviewing, and correcting questionable records

40%

Operational impact

Expected outcome

Administrators received safer tools for identifying, reviewing, and correcting questionable records

30%

Business impact

Measured outcome

The work supported customer confidence, downstream product adoption, renewals, and XOi’s broader platform direction

30%

Business impact

Measured outcome

The work supported customer confidence, downstream product adoption, renewals, and XOi’s broader platform direction

The high-fidelity screens shown here are part of a personal practice exercise. After the initial project shipped, I chose to redesign the entire interface on my own to explore how a refreshed design system and improved visual hierarchy could elevate the final product. Detailed case study is liked here

rabia razzaq

Twitter

LinkedIn

Reach out at

LinkedIn

Twitter

Email

Contact