Khushboo
Portfolio case study / AFL - TELUS

Financial Operations Platform

I was both the project analyst using the workflow and the product designer who saw the automation opportunity: turning Jira, Solomon, email, and 10+ Excel sheets into an internal MVP.

Quote generation
~30m -> ~2m
Budget updates
~2h -> ~30m
Sources
10+ -> 1
Quote Generator interface for AFL TELUS financial operations
Operating role

Why I saw the opportunity: I was inside the workflow

This was not a handed-off design brief. I experienced the lookup burden directly as a project analyst, then used design and working software to turn the pain into product logic.

Manual lookup

NGMR to project code to PO to cost to revenue required repeated copy-paste and reconciliation.

Hidden risk

Missed billables, wrong PO checks, delayed budget updates, and partial status visibility.

Product bet

Centralize relationships first, then automate repetitive work with human review and auditability.

Current state

One project check required 8-12 manual steps

The workflow worked only because experienced team members memorized where to look. The product needed to connect fragmented business logic spread across Jira, Solomon, Excel, and email.

  1. JiraFind NGMR.
  2. ExcelFind project code and related codes.
  3. SolomonPull POs and vendor references.
  4. ExcelVerify costs, revenue, and missed scope.
  5. Email / trackerSubmit updates and status.
Systems thinking

NGMR became the primary project object

01 Data layer

NGMR, project codes, POs, vendor cost, revenue, budget, quote, and change order.

02 Workflow layer

Lookup, verify, compare, identify missed cost, submit to TELUS, approve or reject, and update budget.

03 Risk layer

Missed POs, missed billable work, delayed reporting, revenue gaps, and budget mismatch.

04 Product layer

Search by NGMR, linked project codes, PO validation, cost and revenue view, CO tracking, and budget updates.

Solutions
AI-assisted quote generator interface with design file and editable line items

AI-assisted extraction generated review-ready quotes

AI was applied where it had clear operational value: reading design inputs and preparing quote items for human review. The analyst stayed in control while repetitive material extraction moved faster.

  • ChatGPT
  • Cursor
  • Codex
  • Human review
Change order dashboard with pending, submitted, approved, and rejected states

Change orders became traceable financial objects

The design made missed costs, TELUS submissions, approvals, rejection reasons, and budget impact visible in one place instead of scattering them across updates.

  • Pending
  • Submitted
  • Approved
  • Rejected
Budget dashboard showing planned cost, actual cost, variance, and alerts

Budget updates moved from reconciliation to live visibility

The dashboard connected planned cost, actual cost, POs, vendor breakdowns, and alerts so daily checks could happen from one project view.

  • Planned vs actual
  • Vendor drilldown
  • PO alerts
Stakeholder buy-in

Working prototypes secured the go-ahead

I convinced the manager, program manager, and director by showing risk reduction, speed, and real workflow proof rather than abstract design artifacts.

  1. Show the painMap the manual path and quantify time loss.
  2. Prototype reliefBuild an AI-assisted quote workflow and portal direction.
  3. Demo real examplesWalk leadership through actual project scenarios.
  4. Frame risk reductionPosition the tool around auditability and missed-cost prevention.
  5. Secure go-aheadAlign leadership on MVP use with the immediate team.
Impact

Faster work, lower financial ambiguity, stronger operational control

~30m -> ~2mQuote generation
~2h -> ~30mBudget update workflow
10+ -> 1Source workflow direction
MVP shippedImmediate team before expansion

The MVP created a practical path from manual financial operations to AI-assisted workflow automation, with clear product objects, workflow states, and governance needs for future integrations.

Previous case study