Featured Project
Designing Teller: Amazons proprietary AP Solution
A global financial intelligence tool for amazon's analysts
Non Disclosure Agreement (NDA):
I am unable to disclose certain aspects of this project. Key details are omitted or substituted to preserve the IP and adhere to the agreements stipulations. The process, challenges, events, and other permissible details otherwise accurately reflect my part in the project.
1. Situation
Context & Challenge
Across its many business entities, Amazon processes more than 20 billion financial transactions each month, including accounts receivable, accounts payable, royalties, amortisation, remittances, payments, and cash.
Teller is Amazons internal proprietary accounts payable (AP) micro service which was created to replace Oracles account payable module and better integrate with Amazons Financial Ledger and Accounting Systems Hub (FLASH). Alongside concurrent FinTech initiatives, Teller performs complex and business-critical functions to substantiate the general ledger, and generates financial statements such as balance sheet, cash flow, and income.
When i joined the project, there was no UI, no product definition beyond a mission statement and architectural diagrams demonstrating how the service would co-op already established micro-services. A lofty goal of displacing Oracle AP module, starting with a 10% migration post MVP.
Amazon operates like no other company i've been apart of. With amazon, everyone embraces their 14 leadership principles, demonstration across the board enables teams to be self directed. After finding my footing, the first move upon joining was to build deep domain expertise, to do this i engaged with stakeholders across Japan, Brazil, the US, Europe and Canada to learn their existing process, goals, outcomes, pain-points, needs and build the empathy needed to craft exceptional products.
My Role
I was tasked with helping course correct the inception stages of the project to meet internal standards for continued investment and lead user acceptance testing (UAT) initiatives. I had joined the project team rather early but not quite from the beginning of the initiative. I quickly came to realise there was little to no design or product previously placed on this project - for point of reference the sparse concept artefacts i could track down were presented as embedded screenshots in a .docx.
2. Approach
Research and Gap Analysis
60+ hours of contextual inquiry across Amazon’s global financial teams
Mapped analyst workflows and regional personas
Conducted a feature gap analysis vs. internal micro tools like Flash
This research clarified the MVP scope and created alignment across teams.
Oracle AP Module
Defining the MVP
With a wide range of input, I designed a high-value MVP that prioritised:
Rapid query access for financial performance data
Clear surfacing of key health indicators and flags
Minimal cognitive load, fast execution
Design Execution
I delivered the full front-end experience and recommended a migration to AWS’s Polaris design system. This unlocked rapid development and high consistency.
Leveraged Polaris Design System to expedite UI ideation, developed bespoke Sketch component library with production-ready fidelity for AWS FinTech expansion, including rich table definition
Enabled dev team to execute with precision and speed
Ensured accessibility and responsiveness from day one
3. Solution
I am unable to share this level of detail publicly, but i can certainly describe what was achieved over a call.
4. Results
Outcomes
Co lead UAT initiative including creation of Training the trainer guides, allowing us to measure MVP adoption viability, identify minor functional gaps for resolution.
MVP deployed to 5% of Amazon’s global AP divisions, representing a substantial financial function
Reduced time-to-outcome by up to 60% for key analyst tasks like reconciliation compared to the equivalent Oracle AP workflows.
5. Reflection
What I learned
Trust is earned through understanding, not just UI clarity, but domain fluency. Taking the initiative to become financially litterate enabled me to empathise with, take a data driven approach and ultimately craft a product that is actively saving Amazon millions of dollars annually.
Design systems drive scale, Polaris enabled us to deliver fast without compromising quality and Product definition is a design act meaning designers are product leaders, especially when the product spec is non existent.
Up next: I’m exploring how embedded AI assistants can surface anomaly alerts in real time—further reducing manual oversight and driving another 20% efficiency lift.
Explore more
Let’s Build Tomorrow’s Products
Foster strategic collaboration to design AI-native systems that amplify human capabilities and experience