expense management
Company
Role
Sr. Product Designer
medium
SaaS Platform
timeline
2020-2022
Overview
Plate IQ is a SaaS platform dedicated to automating accounts payable tasks, primarily serving the restaurant and hospitality industries. Its core product streamlines back-office operations, allowing customers to upload invoices by either forwarding them to a dedicated email address or uploading images directly to Plate IQ for digitization. Once uploaded, customers can sync these invoices with their accounting systems according to their preferences and set up automation flows, significantly reducing the time required for accountants and bookkeepers.
Spend Management (Spend) is a new expense management product from Plate IQ that facilitates petty cash management and other business expenses by linking the company’s bank account as a funding source. Spend enables customers to create both physical and virtual cards for their employees, ensuring seamless integration with their accounting software.
Challenge
This was a brand new product that began near the start of the Covid-19 pandemic in April 2020, so including the challenges of a pandemic, we also had a goal of releasing our V1 by Labor Day weekend (early Sept. 2020) in a couple of months. With a short timeline, this meant focusing on the core functionality only. We also had to provide card artwork and simple wireframes to our banking partner (Marqeta) to fulfill legal requirements with a 6-8 week lead time.
We had several key goals of:
Providing an easy way for users to create and request virtual and physical cards for company uses
Give users the ability to customize and edit card controls
Sync with users’ accounting software just like the main product
Gain new customers with Spend who are not familiar with Plate IQ
Post launch I would also face the challenges of having frequent changes on product management and having to adjust feature plans and goals.
Users
The main customer base for Plate IQ were in the restaurant & hospitality industry, which ranged from main chain restaurants to local mom & pop shops. There also were various amount of other companies ranging from accounting houses and retail chains too.
Employees from those companies ranging from owners, CFOs, and in-house accountants we’re the main users logging onto Plate IQ at least a few times a week.
Role
Like many startups, I wore multiple hats as the sole designer when I first started. I worked closely with a small team—one product manager (PM) and four engineers—as we pushed toward our V1. Being the only designer at Plate IQ also meant balancing marketing and visual work for Spend alongside product design.
My responsibilities included:
Designing marketing materials and card designs for the product, adhering to regulations
Creating wireframes for each workflow and building simple prototypes
Collaborating with the PM on copy and product experience improvements
Delivering designs to engineering, tracking progress, and analyzing user data
Conducting competitive and user research post-launch
Example wireframes and explorations
Product Audit
With limited time for design and testing, my PM and I agreed to prioritize the main user flows for the product. We also had to meet the deadline for delivering a draft of these flows, along with the virtual and physical card designs, to our banking partner for approval. Our focus was narrowed down to:
Card creation and requests
Editing/canceling cards
Viewing transactions (including sorting, filtering, and search)
User and company settings
Our goal for V1 was to onboard existing Plate IQ customers. As a result, we reused several flows, such as onboarding, accounting setup, and company details, from our core product. We concentrated only on the settings specific to Spend, like connecting a bank account for funding. For V1, we developed basic versions of two user personas: the admin and the employee (who can request and use cards but not edit).
After V1, I collaborated closely with the PM to define the next flows, focusing on expanding accounting features, implementing receipt capture, and integrating Spend with our mobile app.
Competitor
Research
In the beginning of this project, I explored a wide range of solutions, from established credit card experiences like AMEX and Discover to newer players targeting tech startups, such as Brex, Ramp, and Divvy. Fortunately, Plate IQ was already a Brex customer, allowing me to dive into their product, understand their workflows, examine features, and talk to coworkers who regularly used Brex cards about their experiences.
After our V1 launch, I collaborated with colleagues in marketing to review demos from competitors whenever possible. I also gained valuable insights from walkthroughs of existing solutions, like SAP Concur, shared by current customers who provided regular feedback.
Wireframe explorations for V1 launch
User
Research
At the start of our process, we didn’t have much time to validate solutions before the V1 release, so most of the user research and testing happened post-launch. Fortunately, our team used Spend internally after launch, which allowed me to gather immediate feedback from coworkers during development. This also gave me the opportunity to work closely with our in-house accountant, whose role aligned with our main persona (the admin), to understand what worked well and what needed improvement.
After V1, I proactively reached out to current Spend customers for feedback and to test new workflows. I also set up bi-weekly to monthly meetings with the heads of our customer support team, which provided a valuable catch-all for both positive feedback and areas that needed improvement. As mentioned during the competitive research phase, some customers even walked me through other solutions they were using, giving me a clearer understanding of their needs and pain points beyond simple conversations.
V1 Release & Onward
After a couple of rounds of changes, shifting priorities to address new customer needs, onboarding a new product manager, and implementing another UI update, we completed the initial launch of Spend Management in the first week of September. We began with 5 customers and grew to 20 by month’s end. Our approach emphasized rapid learning and iteration based on customer feedback, while also expanding our feature set to better address key customer workflows.
Next, we focused on adding features with quick implementation potential, particularly around settings. Features like policy management, alert setup, and custom user permissions were prioritized, as they leveraged existing tech stacks from our core product, making the experience familiar for our customers. This strategy allowed us to make steady improvements while giving product and design time to consider the next set of features for Spend.
With the quick wins behind us, our next focus shifted to aligning with key features that current customers loved in our core product but needed tailored adjustments for expense management. Customers frequently requested expanded accounting software integrations (including transaction coding), receipt upload and matching capabilities, and automatic syncing with more accounting platforms.
Some solutions, like coding transactions for accounting, were straightforward to adapt for Spend. Others, like automated invoice matching, required more complex engineering. We wanted to offer a similar experience in Spend, allowing receipts to be automatically matched to transactions even if uploaded separately. However, without the capacity to build this fully, we needed a bridge solution in the interim.
Step 1: User selects an uploaded image, Step 2: User sees list of transactions to match receipt to, Step 3: User selects transaction to tie receipt too.
Receipt Capture
After analyzing the receipt capture landscape and identifying common patterns and flows, my team and I developed an easy-to-implement, scalable solution to streamline the process for customers.
Previously, users could manually upload images to transactions in our web app. We also revamped our mobile app to include push notifications, prompting users to photograph receipts immediately after swiping their card. Still, customers faced scenarios where employees didn’t upload receipts immediately, often due to app access issues or simple forgetfulness. This left admins and bookkeepers with the task of manually matching receipts to transactions, which could take hours. Our goal was to reduce this time by at least 25%.
Our solution allowed users to email receipt images to a dedicated company inbox, where an admin could view each image along with upload details, such as timestamp and sender. Next to each receipt, the admin could see a list of recent transactions without receipts. They could then select the matching transaction and apply the appropriate accounting codes. We tested this prototype with high-volume customers to validate the approach and moved swiftly into development based on positive feedback.
Keep Going
Over the next year, we added key functionality and features, including a reimbursement system, cashback rewards, and several other enhancements:
Reimbursements (currently supports single-use and mileage reimbursements)
Transaction exporting (backend improvements)
Linking external company cards to the platform via Plaid or CSV import (now discontinued)
Custom alerts for failed transactions, spending limits, and more
Exporting transactions directly to our invoice management product
Automatic syncing and mapping of transactions with customers' accounting solutions
Cashback rewards on all purchases (1-3%)
Multi-location support for single accounts (ideal for businesses with multiple sites)
These features aligned us more closely with industry standards, setting us up to tackle the next challenge: expanding our customer base. We collaborated with Marketing, who was already executing a company-wide initiative to promote all products as part of a comprehensive engagement and adoption push.
Outcomes
Between March 2021 to Nov 2021, our top customers averaged $80k-$300k each month of usage, majority coming from virtual card use. Our customers ranged from regional coffee chains, internet ad companies, hospitality, and various restaurant groups large and small.
Settings and Reports screens. These expanded further after the initial launch
Looking Back
There’s a lot I’m proud of with this product, but I also feel there were missed opportunities, many of which were tied to short time frames, an unstable roadmap, and unnecessary stress. From a product and design perspective, we could have improved our market research and done a better job of mapping customer needs. We were quick to respond to customer requests, but this often hurt us in the long run. At times, I didn’t have enough time or resources to properly design features, which led to recurring issues down the road. This was a challenge not only for product and design but also for the engineering team, especially when it came to back-end fixes. Often, the team felt too rushed to think things through thoroughly, resulting in headaches for customer service later on.
Despite the challenges, there were many successes with Spend. One standout was the launch of our first design system at Plate IQ. I designed Spend to be more modern, clean, and visually appealing than our core invoice management platform, which received rave reviews from both internal teams and customers. It also led to the implementation of better design-to-engineering processes within the company—something that didn’t exist before I joined. The engineering team felt comfortable reaching out to me with questions, concerns, or for help, which fostered better collaboration. Spend also introduced new design patterns, like the “side car” for Card and Transaction Details, which became popular in other parts of our product suite.
On a personal note, I’ve gained a much deeper understanding of credit card systems than I ever expected. This kind of knowledge became be valuable in the future. Spend is also a product I take pride in, especially since it was my main project during the height of the pandemic. In a way, it’s my pandemic baby, particularly given that platforms of this size often have entire design teams, whereas this was primarily my own.
What I Would
Do Next Time...
No product is built without challenges; there’s always a need to pivot and adapt. Looking back, several key areas stand out that I would approach differently next time.
For building a product from the ground up, I would prioritize running a series of workshops and conducting more research at the outset. Workshops would bring stakeholders together to identify essential features, define the problem, clarify what the product should and shouldn’t include, align on target users, value propositions, and both business and UX strategies. Having these elements established upfront—rather than piecing them together as we went—would have led to stronger alignment among stakeholders and a clearer trajectory. It’s one thing to pivot in response to new evidence; it’s another to change course due to minor opinions without supporting data.
In terms of research, my initial competitive research was solid, but our user research could have been more robust. Being able to speak directly with customers and sitting in on meetings where the founders pitched Spend to current and prospective customers would have been incredibly valuable. This would have given me early insights, helped validate key features, and strengthened our value proposition. It could have also provided foundational support for the roadmap, giving design a stronger role from the start. Post-launch, a more consistent schedule for gathering feedback from customers or support staff would have been beneficial, rather than addressing it on an ad-hoc basis.
Finally, while I did my best to stay organized under time constraints, I would establish strong file organization practices from the beginning rather than trying to catch up along the way. This was a key lesson I’ve since applied to projects after Spend. Sometimes, it takes navigating a bit of chaos to truly appreciate the value of those “small” organizational details.
