Centsible
Improving the end-to-end expense management process by being bold and creative - leveraging the power of automation and AI.
overview
In a nutshell
Between March and June 2025, I led a six-week group project as part of my UX Design Master's at Loughborough University. Partnered with MHR, a leading HR and payroll solutions provider from Nottingham, our brief was:
"Improve the end-to-end expense management process by being bold and creative - leveraging the power of automation and AI."
Working in an agile team of six, I took on the role of team leader, co-ordinator, and main client communicator. I organised meetings, delegated tasks, and led all client presentations, while also contributing to research synthesis, ideation, and design development to create a user-centred solution within a tight timeframe.
6 weeks
Lead of a 6-person UX Team
Secondary Research, Interviews, Affinity Mapping, Personas, User Journey Maps, Prototyping, User Testing
Pen & Paper, Figjam, Figma, Miro, Microsoft Teams
Group Academic/Client Project
introduction
The Challenge
MHR tasked us with reimagining the expense management process from the ground up. Rather than improving existing software, they encouraged us to create a completely new solution, exploring how AI and automation could transform the experience for users.
We focused on three key user groups:
- Employees who create and submit expense claims
- Managers who review and approve claims
- Finance teams who oversee policy adherence and report on company expenses
How might we improve the end-to-end expense management process by being bold and creative - leveraging the power of automation and AI?
The Approach
We adopted an agile approach throughout this six-week project to ensure continuous progress and alignment with MHR's expectations. Each week, we held:
- Team meetings with ice breaker activities and retrospectives to reflect on achievements, identify blockers, and plan priorities for the next sprint.
- Progress meetings with the client, which I prepared and where I presented our work, gathered feedback, and discussed the best way forward.
Our team, which we named OneDesign, used a range of collaborative tools including Figma, FigJam, Miro, Microsoft Teams, OneNote, and WhatsApp to support remote and in-person collaboration efficiently.
As team leader, I facilitated meetings, maintained our project plan and timeline, and ensured tasks were delegated effectively based on individual strengths. I also put a big focus on good team morale by organising weekly emotional check-ins, two team socials, and peer evaluations to maintain a supportive, motivated working environment.
discover
Secondary Research
To ensure our design decisions were grounded in real user needs, we began with secondary research, which allowed us to build foundational knowledge efficiently by dividing multiple research topics across the team.
Background Research
Understanding the expense management process, from the point at which the expense occurs to the reimbursement, with key stakeholders involved throughout process.
Common Pain Points
We started to see some key problem areas emerge around lack of automation, long reimbursement processes and heavy reliance on paper receipts.
Competitor Analysis
We analysed SAP Concur, Expensify, Zoho Expense, and Rydoo and found them to have cluttered interfaces, steep learning curves, minimal utilisation of AI and a lack of a chat-first approach.
Technology and Trends
We saw potential for artificial intelligence and machine learning to automate the expense management process and to identify spending patterns and provide personalised insights.
Primary Research
The knowledge from our secondary research helped us plan what we wanted to talk about with real target users. Considering the short project timeframe, we opted for semi-structured interviews only as they allowed us to explore predefined topics that built upon our existing knowledge and assumptions in greater detail, making the most out of the limited number of participants and their availability as busy working professionals. We ran eight semi-structured interviews in total, splitting it up across two rounds.
Round 1
4 employees and 1 participant with both managerial and finance team experience.
These interviews validated our secondary research findings and gave us deeper insights into first-hand challenges.
Round 2
3 additional employees.
By this stage, we had begun developing early concepts, so we used these interviews to gather targeted feedback on our initial ideas alongside exploring further user needs.
define
Insight Synthesis
To make sense of all our data from secondary and primary research, we first utilised affinity diagramming to extract key insights from all our different research findings.
Patchwork of tools
Organisations rely on multiple disconnected tools with minimal integration, leading to inefficient, manual processes.
No automation
A lack of automation leads to expense tasks being repetitive, time-consuming, and prone to human error.
Unclear process
Employees often lack clear guidance on company policies, resulting in unclaimed expenses or incorrect submissions that complicate approvals.
Nothing quite fits
Existing solutions are inflexible, have a steep learning curve, and fail to harness AI to its full potential, highlighting an opportunity for a more intelligent, user-friendly alternative.
Defining Our Target Users
From our research, we developed personas representing each target user group: employees, managers, and finance teams. To build empathy and map out pain points clearly, we also created "as-is" user journey maps for each persona, visualising their experiences throughout the current expense management process.
The Employee
The Manager
The Finance Team
develop
Ideation
Building on our research insights, personas, and user journey maps, we began ideating solutions with a strong user-centred foundation.
We ran a rapid ideation sprint as a team, starting by generating as many "How might we?" questions as possible in response to our research findings. We then ranked these in a Feasibility vs. Potential matrix, prioritising bold and creative ideas over feasibility, as encouraged by MHR's brief. We then combined our most promising HMWs to arrive at a guiding question to frame our ideation activities:
"How might we design an AI-powered expense tool that integrates effortlessly with the tools people already use, takes away the hassle of scanning receipts, and feels like having a friendly personal expenses officer who's always there to keep things on track and provide insights when asked?"
We continued with a "What if?" activity to spark creativity through coming up with bold imaginative scenarios that could support our guiding HMW statement, before moving into Crazy 8s sketching, rapidly exploring divergent ideas individually before discussing them as a team. Some of the initial ideas included:
- An AI assistant capable of logging, creating, submitting, and reviewing expenses while offering real-time policy guidance
- Automatic detection of expenses through calendar events, card feeds, and bookings
- A digital expenses wallet akin to Apple Wallet, where expenses are stored as wallet items
- Intelligent integrations that can analyse output from or generate input for other tools
We discussed, voted on, and combined the strongest aspects from our ideas to shape an initial concept, ready for low-fidelity prototyping and iterative development.
Prototyping
With our initial concept in place, we moved quickly into prototyping to bring our ideas to life and gather user feedback early.
We began with hand-drawn sketches, translating them into low-fidelity digital interactive prototypes to facilitate easy testing with our participants, who were only available remotely. For employees, we created two alternative user flows to evaluate different approaches. Our prototyping process included two rounds of user testing focused on qualitative feedback, in line with MHR's brief to focus on ideas over polished designs.
Throughout this stage, we iterated collaboratively, integrating team feedback and client input to ensure our solution remained both innovative and grounded in user needs.
Round 1: Low-Fi
Three employees.
Based on their feedback, we combined the best elements of our two alternative designs into a unified, refined concept which also guided the refinement of the design for the other user groups. You can see some of our low-fi prototype screens from this round of testing below.
Round 2: Low/Mid-Fi
One employee and one participant with managerial and finance experience.
This round focused on refining features and aligning better with user expectations for not only employees, but also our other two user groups. Participants highlighted areas such as decluttering dashboards and allowing expense tracking to be switched off when not needed (e.g. at weekends), which we incorporated into our final designs later on. See some of our prototype screens from this stage below.
Throughout this stage, we iterated collaboratively, integrating team feedback and client input to ensure our solution remained both innovative and grounded in user needs.
At the core, each refinement was guided by our validated belief in a concept that focused on an AI- and chat-first approach, setting ourselves the goal of creating not just another expense management solution, but something that feels like your personal expenses officer.
In line with these refinements, we also reworked our "as-is" user journey maps into one unified, "to-be" user journey map going across all three user groups in a scenario enabled by our developed concept, capturing how our "personal expenses officer" would help our target users.
deliver
The Final Idea
Centsible reimagines expense management as a chat-first intelligent, AI-powered experience. At its core is "Penny", an AI assistant designed to feel like a friendly, personal expenses officer who does most of the work in the background for you.
In line with MHR's brief, Centsible's design is purposefully simple and the colour palette neutral, focusing more on seamless flows and meaningful ideas and allowing our client to easily adopt the underlying concept and turn it into MHR's next game-changing product.
AI-Chat with Penny
Users can control the entire app through text or voice, from creating expense claims and generating visual reports to asking policy questions and getting real-time budget insights.
Expense Chat
Every submitted expense creates a real-time group chat with involved employees, managers, and finance teams, with Penny assisting discussions, suggesting adjustments, and automating approvals.
Automatic Expense Creation
The system gathers data from calendars, bookings, card feeds, and location information to detect and draft expenses automatically.
Digital Expenses Wallet
Similar to Apple Wallet, expenses and receipts are stored as individual cards, ready to be submitted with a single tap.
Flexible Policies & Workflows
Finance teams can create custom workflows using simple “If this, then that” logic blocks, adapting processes to organisational needs.
Smart Integrations
Leveraging AI to read and export data to other tools without requiring custom-coded integrations.
Support for Paperless Receipts
Including QR codes, NFC, and digital booking confirmations, with scanning capabilities for paper receipts if needed.
outcomes
Our final presentation to MHR was received with enthusiastic praise. The client highlighted the originality and real-world potential of our concept:
"Centsible is a strong concept with real-world potential. You’ve clearly identified a gap in the market and responded with original, user-focused thinking. The combination of AI support, conversational UI, and seamless workflows makes this a highly promising idea for enterprise adoption."
Sean Lynch, Head of UX, MHR
They also commended our professionalism and deep understanding of user needs, adding:
"We genuinely enjoyed working with you and will reflect and share the design ideas across the organisation."
Sean Lynch, Head of UX, MHR
This positive feedback validated our approach and demonstrated the impact that bold, user-centred, and AI-integrated design can have in addressing organisational challenges.
reflections
This project was a defining experience in my development as a UX designer and team leader.
My Leadership & Teamwork
As team leader, I:
- Facilitated weekly meetings, retrospectives, and check-ins to keep the team aligned and motivated
- Managed all client communications and presentations, ensuring our work met MHR's expectations
- Delegated tasks effectively based on individual strengths and maintained our project plan and timeline
- Fostered a supportive team culture by organising peer evaluations and team socials to build trust and morale
My UX Design Contributions
Beyond leadership, I contributed to:
- Synthesising research insights into personas and user journeys
- Facilitating ideation activities to drive bold, user-centred concepts
- Developing prototypes and incorporating user feedback to refine our designs iteratively
What I learned
My biggest success was embracing and developing my professional identity and values: human-centredness, making a difference, and leadership. This project showed me I can thrive under pressure, unite a team around a shared vision, and deliver impactful solutions within tight timeframes.
I also learned the importance of balancing leadership with openness. While my team viewed me as a competent and inspiring leader, I became aware that I can sometimes focus too strongly on my own ideas. Moving forward, I will continue to actively create space for every team member's voice to strengthen collaboration and outcomes.
