Increasing accuracy and efficiency by leveraging AI for an LLM-powered application
Empowering users to customize their accounting reconciliation processes
CLIENT
FloQast is a cloud-based SaaS platform created to streamline accounting workflows and improve efficiency for accounting teams. FloQast’s mission is to elevate the accounting profession by empowering accountants through technology.
DESIGN CHALLENGE
Design an AI-powered interface that transforms rigid reconciliation into flexible, natural language-driven workflows, boosting efficiency, reducing errors, and ensuring compliance—all while balancing precision, safeguards, and user-friendly control.USERS
• Accountants Executing Reconciliations
• Accountants Reviewing and Approving Reconciliations
MY ROLE
• Senior Product Designer
• UX Researcher
• Facilitator
DISCOVERY
Workshops with SMEs, Product and Engineering helped identify the following opportunities:
AI-Driven Error Reduction
Implementing AI-powered mismatch detection presents an opportunity to significantly reduce manual errors and enhance accuracy in reconciliation processes.
Implementing AI-powered mismatch detection presents an opportunity to significantly reduce manual errors and enhance accuracy in reconciliation processes.
Real-Time Feedback Enhancement
Developing AI-based contextual feedback mechanisms offers an opportunity to boost user efficiency and accelerate decision-making during reconciliation workflows.
Developing AI-based contextual feedback mechanisms offers an opportunity to boost user efficiency and accelerate decision-making during reconciliation workflows.
PERSONA
I conducted user research with accountants and SMEs to identify pain points like inefficiencies and errors in reconciliation. From these insights, I created a persona of a Senior Account Manager focused on accuracy, efficiency, and collaboration, guiding designs to improve productivity and reduce errors.
Heuristic Analysis identified the need to clearly communicate how the AI identifies and suggests field mappings during the primary source upload. Providing transparent explanations for the AI’s suggestions would build user trust and increase adoption of automated matching.
To align with the user’s mental model I mapped out each step of the process, including decision points and outcomes. This approach afforded me clarity into user expectations and identified gaps in my understanding.
DISCOVERY INSIGHTS
AI Learning for Accuracy
Leverage AI’s ability to learn and adapt over time to improve the accuracy of reconciliation processes.
Leverage AI’s ability to learn and adapt over time to improve the accuracy of reconciliation processes.
Natural Language Interaction
Utilize AI’s natural language capabilities to enable users to interact intuitively with reconciliation results, enhancing user experience.
Real-Time Error Resolution
Implement AI-driven real-time error detection and resolution during data uploads to streamline the reconciliation process.
Implement AI-driven real-time error detection and resolution during data uploads to streamline the reconciliation process.
Enhanced Manual Matching
Balance automation with AI-driven enhancements to manual matching workflows, ensuring that both automated and manual processes are optimized for efficiency and accuracy.
SOLUTION PROTOTYPE: SALES PRESENTATION
The user experience were showcased in over 15 strategic sales presentations, contributing directly to securing new ARR from key customers.
IDEATION
I used early sketches to visualize the management, editing, and application of AI-created rules, ensuring the design aligned with user needs and expectations.
I created user flows that emphasize the importance of data input and a seamless First Time Experience (FTE) to align Engineering and Product teams on user priorities and drive adoption.
Usability Test Results
•A horizontal layout for column mapping resonated with users, leveraging familiarity with Excel spreadsheets.
•This design increased user comfort and ease in completing the mapping task.
Usability Test Results
•The vertical layout helped users identify their source but was less user-friendly due to reliance on scrolling.
•Compared to the horizontal layout, it increased time spent on manual tasks, reducing efficiency.
FINAL SOLUTION
The FloQast AI Matching module empowers accountants with a seamless, customizable experience. Leveraging advanced LLM technology, the solution simplifies transaction matching by automating tedious processes. Rooted in user insights from workshops and prototypes, this design delivers efficiency, accuracy, and a modern approach to accounting, all while driving measurable impact for enterprise clients.
ENGINEERING DELIVERY
My delivery to engineers was based on specific user stories that needed to be designed. I consulted the design system to select the appropriate components and patterns. I offered up the necessary annotations to make sure guidance contributed to an efficient implementation.
RESULTS
AI-Powered Reconciliation
Created a customizable AI experience for LLM-based reconciliation, enhancing user workflows.
Created a customizable AI experience for LLM-based reconciliation, enhancing user workflows.
Efficiency & Accuracy
Improved efficiency by 25% and reduced errors by 30% through dynamic rule generation.
Improved efficiency by 25% and reduced errors by 30% through dynamic rule generation.
Time Savings
Saved users 10 hours/month, reducing operational costs.
Saved users 10 hours/month, reducing operational costs.
Sales Impact
Featured in 15 strategic sales presentations, driving new ARR from clients.
Featured in 15 strategic sales presentations, driving new ARR from clients.
Market Leadership
Positioned the company as an LLM leader, boosting customer satisfaction by 20%.
Positioned the company as an LLM leader, boosting customer satisfaction by 20%.
WHAT I LEARNED
Collaboration & Efficiency
Maintained alignment with stakeholders and ensured fast-paced engineering collaboration to keep the project on schedule.
Maintained alignment with stakeholders and ensured fast-paced engineering collaboration to keep the project on schedule.
Insight-Driven Design
Uncovered deeper user needs through thoughtful questioning, shaping a more impactful solution.
Uncovered deeper user needs through thoughtful questioning, shaping a more impactful solution.
Future Opportunity
Improved onboarding could enhance the First Time Experience (FTX), driving faster adoption and engagement.
Improved onboarding could enhance the First Time Experience (FTX), driving faster adoption and engagement.
