Enhancing agent workflows to provide better support for claimants
Organisation
UK Government
Role
Interaction Designer
Timeline
February - July 2024
Overview
Universal Credit (UC) is a UK government service designed to provide financial support to individuals and families facing living costs. The Additional Support section is a crucial part of the service, it's used to capture and address the varied needs of vulnerable claimants and those requiring help with accessibility needs. By streamlining this part of an agents workflow, we improved the accuracy of claimant data and ensured that people with additional support needs could access the help they needed.
I led the redesign of this section, collaborating with a cross-functional team including product, tech, ops and policy. My focus included UX research, workshop facilitation, defining product scope, prototyping, interaction design & user testing.
> Over 5x more page visits & actions per agent after launch
> Agents have ~80% higher confidence in data accuracy, leading to claimants receiving the right support
Field research
Workshop facilitation
Product scope
Rapid prototyping
Interaction design
User testing
The Challenge
How might we improve the agents workflow and ensure that claimants get the support they need?
The Additional Support section had become widely underused by agents, resulting in outdated claimant data and a lack of proper support. This issue affected both claimants who weren't getting the right support and service agents unable to effectively do their job.
As part of a wider initiative within the Department, extensive research identified the need to improve the functionality of the agent-side service to ensure agents were able to capture claimant data. The scope of this work was to make the support section more accessible, user-friendly, and aligned with agent and claimant needs.
Existing additional support section in Universal Credit
Discovery
Uncovering pain points & user needs in the agent workflow
Partnering with the UX Researcher and a Business Analyst in the team, we visited a Job Centre in South London. Whilst there we shadowed agents, gaining insight into their daily workflows, understanding pressures they face and the challenges of managing multiple cases and urgent requests. It became clear how important it was for their entire workflow to be streamlined, helping avoid them becoming overwhelmed and to ensuring timely support to claimants.
These observations, along with mapping workflows and reviewing content, highlighted usability issues and functionality gaps in the additional support section. Agents reported issues with poor content structure and frustration that they could only add or remove data, often losing valuable claimant history. This ultimately led to a lack of trust in the system, with agents often ignoring information or disregarding it.
After surveying agents, we found that ~97% would immediately start using the additional support section if the experience was improved and they could see a history of any changes, and 100% of agents agreed that it was a necessary part of their workflow that would be used multiple times a day.
Creating a shared understanding of the problem
To align the team and our stakeholders, I co-facilitated a cross-functional discovery workshop with the Product manager and UX Researcher. We combined research insights and data analytics to create a shared understanding of the problem space. The agenda included presenting research findings, engaging in discussions, capturing user needs and motivations, and defining hypothesis and solution statements. After discussion in the team, we prioritised scope ideas based on effort vs impact.
Following the workshop, I worked with the product manager to refine the initial scope of user needs, before presenting our findings back to the team, ensuring alignment and clarity moving forward.
Research from a Job Centre in South London
High level scope formed from the discovery workshop
Design Iteration
A modular approach to delivering updates across multiple sprints
I designed a user flow that started with understanding the big picture and the end state we were working toward. The goal was not just to map out the high-level experience and user steps, but to help the team see how each journey and feature would intersect.
This approach allowed us to break down complexity into smaller, manageable features that could exist independently, making the overall experience more flexible and scalable. I created a modular user flow and worked closely with BAs and the tech team, ensuring that updates could be delivered over multiple sprints without introducing new problems or leaving gaps.
As we ran tech spikes and wrote user stories, the flow evolved, and I designed multiple versions to illustrate how each feature should evolve over time, ensuring we stayed aligned with both the current needs and future goals.
Modular user flow mapping out high-level user experience
Evolving the design with testing & feedback
Throughout the design process, we continually evolved the solution by running multiple user testing sessions at different stages, gathering valuable feedback that informed each iteration. Regular sessions with the team and weekly design critiques with other designers allowed us to stay aligned and refine the design whilst ensuring consistency across the service.
To further validate our ideas, the operations team helped facilitate focus groups, providing a platform to share the changes with agents, ensuring their feedback was integrated into the design. This collaborative approach allowed us to make informed adjustments and create a more effective, user-centred solution.
UX decisions & trade-offs
Removing agents ability to permanently delete
In the existing experience, agents could remove support needs entirely, causing important information to vanish without a trace. This lack of historical data left agents with an incomplete view of a claimant's needs and ultimately led to lack of trust and usage. Following guidance from senior stakeholders and policy teams, we were instructed to limit the ability to permanently delete support needs. The challenge with this request, was that nobody could agree who should have the ability to delete and how easy this should be.
In response, I proposed a design for a new data structure that seamlessly moved support needs between active and archived states. This change allowed agents to easily add or remove support needs while preserving a historical record, providing a fuller view of each claimant's needs.
This trade-off meant that support needs couldn’t be permanently deleted, which could pose an issue if a mistake was made. To address this, agents found a short-term work around of making a note when updating, with the long-term goal of defining user access roles and implementing a 'destructive' delete option for necessary cases.
UX of moving support needs between active and archived state
Introducing friction to encourage more accurate data entry
Our research revealed that agents often created duplicate entries or used the custom "Other" option, even when a more appropriate pre-determined type was available. This resulted in inaccurate & messy data. To address this, we introduced a limit of 'other' and displayed a 'remaining number' counter, giving agents visibility into the system’s limitations. After testing, we found that this small adjustment encouraged agents to first check both active and past support needs before adding new ones.
While this change added some friction to the agent's workflow (something we typically try to avoid), it was essential to prevent the creation of unlimited custom entries, which were leading to inaccurate data and increasing frustration later in the process.
The solution
Enhancing Usability Through a Linear and Modular Design
The final solution combined linear task flows and simplified content design, ensuring users always knew where they were in the process and what to expect next. To enhance usability and streamline the experience, we adopted a modular layout that separated content from actions, making tasks easier to complete and reducing the time required.
The design reused components from the GOV.UK Design System and the Universal Credit pattern library, ensuring consistency and familiarity across the service, which helped users feel more comfortable navigating the system. Accessibility was a key focus throughout the design, with careful attention paid to making the service compatible with assistive technologies and meeting WCAG standards, ensuring that all users, regardless of their abilities, could access and benefit from the service.
Results
Boosting usage and delivering better support for claimants
Working on a live service meant that even small changes could have a significant impact. We adopted a "fail fast" approach, iterating multiple times based on user testing and feedback. Once live, we saw a noticeable increase in usage, with more page visits and actions in the support area. Agents began to frequently update information, reporting increased confidence in the data.
This improved process resulted in more accurate claimant data being captured and updated, leading to better support being offered to claimants. The iterative approach, combined with user feedback, drove clear improvements in both the usability of the support section and the accuracy of claimant data.
After releasing the full set of changes, we saw the following impact:
> Over 5x more page visits & actions per agent after launch
> Agents have ~80% higher confidence in data accuracy, leading to claimants receiving the right support
Reflections
Field research was incredibly valuable, providing direct insights into agent challenges and helping build empathy for their needs. By expanding beyond my role as an interaction designer, I was able to bring UX strategy into the project, helping to shape its scope and guiding the design process. During the project, I collaborated very closely with a content designer, which not only ensured a simplified, holistic experience – but also made for a great experience when iterating design to have a partner that co-owned and debated the design decisions and trade-offs with me.
Lastly, working with a mature design system was a highlight, as it allowed us to focus fully on enhancing the user experience while ensuring consistency and scalability. With accessibility at its core, we were able to make impactful, user-centred improvements that directly addressed both agent and claimant needs.