A pandemic-driven challenge
Universal Credit (UC) is a means-tested benefit administered by the Department for Work and Pensions (DWP). It simplifies the welfare system by combining several elements into a single payment for those requiring support with housing, children, disability, and caregiving.
Before the pandemic, claimants attended in-person appointments. However, national lockdowns made this impossible, creating significant identity verification challenges. As a temporary response to COVID-19, the standard allowance increased by £20 per week and housing benefit limits were relaxed.
These changes, combined with a surge in new claimants, led to a rise in fraud and error. In 2019/20, 9.4% of UC payments were lost (£1.7 billion), jumping to 14.5% (£5.5 billion) the following year.
The goals
Hippo joined the Universal Credit Fraud and Error project in January 2021. Our team of over 25 consultants (spanning product, design, research, data, and development) collaborated with DWP colleagues in a blended environment. The team supported DWP in designing and developing a service intended to help UC agents investigate and detect fraud and error cases.
The project aimed to create a service that:
- Ensures claimants receive the correct payments at the right time.
- Detects overpayments due to fraud or error.
Secondary goals included:
- Guiding agents through a structured review process.
- Protecting claimants with access barriers and securing sensitive information.
- Prioritising reviews for maximum efficiency and scalability.
- Gathering data to inform future UC developments.
Understanding the challenges
Our initial focus was on understanding the existing service and its shortcomings through on-site observations and user research sessions with agents. The discovery work identified several key issues:
- Lack of Process: Agents had no established review process, minimal guidance, and limited powers.
- Security Risks: Claimants lacked a safe way to share bank documents digitally; manual processes caused both security and efficiency issues.
- Inefficient Evidence Gathering: Agents manually requested ad-hoc documents with no secure upload method, making “Freedom of Information” requests nearly impossible.
- Communication Gaps: Appointment reminders were sent via “journal” messages with no clear instructions or integration into the claimant’s UC diary.
- Manual Tracking: Agents spent 30-80 minutes daily checking for evidence updates because the system lacked automated notifications.
- Ambiguous Naming: The original name, “Targeted Case Review,” was inaccurate as claims were not initially “targeted” and reviews were not “cases.”
- Accessibility: The existing service had limited information hierarchy and had not undergone structured user testing.
A user-centred delivery approach
The Hippo team adopted a two-week sprint cycle to ensure rapid, user-friendly improvements. We explored technical solutions to transform how agents complete reviews, how claimants receive notifications, and how data is captured.
To ensure these solutions were accessible and scalable, we conducted regular visits to UC hubs across the UK. These in-person sessions fostered rapport with agents, who proactively helped us identify pain points, allowing us to refine our design and development approach in real-time.
Blended teams
Hippo integrated into a blended team of DWP civil servants and contractors. We ran workshops to define working practices and collocate assets, ensuring the Hippo team could seamlessly navigate the project’s complexity within the wider UC community.
Navigating complexity
The project’s primary hurdle was the convoluted nature of the existing journey. With no single “right” way to conduct a review, defining a universal solution was challenging. Furthermore, Universal Credit is a technically intricate system, the work required GDS alignment to fit service-specific style guides and codebases.
The solution
The team contributed to developing an end-to-end solution within the complex UC codebase, delivering iterative improvements alongside long-term service features:
- Agent Process: A pragmatic tool using “Digiman” (a proprietary UC solution) that guided agents through reviews using binary questions, simplifying the process for both English and Welsh speakers.
- Secure Document Upload: A simple, secure method for claimants to upload bank statements and evidence, featuring formatted instructions to improve accessibility.
- Analysis and Interview Tool: Collated all uploads into one location, allowing agents to prepare notes and record interview answers securely.
- Bespoke Appointments: A new system for invites and reminders that integrated directly into the claimant’s UC calendar.
- Agent Preview Tool & Dashboard: A consolidated summary of claim info and tasks. This eliminated the need to hunt for data across multiple sources, saving agents 30-80 minutes per day.
- UC Claim Review (UCR): The service has been enhanced with a dynamic task‑list model that adapts to different claimant scenarios and provides a stronger audit trail, forming a solid foundation for continued automation.
Third-Party Resource Augmentation
In September 2024, the DWP began scaling reviews using third-party supplier agents. We modified the service to facilitate this “Resource Augmentation,” building bespoke access requirements and handover solutions to ensure a smooth transition when a case requires a change to a claimant’s payment.
Outcomes
The Universal Credit Fraud and Error project delivered a brand-new service that has supported DWP operations:
- National Scale: Enabled the service to scale to 3,500 agents.
- High Volume: 56,000 claim reviews are now completed every month.
- Significant Savings: Helped the programme save £69.8 million between April and June 2024.
- Efficiency: Reported reductions in agent processing time were observed.
- Contributed to broad DWP initiatives that successfully prevented £1 billion in incorrect payments, with savings set to rise to £13.6 billion by 2030.
