Recommendation 13

We asked the Department how it ensured that vulnerable claimants were taken into account, in particular when it uses data analytics and machine learning. The Department told us that it was considering and testing for vulnerability “at every stage”.23 It acknowledged that that its data analytics tools focused on characteristics for fraud rather than vulnerability, but that it had other processes to identify and signpost support to claimants with complex needs before a claim would be suspended due to suspected fraud. It explained that this process started with the initial assessment of the claimant by a case worker, and that if evidence of vulnerability was suspected, there were markers within the Department’s systems that can be used to flag this in the claimant’s case. We asked whether this flagging process is automated, and the Department told us it was entirely manual and relied on the discretion of the case worker. It further explained that where there were suggestions of vulnerability, the next stage would be a ‘case conference’ in which a wider team, which it calls ‘Advanced Customer Support Leads’ would assess the case to help ensure the right decision was made about whether to suspend payment. In a case where the claimant had failed to engage with the Department’s processes, this wider team may contact people who might know the claimant—such as their landlord, Local Authority, or GP—to establish what is going on. The Department told us it has expanded the team that specialises in the assessment of these complex cases from 30 to 36. It added that it also had an automated ‘call listening’ service in place to detect key words and phrases associated with vulnerabilities, and that this covered 30% of claimant calls and was planned to expand to 100%.24 20 DWP ARA 2021–22, pages 64, 66, 223–224 21 Q 26 22 Qq 27–29, 78–80, 82–83 23 Q 61 24 Qq 61–65, 79–81 The Department for Work and Pensions’ Accounts 2021–22 – Fraud and error in the benefits system 13