Measuring AI inclusion readiness
Firms report high AI inclusion readiness, but reported practice tells a more nuanced story.
The headline score is strong, but the evidence behind it is less clear.
The AI inclusion readiness index provides a snapshot of how prepared firms believe they are to deploy AI in ways that support vulnerable customers. Across the 110 senior decision-makers surveyed, the average score is 4.1 out of 5, with financial services and retail and wholesale firms reporting similar levels of readiness.
The index considers current practice, inclusion by design and future readiness. Scores are consistently high across all three areas, suggesting broad confidence in firms’ approach to inclusive AI.
Looking beneath the headline score, however, reveals a more complicated picture. Firms’ self-assessments are consistently higher than their reported practices, while a significant minority do not monitor outcomes for vulnerable customers separately. The findings therefore raise an important question: how much of firms’ confidence can be supported by evidence?
The average AI inclusion readiness score is 4.1 out of 5, with little difference between financial services and retail and wholesale firms. Scores are also broadly consistent across current practice, inclusion by design and future readiness.
Inclusion by design scores slightly higher than current and future readiness, but the differences are small. The headline picture is one of consistently high reported readiness across the three areas measured.
Key finding
High confidence, limited visibility
Nearly one in five firms report high confidence in their AI today despite having no monitoring specifically focused on vulnerable customers. This makes it harder to evidence whether that confidence is reflected in customer outcomes.
of firms report high confidence in their AI today despite having no monitoring specifically focused on vulnerable customers.
Firms assess their readiness more highly than their reported practices suggest across all three areas. The gap is widest for future readiness, showing that the headline score does not tell the whole story.
Firms are overwhelmingly positive about AI’s impact on vulnerable customers. However, a third do not monitor outcomes for these customers specifically, making it difficult to demonstrate the improvements they report.
TPA industry perspectives
Industry leaders reflect on the report’s key findings.
Sarah Hunnings, chief operating officer, Blue Train Marketing
It’s not enough to be confident that your AI is inclusive - it must work for every customer. The widening gap between perceived readiness and reality is a red flag. Financial inclusion must be measured through customer outcomes, segmented by vulnerability, not just inferred from policies or aggregate data. Businesses must identify failing AI journeys before they become embedded at scale and inclusion is lost in the rush to innovate.
Cressida Stephenson, CEO and Founder, EdenChase Associates
Offering a human is an activity, not an outcome. The FCA's test is comparative: outcomes for vulnerable customers should be as good as for everyone else. Monitor nothing by cohort and you cannot run that comparison. Firms are measuring the contact when the regulator is asking about the outcome.
Nick Quin, chief corporate affairs officer, LINK Scheme Ltd
“Millions of people still rely on cash, and part of that is because they don’t trust digital alternatives. As we build the next generation of payments systems, we need to focus on how to build inclusion throughout them to ensure we don’t leave anyone behind.”