AI feature research for a UAE bank

Background

‍

A major UAE bank, in partnership with a global payments network, was preparing to launch a first-of-its-kind generative AI feature within their mobile banking app. The feature would allow customers to discover personalised lifestyle offers - from dining and entertainment to retail - through natural voice or typed conversation with an AI assistant.

Before committing to a full build, the bank needed to know whether the concept would land with real users. Digital of Things was engaged to validate it with customers.

‍

The challenge

‍

A generative AI assistant embedded in a banking app was new territory for the client and for its customers. Before investing further, the team needed evidence, not assumptions, on a set of open questions:

• Would customers understand they were being offered lifestyle perks, and not mistake the feature for a financial product?
• Would they trust an AI assistant enough to act on what it recommended?
• Could they complete the journey, from first prompt to offer redemption, without guidance?
• Would voice interaction feel natural inside a banking context, or introduce friction of its own?

Getting this wrong after launch would mean a costly rebuild and a missed opportunity to differentiate in a crowded banking app market. The client came to DOT with a clear brief: pressure-test the concept before it went any further.

‍

The approach

‍

The work moved through three phases.

‍

1. Research design and recruitment

DOT designed a moderated usability study built around a prototype of the feature, structured to test every key touchpoint end to end. Ten participants were recruited to match the target demographic precisely - a mix of existing and non-customers, spanning multiple nationalities, age ranges, and device types.

‍

2. Fieldwork

Sessions ran across three days in DOT's research lab in JLT, Dubai, with each participant completing a structured one-hour session. The lab's eye-tracking and emotional response capabilities captured both what participants said and what they actually did, revealing gaps between stated preference and real behaviour that surveys alone would miss. The team assessed the experience across the homepage entry point, onboarding screens, the AI conversation interface, microphone permissions, and individual offer detail pages.

‍

3. Analysis and reporting

Every finding was severity-rated, from major friction to positive signals worth amplifying, and paired with a recommendation the product team could act on. The concept itself tested well: participants were receptive to AI-powered offer discovery in a banking context, and the proposition generated genuine interest. But the research also surfaced meaningful gaps between what the prototype communicated and what users understood:

• Offer copy on the homepage created ambiguity - several participants expected financial products rather than lifestyle offers, so the proposition's value wasn't landing from the first moment of exposure.
• The onboarding flow was perceived as lengthy, with progress indicators not pulling their weight.
• Hesitation around the microphone interaction - tap versus press-and-hold - created unnecessary friction at a critical point in the journey.
• Offer relevance was a concern: when the AI surfaced results that didn't match the user's intent, trust in the feature took a hit.
• Offer detail pages lacked the clarity users needed to confidently redeem benefits, particularly where no promotion code was present.

‍

The impact

‍

The client received a detailed findings and recommendations report covering 20+ screens and interactions, with fixes prioritised against effort and severity. It gave the product team a clear, evidence-based picture of where to invest before MVP release, including several quick wins that could meaningfully increase feature uptake without significant development effort.

Participants expressed genuine enthusiasm for AI-curated offers once the journey friction was resolved, validating that the underlying proposition had strong commercial potential.

The client acted on DOT's recommendations and successfully launched the POC. DOT's research gave the team the confidence and direction to move from a promising prototype to a live product, shaped by real user behaviour rather than assumptions.