Case Study: Kalgera / Fintech Scotland
Financial Vulnerability Research: Kalgera / Fintech Scotland
End-to-end qualitative user research validating AI-driven early warning signals for financially vulnerable adults in Scotland. Three structured outputs delivered: signal validation, acceptability framework, and Finance and Health Lab summary.
Screening Survey
Quant + qual
qualification mapped to signals
In-Depth Interviews
1:1
60 minutes each, semi-structured
Signal Categories
8
mapped to Kalgera architecture
The Challenge
Kalgera's AI-driven early warning system detects financial vulnerability through transaction data signals. Before scaling the product, Kalgera needed primary qualitative research to validate that the signals reflect real lived experiences, not just statistical artefacts. The work was delivered within the regulatory context established by the FCA Consumer Duty, which places explicit obligations on firms to understand the needs of customers in vulnerable circumstances, and the Financial Conduct Authority's Guidance for Firms on the Fair Treatment of Vulnerable Customers (FG21/1).
The research had to capture the experiences of financially vulnerable adults in Scotland: people experiencing cognitive decline, scam victims, carers managing money on behalf of others. Reaching this population ethically and reliably required specialist recruitment and a robust ethical framework aligned with the Adult Support and Protection (Scotland) Act 2007. Fintech Scotland, in partnership with the Scottish Government's Financial Health Lab, supported access to the target participant cohort.
Research Programme: Three Stages
| Stage | Method | Detail |
|---|---|---|
| 1 | Paid Social Recruitment | Facebook and Instagram campaigns. Primary target: financially vulnerable adults in Scotland (50+). Secondary: 35–49 age group. |
| 2 | Screening Survey | Quantitative and qualitative data. Qualification criteria mapped to Kalgera signal categories. |
| 3 | In-Depth 1:1 Interviews | 60 minutes each, 1:1. Semi-structured protocol mapped directly to all eight signal categories. |
Kalgera Signal Categories: Interview Protocol
Each interview was structured around all eight of Kalgera's signal categories. Participants were asked to describe experiences relevant to each category in their own words.
| Signal Category | What it captures |
|---|---|
| Spending pattern changes | Sudden shifts in regular spending behaviour. |
| Income depletion | Faster-than-expected drawdown of available funds. |
| Credit reliance | Increased use of overdraft, credit cards, or BNPL. |
| New payees | Payments to previously unseen accounts. |
| Cash patterns | Unusual ATM withdrawal frequency or amounts. |
| Bill changes | Missed direct debits or new recurring charges. |
| Account access | Login frequency, timing, or device anomalies. |
| Scam indicators | Transaction patterns consistent with known fraud typologies. |
Ethical Framework
- •Consent and distress protocols aligned with the duty-of-care principles of the Adult Support and Protection (Scotland) Act 2007, applied throughout the research programme.
- •Distress protocol in place with trained facilitators for all interviews involving adults at risk.
- •All data encrypted, UK-hosted, and anonymised within 7 days of collection.
- •Verbal and written consent obtained for participation, recording, and data use.
"Qualitative research with financially vulnerable adults demands the same care and ethical rigour we would apply to any research with adults at risk: distress protocols, informed consent, and a methodology grounded in lived experience. Aligning our protocols with the duty-of-care principles of the Adult Support and Protection (Scotland) Act 2007 gave Kalgera's signal validation genuine human validity."
Dr Stylianos Kampakis, Managing Director, Tesseract Academy
Three Structured Outputs
Output 1
Signal Validation Report
Confirms which behavioural markers are observable in transaction data. Grounds Kalgera signal architecture in lived experience.
Output 2
Intervention Acceptability Framework
Documents what vulnerable people consider helpful versus intrusive. Defines the acceptability spectrum for AI-driven nudges.
Output 3
Summary Findings Report
Condensed findings for the Finance and Health Lab. Direct participant quotes used to support product decisions.
