Improving oversight and quality assurance across mortgage advice conversations

The challenge

A major UK retail bank handled around 5000 mortgage advice conversations each month across telephone and face-to-face adviser meetings. While the bank recorded telephone calls, it lacked consistent oversight of in-person conversations. 

As a result, fewer than 5% of mortgage sales underwent quality assurance (QA) review. Manual observation was costly and difficult to scale, leaving most mortgage advice conversations without a meaningful check. This limited the bank’s ability to identify potential issues, including unsuitable advice and incomplete disclosures, before they led to complaints or customer harm. 

The bank needed a scalable way to capture mortgage advice conversations across every channel, identify higher-risk cases and build a consistent evidence base for QA and regulatory oversight.

The Recordsure solution

The bank introduced Recordsure technology to capture, analyse and review mortgage advice conversations across its telephone, video banking and face-to-face channels. 

The solution brought together Recordsure’s: 

• Conversation Intelligence 

• Workflow automation 

• Reporting & Analytics (MI)  

to give the bank a consistent view of interactions and help its QA teams focus attention where it was most needed. 

Capturing conversations across every channel

Recordsure’s Conversation Intelligence technology enabled advisers to securely record face-to-face mortgage meetings. These recordings were automatically uploaded to a central portal and analysed alongside the bank’s existing telephone and video banking recordings. 

Bringing these interactions together gave the bank a consistent view of mortgage advice conversations across its distribution channels and extended oversight to conversations that had previously been difficult to review. 

Prioritising higher-risk conversations

Automated risk scoring and quality checks were applied to every mortgage conversation. Higher-risk cases were prioritised for human review, allowing QA teams to focus on the conversations most likely to require closer attention. 

The automated scoring achieved a reported user-verified human-AI agreement rate of 95%, giving reviewers confidence in the prioritisation process while retaining human oversight. 

Structuring the review process

Mortgage conversations were automatically classified by topic and organised into structured, colour-coded themes aligned with the bank’s existing QA checklists. 

This combination of Workflow automation and Reporting & Analytics (MI) helped reviewers work more methodically and consistently. Mortgage advice reviews were completed in half the previous time. 

The outcomes

By deploying AI-led oversight across the full population of mortgage conversations, the bank achieved outcomes that would not have been possible through manual review at any reasonable cost:

• QA coverage rose from fewer than 5% to 100% of all mortgage conversations.
• Review times were cut by more than 50%, freeing QA resource for higher-value work.
• Advice standards improved, with a measurable reduction in complaints.
• QA and compliance costs fell meaningfully, with substantial longer-term savings.
• Today the bank analyses approximately 60,000 mortgage conversations per year.

Evidence

By applying technology-led oversight across the full population of mortgage conversations, the bank increased QA coverage from less than 5% to 100%. This gave the bank a more complete view of the advice being provided across its distribution channels. 

QA review times were reduced by more than 50%, freeing reviewers to focus on higher-risk cases and other higher-value work. The bank also reported improved advice standards, fewer complaints and lower QA and compliance costs, with meaningful longer-term savings. 

 The bank now uses Recordsure to analyse approximately 60,000 mortgage conversations each year, supporting a more scalable and evidence-led approach to mortgage quality assurance.