Year

2025

Product

Mortgage Assistant

Designing AI-Powered Digital Assistance in a Regulated Mortgage Space

Enables brokers to efficiently find suitable mortgage products

Specialist mortgage lenders offer a wide range of complex products, making it difficult for brokers to quickly identify the best options for their clients. To address this, an AI-powered digital assistant is designed to enable brokers to efficiently find suitable mortgage products through a chat-based interaction while ensuring compliance with strict regulatory requirements.


Shaping the vision and execution of AI-powered assistance in mortgage product discovery

• Leading cross-functional collaboration between design, AI engineering, legal, and business stakeholders to create a scalable and compliant solution.

• Conducting extensive user research to understand broker workflows, challenges, and expectations.

• Defining AI interaction models and ensuring a balance between efficiency, transparency, and trust.

• Establishing regulatory guardrails in collaboration with compliance teams to mitigate risks.

• Driving the iterative design process through prototyping, testing, and refinement to optimise usability and adoption.

Business Problem & Strategic Approach

Key Challenges:

• Time-consuming process - Brokers spend significant time searching for suitable mortgage products.

• High query volume - Many queries require manual handling, adding operational cost.

• Complex product variations - Lenders offer diverse mortgage products with intricate eligibility criteria.

• Regulatory constraints - AI-generated responses must align with financial regulations while maintaining usability.

Strategic Approach:

• Defining AI Use Cases & Boundaries - Ensured AI-assisted responses were informative rather than advisory, minimising compliance risks.

• Enhancing User Transparency - Implemented reason-based responses and disclaimers to build trust and clarity.

• Data Governance & Privacy - Worked closely with legal teams to establish robust data handling protocols.

• Stakeholder Alignment - Balanced the needs of lenders, brokers, and AI capabilities to drive product adoption.

Design Process & Solutions

To ensure a seamless experience while maintaining compliance, I used design process that included research, prototyping, testing, and iterative refinements. Addressing the problems with design methodologies to implement optimal solutions.

One of the major concerns was to build trust of the agent. By providing reasoning and sources, it ensures brokers can check over the response or recommendation by providing sources and justifications.



By making sure the tool align with the brand, it create brand recognition in building trust from the brokers.


Conversational Experience Design

Understanding User Expectations - Conducted research to align AI responses with broker mental models and typical chat interactions.

• Guided vs. Open-Ended Queries - Designed a hybrid approach, allowing structured and free-text inputs for flexibility and efficiency.

• Handling Uncertainty - Implemented fallback mechanisms and clarification prompts to prevent dead-ends and user frustration.

• Reason-Based Responses - Integrated transparency-driven explanations to help brokers understand why certain products were suggested.


Product Discovery Experience

• Optimised Information Hierarchy - Designed clear, scannable layouts to highlight key mortgage details.

• Dynamic Filtering & Refinement - Allowed brokers to interactively refine product matches to meet client needs.

• Contextual Explanations - Integrated tooltips and inline guidance to help brokers understand product eligibility and terms.



Regulatory Compliance & AI Governance

• Transparency & Disclaimers - Ensured AI outputs included regulatory-compliant messaging to prevent misinterpretation as financial advice.

• User Testing for Compliance - Conducted usability tests with compliance teams to refine AI responses and reduce legal risks.

• QA & AI Validation - Established rigorous testing frameworks, including real-case scenario simulations, to fine-tune accuracy and reliability.

Impact & Outcomes

• 24/7 Availability - Enables brokers to handle inquiries anytime without human dependency.

• Operational Efficiency - Expected to significantly reduce the time brokers spend on product discovery.

• Resource Optimisation - Anticipated reduction in manual workload for support teams.

• Accuracy Improvement - Reduction in errors caused by manual product selection.

• Regulatory Compliance - Successfully implemented AI guardrails, disclaimers, and auditing processes to meet financial regulations.

This initiative represents a significant step in transforming broker workflows, making mortgage product discovery faster, more accurate, and fully compliant in a highly regulated industry.

The strategic design framework established in this project also sets a foundation for future AI-driven initiatives within the company.