GenAI mortgage broker MVP

It all started when

The client's venture team wanted to identify GenAI-unlocked opportunities that leveraged the bank's privileged data assets and served under-served customer segments — deliberately steering away from well-trodden internal use cases like customer-service optimisation. The engagement was structured as an investment case: BCG's deliverable was market analysis plus a POC of the recommended solution.

What Monica built

Monica ran discovery workshops that generated 200+ GenAI business ideas across high-touch, high-cost advisory services (financial advisors, loan brokers, lenders), then narrowed to the top three on two objective criteria — revenue potential and market uniqueness — each with a POC:

• Disruptive mortgage broker — streamlines home financing, improves offer transparency, and simplifies applications; selected because lead generation and application review/KYC were still largely manual in-market.
• Always-on business partner — 24/7 business insights for SMBs to support faster borrowing decisions, built around context-aware, RM-style servicing.
• Personal wealth coach — democratises tailored financial advice at low cost, positioned to differentiate against existing robo-advisory competitors.

Monica also explicitly rejected use cases that were already well-served or non-differentiated in-market — fraud-detection bots, hyper-personalised marketing, and commoditised OCR document processing — to keep the shortlist genuinely fundable. On the winning concept, she:

• Shaped the product-market fit analysis for the mortgage-broker concept, including market sizing: a top-down upper bound off the total Australian mortgage market (~AUD $100B) and a bottom-up lower bound (bank users with no mortgage × average mortgage value), landing on an estimated addressable opportunity of $100M+ for the concept.
• Designed the mortgage-broker product concept itself — automated document review, a status-update customer-service bot, social-media-driven lead generation, and multi-channel lead follow-up (calls, texts, emails).
• Directed the build of a GenAI demo showing conversational query of complex financial data, benchmarking feature sets against comparable GenAI financial products globally.

Key Stakeholders collaborated with

• The bank's venture team — the engagement sponsor, which set the mandate to find GenAI-unlocked opportunities built on the bank's privileged data.
• Bank employees — interviewed in structured sessions to surface the pain points that fed ideation.

Impact

• Pitched three GenAI business ideas to bank executives with a combined revenue opportunity of $400M+.
• Shipped a working demo in ~2 weeks using GPT-4, low-code tooling, and live financial data (Yahoo Finance, ASX APIs) — showcasing a conversational interface over real-time financial data.
• Produced reusable IP that accelerated later BCG GenAI sprints (unstructured data ingestion, LLM-based insight surfacing).
• The client kept socialising the ideas across business units for months after the engagement ended.

Major Takeaways

• Fast, structured ideation-to-prioritisation works best on clear, objective criteria (revenue size, uniqueness) rather than subjective preference.
• Prototyping with LLMs and financial-data APIs under a tight timeline means constantly balancing technical feasibility against the commercial narrative.
• Market sizing in financial services is strongest when top-down and bottom-up methods are combined to bound a defensible estimate.