George Dickinson, CEO and founder of Latimont, worked as a mortgage broker for four years and slowly gained an ambition for technology. He has been programming for the last two years and has spent around a year and a half developing the Latimont platform.
Users can input queries about specific lenders using natural language, and the platform provides results based on current criteria.
The core system is built around multiple LLMs, not a single one, Dickinson said, adding: “My thinking is as we start to see these systems applied effectively and add value to the economy, we won’t be building systems with just one large language model; there will be systems where many are interacting with each other.”
Dickinson said he considered the user experience and how brokers wanted to receive information, then built the technology with this in mind.
When asked what alternative this offered compared to existing platforms, Dickinson said: “When I look at existing criteria sourcing systems, the ones I’ve interacted with as a broker were ones where the product is the database itself.
Aldermore Insights with Jon Cooper: Edition 5 – Feeling enthusiastic about next year’s run-of-the-mill market
Sponsored by Aldermore
“A broker will be interacting with the database directly, whereas this is essentially building a layer of intelligence in between the user and the database. The user can then express what they want without having to think too much about how to accrue specific information; the platform does that in the back end.”
The platform pulls the information published on lender websites and Dickinson plans to eventually have direct lender integration.
