Back in the 70s getting mortgages took months – there was a mortgage queue (&before techies get over-excited about how “bad” this was it was one of the factors restraining insane bull markets in residential property). Fast forward and the median mortgage in the UK is granted in 19 business days which is still a long time to wait given the stresses associated with the UK legal structure of buying a house. MPowered Mortgages, a regulated mortgage lender in the UK, have reduced this to a median of 3 days to make a decision and are aiming in the next 18mts to make half of those done in a few hours.
Stuart is the CEO of MPowered (and its sister company MQube seller of AI products into regulated industries more globally) and guides us through this astonishing outperformance. Like many a hardcore Fintech founder he has a long history in FS behind him having been CEO of several banks in Asia for Lloyds Bank for many years.
Even so this is no mean achievement as Stuart himself found out recently when he needed a mortgage but – under the “rules” – was not allowed to use his own company to borrow from.
Topics discussed include:
- Stuarts recent experience of buying a house and getting a mortgage
- running banks in Asia and being a nomad
- the UK Housing market
- managing banks in a crash period
- “inventing the motor car and a regulator telling you you must use a horse for travel as you have a vested interest in motor cars” 😀
- pivoting from prestigious senior roles to becoming a founder
- the stark differences between managing banks and tech firms along with the internal culture
- banning GANT charts and project plans
- autonomy managing businesses 8 hours time difference away from HQ
- challenges of returning from such to HQ
- cofounding a new business starting with just the motivation to do so yet a blank sheet of paper re area/plans
- the business was formed about 6yrs ago
- the value of clear and simple propositions as the corporate mission
- desire for fast mortgages vs certainty – what consumer research shows
- “a mortgage is a necessary evil for buying the house of your dreams”
- challenges of using technology in a highly-regulated world
- challenges of the UK process of buying a house
- importance of surveys and the delay caused by those
- “making fast decisions in the industry isn’t the problem the problem is getting fast access to clean data”
- fixing the GIGO problem
- three types of survey – automated digital, desk-based one or physical visit
- about 50% of the time physical surveys are done – hence the other half are way faster – ~immediate or a few hours
- comparison with insurance price comparison sites
- do loan to value ratios vary depending on the type of survey?
- collecting all the other requisite data for the decision
- anomalies in the input data as a major challenge – using AI to cut through this
- automating data collection
- “when you are filling in our application form which takes around 10 minutes we collate around 19,000 pieces of data (!!!) in that 10 minutes”
- super-advanced autocomplete – only asking the borrower for information not known rather than tons of info your APIs can find
- examples of the type of automated data that MPowered autocollect while you are filling in the form
- the regulatory constraints on tech development
- AI is never used to make a decision rule-based engines are always used for that
- AI is to supplement/replace the human intervention around the nature of data and the inconsistencies
- shoutouts for MPowered and MQube and current products, services and needs
And much much more 🙂
Share and enjoy!