Mortgage Buddy: AI in Finance
Filling out forms can be stressful. There’s even a social media term for it: formophobia.
But, for most people, it’s unavoidable. Taxes, financial applications, healthcare all require that we concentrate hard and answer questions on subjects most of us have barely a passing acquaintance with.
Perhaps the worst part is feeling alone, lost in unfamiliar terminology, and knowing that a seemingly small mistake can have dire consequences.
Thankfully, the form fillers of the world are not alone! If you need your customers to provide detailed and, let’s face it, often arcane information, then it is now easier and cheaper than ever to give them the help they need.
Applying for a Mortgage Piles Stress upon Stress
Put yourself in the shoes of someone looking to purchase a home. Moving is nerve-wracking. It’s that rare combination of hopes, dreams, and bureaucracy.
So, no wonder that a quarter of mortgage applicants find even the application process stressful. Mortgage applications give us a glimpse into what’s wrong with so many similar paperwork situations:
- there’s an imbalance of knowledge: the mortgage applicant can’t even be sure what they don’t know
- terminology is unfamiliar
- lack of knowledge lets the imagination inflate the potential consequences of an error: could it mean losing the dream home or even a downgraded credit score?
Filing taxes is even worse; tax collection authorities the world over are rarely known for their easygoing attitudes.
Perhaps one solution would be to sit down with each customer and coach them as they gather and provide the information needed. But, in the age of self-service apps, that would be impractical and unaffordable for most products.
There is, however, another way. Artificial intelligence techniques delivered through voice assistants and chatbots could provide 80% of the help that most people need. And for the rest? A video or voice call with a human expert would answer the questions that AI cannot and reassure those people who are not yet comfortable with AI. Let’s look at how that might work.
AI is an Inexpensive Way to Help Customers
Alejandra needs a mortgage for her new apartment.
She uses comparison sites to get started but knows that can’t give her the full picture. For one, she’s worried that she doesn’t have enough knowledge to tell a superficially good deal from a bad one. But there’s also the problem of not being sure if a comparison site is working to help her or to maximize its commissions.
She considers hiring a broker but again there’s always the question of whether they really work for her or for the lender.
Then she spots a new service called Mortgage Buddy. It’s an app that does the job of a mortgage broker but for just a small fee. And it doesn’t just help find the right mortgage but it promises to make the entire process as easy as possible.
The Mortgage Buddy app starts by asking Alejandra a few details about herself and the place she wants to buy. She’s surprised at how little it needs to know; at first, just some basic details. Almost straight away, it comes back with some details for her to confirm.
Next, the app asks to connect to Alejandra’s bank account, her tax software, and her realtor’s app. After a few seconds, Mortgage Buddy presents her with the best mortgage it can find for her circumstances.
It looks good but Alejandra is unsure. There’s a call button on the screen. She presses it and within a few minutes she’s in a video call with a human mortgage advisor. They talk through the offer and, based on some additional information Alejandra provided, there was an even better deal available.
Alejandra feels reassured by the friendly face and she says she’ll go ahead.
Later than night, Alejandra remembers a question she forgot to ask. She opens the app and finds she can chat with the service via WhatsApp. She asks her question and the reply comes back a few seconds later.
It’s the same story throughout the entire process. The app prompts Alejandra when it needs her input. When she’s not sure, she asks a question in WhatsApp. When she’s really not sure, it’s easy to speak to a friendly face in a video call.
Some things don’t change, though, and there’s still a paper form to complete. Alejandra asks the app for help. As the broker, the app has a copy of the form and talks her through each part using her phone’s loudspeaker. There’s one question the app can’t help with and so it connects Alejandra to a human who can help.
Not long after, Alejandra moves into her new apartment.
Data, AI, and APIs
So, how could an app like Mortgage Buddy work? Gathering publicly available data about Alejandra and the apartment would be relatively easy. Diving deeper into her bank accounts and so on would be a matter of using her providers’ APIs, with Alejandra’s consent.
Finding the right deal would almost be the easy part. With so much data, the app could make the same searches as a human broker. The difference is that an app like Mortgage Buddy could use machine learning to put customers into different profiles and then see which profiles are best suited to which mortgage products.
From there, it’s all about communication. The truth is, AI and service can go only so far. When Alejandra needed it, the app was able to use an API like OpenTok to connect her face to face with an expert.
Later, when she had follow-up questions the app used a product like Nexmo’s Messages API to tie its customer management system into WhatsApp. And at no point was it clear to Alejandra whether her WhatsApp messages were answered by a human or a chatbot.
Finally, when completing the form, the app used a Web RTC connection to conduct a voice call with the app’s voice assistant and, later, connected Alejandra to a human using a product such as Nexmo’s Voice API.
You Could Build This Today
Almost everything described in Alejandra’s story is available today. The Royal Bank of Scotland uses video calls, via OpenTok, to connect account managers with its private banking customers.
Airlines, parcel companies, and ride-sharing providers all use Nexmo’s messaging and voice APIs to connect customers to both chatbots and human agents.
It’s clear that we’re moving to a world where self-service, supported by AI, is the norm. But that won’t replace human interaction. Instead, human interaction will be reserved for those times when it really makes the difference. In those situations, your customer communication infrastructure needs to be ready to go on time, every time.