The imminent real estate chatbot uprising

home-curve-01
November 10, 2016
4
min read

Chatbots are software programs that integrate with platforms like Facebook messenger and Slack, and live in your devices in the forms of Siri, Cortana or Alexa. The basic premise is that rather than using our software in the old point-and-click way, these tiny-little computer personalities can deliver a more intuitive, human method of interaction. Good old fashioned chitchat.Imagine if, rather than having to run through a portal yourself, you could ask an AI to find you properties that match your specific needs?The idea is that, sooner than we think, we won’t be ‘using’ a lot of our day-to-day software. Rather, we’ll be shooting the breeze with it. We’ll be booking flights and hotels, and ordering pizzas and Ubers by ‘talking’ to a company's proprietary robotic assistant.I know. Sounds pretty dystopian. You have to actually talk to your paying customers? Bah! Let’s find a way to automate this en-masse without the whopping financial burden of outsourcing to the Philippines. One day we’re cheerfully ordering pizza via bot, the next our children are trying to explain their impending sense of existential dread to an android therapist who only knows how to fire pellets of valium at anything whose face appears to be leaking.Not quite. As chatbots become more and more advanced, they’ll eventually reach a human-level of conversation. The benefit then is speed. Rather than having to wait for an agent to get back to you (within business hours), chatbots will be able to run 24/7. And they’ll be able to pull up information faster than even the speediest of agents ever would.

Chatbot applications for real estate

Real estate is, and will remain to be for the foreseeable future, a people-to-people business. A chatbot will never be able to replace an agent’s role in the process, though they can be used in some pretty useful ways.Domain, for instance, has just released an Australian-first Facebook Messenger chatbot to help buyers find their dream home. Users can share their location, and the bot will pull up a list of properties for sales surrounding them. It has some pretty impressive functionality. For example, in Domain’s release of the product, they say: “Even if a specific address is not listed for sale, the bot will utilise Domain’s Home Price Guide to provide an estimated price, previous listing photos and a sales and rental history summary.”Chatbot applications reach far further than just finding buyers the right property.Imagine if a visitor to your website could be greeted by a chatbot assistant. You could assign some free, bit-based labour to ask leading questions and mine information to create a prospective client profile:

  • “Hi - anything I can help you with today?”
  • “Are you looking to buy or sell?”
  • “Do you have any preferences on location, architecture or size?”

Or how about directing clients to your agency expert on a topic? What if your chatbot had responses to commonly asked queries, such as:

  • “How can I save money selling my house?”
  • “I’m trying to find a low-rate mortgage”
  • “What are your commission rates?”

But thanks to machine learning, chatbots can get even smarter than that…

Holmes

Chatbots can be taught - sure. But they can also teach themselves through what is known as ‘machine learning’. This involves huge amounts of data and some clever algorithms to. Simply put (extremely simply, that is) if you give a pattern-matching AI enough conversations, after a while, it’ll start figuring out the rules of the game itself.Obviously this is heavily dependant on what sort of data you feed it - and the results are pretty dicey. Microsoft, for example, created an AI bot named ‘Tay’ which they raised largely on Twitter. This is the machine learning equivalent of leaving a child under Pauline Hanson’s care during their developmental years. Subsequently Tay, suddenly and without warning, became a massive, unashamed racist, and the plug was quickly pulled.But what if, rather than being fed torrential amounts of hatred from Twitter, a chatbot could learn from the conversations of normal people looking for homes, and even start learning about individual house-hunters themselves?It sounds like something Isaac Asimov would come up with, but a particularly clever startup Structurely has done just that. Their real estate chatbot Holmes is a miracle of development - and probably, the first of many just like it.Clearly expressing what you want in a home to your real estate agent is usually an exercise in futility. In fact, the vast majority of the time, buyers themselves don’t know what they want outside of their subconscious. To be sure it can recommend properties as intuitively as an agent, Holmes tracks the properties and places you’ve viewed and remembers the questions you’ve asked about them. Subsequently, it picks up you like terracotta roofing tiles even if you actually never said it.Holmes promises to grow and nurture leads until they’re ready to pick up the phone and speak to a real agent. It also promises never to turn into a radical xenophobe. Probably.

Final thoughts

There’s definitely a future for chatbots in plenty of businesses - and real estate is no exception. We’ll no doubt be noticing a growing trend of chatbots appearing on forward-thinking real estate agencies’ websites in the near future.There’s no doubt they can serve to streamline the house-hunting process for buyers, answer questions for vendors before they pick up the phone and provide agents with extremely valuable, granular data.

Here's some more article you might be interested in

For hints, tips, and community support follow