In October 1965, Lawrence Roberts and Thomas Merrill communicated with one another via two distant but connected computers, a seminal event.

From that moment – followed by arduous efforts and big milestones, i.e. ARPAnet (1969), TCP/IP (1974), NSFNET (1986), etc. – to the point in 2022, where 5 billion people were online worldwide, the furniture of our human minds has undergone a lot of re-arranging.

During all of that mental rearranging of furniture, the Internet's nature has powered both technology's exponential change and people's exponentially-evolving adoption and demand of technology in beyond-breathtaking, mind-boggling ways.

Now, with ChatGPT and other generative A.I. tools that have enabled computers to talk to and learn from one another as well as from us, an awful lot of new rearranging of our mental furniture has begun.

Excitement about the power A.I. will unleash – relieving tedium, performing grunt research, generating best-practice recommendations, and other tasks, securing time for meaningful, productive focus, etc. – is, among homebuilding and development professional friends and associates, at a fever pitch.

Problem many of them express is befuddlement about where to get started. Particularly, how not to fall into a rabbit hole of exploring how wonderful the new A.I. tools are at doing so many things that don't have absolute and immediate-return relevance with respect to building, marketing, selling, planning, financing, and delivering new homes.

The "general knowledge" nature of the training of most A.I. tools can be fascinating and at the same time tantalizing.

Mosaic co-founder and CEO Salman Ahmad described to The Builder's Daily audiences the challenge this way:

The next step is getting domain specific data that allows the A.I.s to talk and converse about vertical specific domains, domain specific information. That's why data is so necessary for A.I.s to function within the real estate world. Once you have that, developers and general contractors will be in a position to use software and technology to do things much more quickly than they were able to do before."

The domain, then is the big nut to crack. Early-on-the-machine-learning curve resources are in development, and one of them Homescribe.ai, focuses on catalyzing a community to rapidly scale a dynamic knowledge whose generative A.I.s are designed specifically for high-volume production builders' and developers' new-home sales and marketing.

Per a background statement shared with The Builder's Daily:

HomeScribe.ai, a new artificial intelligence (AI)-powered company created specifically for the home building industry, is proud to announce the launch of its Beta version. HomeScribe.ai is driven by the community and optimized with real-time data. The AI is trained using publicly available data from the top 200 home builders.

Built to cater specifically to the sales and marketing community, HomeScribe.ai is loaded with 100 examples, over 20 training videos, and recipes to help users learn how to use AI. Think of a recipe as a chat prompt that is already created, and all a user must do is plug in variables.

HomeScribe.ai is free to use, and the community has free access to weekly training sessions to improve their AI skills.

We had an opportunity to speak with HomeScribe.ai founder Barrett Davis, and what became clear in his initiative is an intention to scale up fast in domain expertise and applicability by building a community around learning, sharing best practices, and galvanizing positive sales and marketing efforts – and outcomes – as a business community nexus.

Here are a few of the highlights of our conversation:

Origin Story

Q: What motivated you to start up Homescribe.ai

Barrett Davis

I met with a lot of people at IBS and they were very curious about A.I. but didn't know how to use it in homebuilding. It seems sort of like a gimmick. So the idea for me was to try to create some sort of free community platform, essentially like a library where they can start to use some of the new A.I. technologies.

Scoping The Domain

Q: Walk us through how it works?

Barrett Davis

We're actually combining A.I. and a lot of the homebuilding datasets. We made the datasets live and online. A lot of the public homebuilder data is essentially public and it exists on all the Google services, you know, in the listing services. There's a lot of information out there that A.I. does a good job of discerning back to getting quality data.

There are large A.I. models that we're using, like the ones from ChatGPT. There are others that we combined to make a perfect secret sauce in the background from information already sort trained on the public Internet.

So, if you're a new home counselor, and you're writing a sales email, you could say 'I'd like to produce a new cold email. I'm a builder, and I sell in this market. I want the email to be funny and entertaining. The A.I. model will go look at what it's been trained on and also use HomeScribe.ai to take a peek at what's publicly out there too. So with this cold sales email goes out, our follow up email will actually throw in some data points around some qualifiers. For example, what's your builder? What community it is. What model home it is, and some of that details already been trained on or on the Internet. It will reference both of them.

Market-Specific Solutions

Q: When you say 'available on the internet,' you're talking about new community listings, selling communities, models, coming soon communities, etc?

Barrett Davis

Absolutely. Anything that they do for a community launch: related newspapers, blogs, etc., all of that ends up back on the Common Internet. It's just a global resource that everybody uses to find information. Their marketing websites or affiliated websites or any sort of advertising they do at all on Google or Facebook and other services like that still pushed to public areas on the internet. A lot of these A.I. models have been trained on those already. What we do is develop extra pointers to where to navigate maybe to certain homebuilder websites, or a certain public datasets, you know, that are specifically around search engines and things like that. A lot of it is just sort of going out there and retrieve information it already has. Or else, learn that it doesn't have the information or doesn't have accuracy. It will go search public repositories and try to find it too.

Initial Focus: Marketing & Sales

Q: Is HomeScribe.ai specifically designed for marketing and sales, sales management, training, etc.?

Barrett Davis

Yes, it's primarily for sales and marketing leaders at the moment. That's really where it's focused at the moment but it can definitely expand quickly. The thing that can help with sales and marketing leaders are prompts for property descriptions. This way you can make it SEO-enhance with a click of a button. A recipe, as we call it, inside the tool has a video so you can learn how to use a command. You can click a simple button like 'run,' and it will show you an example on how to copy and paste your property description into the variable, click run, and it will generate you new property descriptions that can be more entertaining or more professional. It can also enhance them for SEO. We use the SEO keywords from Google, we look at the public repositories for the high level of SEO rankings and things like that. And then we just sort of hand it over to the A.I. model, and it will actually spit back a better SEO-enhanced property description than what was on the website.

So, a marketer can copy and paste their property description into HomeScribe, click a button and they'll get a better property description. For homebuilders, let's say, in Texas or Southern California – that might have a lot of multilingual buyers -- you can take your property description, click our multilingual Property Description prompt in the library, and it can turn into Spanish, Chinese, etc.

So, from the simple clicking of a button, you can enhance your marketing website and some of the day-to-day, tedious tasks for somebody to do some of that manually, where they go online, look up keywords and figure out how to change this into Spanish. They'd have to spend 30 minutes or more. Homescribe is just trying to automate some of the manual tedious tasks that impact a lot of their day-to-day work. If you need to write a blog, this can write a blog for you. You know, it will produce 80% of a great blog the other 20% is the editorial piece that you need to come in with. But it starts with an example, and we have prompts and recipes around how to use the tools to write a blog, etc.

Tipping Point In Real-Time: Now

Q: So, would you characterize the platform as a how to become A.I. fluent in in today's selling environment with respect to sales and marketing practices?

Barrett Davis

Absolutely. There's a changing of the guard. People are starting to beta test the platform. People tend to use these new A.I.'s as their search engine as their first place where they they'll try to find information because they're just getting facts and ads. And so you can actually search for homebuilder data and you can search result listings, etc. using HomeScribe because it's probably on the internet, it's sometimes a little bit more clear than what you would get on other traditional means. You just flip us a question, 'is it better to buy a new home or old home?' And the results come back with data from homebuilders, websites, just as an example. So the goalpost is shifting for, you know, how how people might find a home builder data and how consumers are sort of searching for new homes as well.

Now, homebuilders can explore the consumers experience, imagining, 'If I was using an A.I. model to look for homes how does our our website stand out? What information comes back? Is it from 2021? Or is it live?' Anyway, it's a tool for the community to become fluent in A.I. It can help their day to day productivity in terms of updating website content or, making new brochures, or getting inspiration. It's really good at solving the blank- or the white-canvas problem, where you have a great idea but you just don't know how to get started. You can kind of type different data in to get inspiration for new content or new ideas around deals or sales or community names, etc.

The HomeScribe community

Q: Could you talk a little bit about the business model for HomeScribe?

Barrett Davis

It's free to use right now. We're building it out as much as we can, like a library. So it's meant to be about data. We do have a pro tier, which is built to cover costs. Sort of, it's not exactly like a big revenue generator. It's meant to be to a community platform. In the pro tier, teams can get group training, where people can ask specific questions to me and my partner, who's a developer as well. So it's just two developers working in the background and taking in problems or ideas for the community, and finding the best sort of prompts are what to tell the A.I. so we can add it to the library. And then, once we find some of the best tips and tricks, we should have a full library throughout the year with a few hundred of those best practices. And that should cover a majority of the use cases for homebuilding."

There's a large open-source community where you can basically build technologies that a lot of people can use free, and donations support the costs. Our business model is being built around keeping operational, and keeping it successful. In that way, HomeScribe is built to be a community tool.