326 Super Juicing Your Lead Scoring With AI - Josh Paul Jul 1 2026

Greg Bray: [00:00:00] Hello everybody, and welcome to today's episode of the Builder Marketing Podcast. I'm Greg Bray with Blue Tangerine.

Kevin Weitzel: And I'm Kevin Weitzel with OutHouse.

Greg Bray: And we are pleased today to have joining us Josh Paul. Josh is the founder and CEO of Homebuilder Ops. Welcome, Josh. Thanks for being with us today.

Josh Paul: Thanks for having me, Greg, Kevin.

Greg Bray: Well, Josh, just, start by giving us that quick background introduction for those who haven't had a chance to meet you [00:01:00] before.

Josh Paul: Sure. So, been working with home builders for almost a decade, and I've been working with HubSpot every single day since 2010. About five years ago, the two of them came together and we spun off Homebuilder Ops from my other HubSpot consulting and training company. It's grown to be HubSpot's number one partner for home builders. It is such a great industry. If you get this right, the ROI is tremendous for builders.

We're not an ad agency. We don't do websites. We don't do content. We focus on the revenue and the plumbing and the reporting and the automation and the customer experience. And so, there's a lot of things that we touch, a lot of things as the HubSpot platform has grown, that we've had to become experts in, but we're seeing tremendous results in this industry.

Kevin Weitzel: All right, before we jump into all that, let's get an interesting factoid about you personally that has nothing to do with work, the home building industry, or family. What do you got?

Josh Paul: All right. Fun fact. Many, many, many lifetimes [00:02:00] ago, I worked at Walt Disney World, my wife and I actually met while both on an internship at Walt Disney World. So, big Disney fans in our family.

Kevin Weitzel: What? So, like, did you do one job or a gamut of jobs? Were you cafeteria? Were you the guy that pushes the safety stop button? What'd you do?

Josh Paul: I was a deep water lifeguard at Blizzard Beach Water Park, not to be confused with a shallow water lifeguard at Blizzard Beach Water Park. It was a wild semester at Disney. Got to meet some of the best people on earth, but also do things that a normal person does not get to do or have access to. But it was intense being a lifeguard with that volume of people.

Kevin Weitzel: So, it's the happiest place on earth, or at least Disneyland says that, and I think Disney World by default is also the happiest place on earth, which I don't know how you can have two of the happiest places on earth if it's the happiest. Happiest indicates first place. Anyway. Were all of your peers the happiest people working there, or were there days where you're just like, "Ugh, I've gotta go save somebody in the [00:03:00] pool again today?"

Josh Paul: The operation is so big that there's always ridiculousness all around you. So, All I can say is I'm glad that we all had each other. I highly recommend the Walt Disney World College Program to anybody out there who is still in college because you just come out a better person, pretty changed.

Kevin Weitzel: Nice.

Greg Bray: Kevin, did you notice how he danced around that NDA that he signed, where he can't really tell us what.

Kevin Weitzel: He did. Yes. I can't tell you secrets about the underground tunnels or anything like that. Yeah, none of that.

Josh Paul: This conversation never happened.

Greg Bray: Well, Josh, let's talk about things we're allowed to talk about then. So, tell us just a little bit more about how Homebuilder Ops works with builders, about the services you guys get into related to HubSpot and helping builders there.

Josh Paul: Sure. Just to go through it briefly, about half the builders that come to us are implementing HubSpot for the first time, and they don't want to get it wrong, and they don't want to lose the momentum that they have or the buy-in from their sales team. They want to get it right from the beginning. They don't have 18 months or 24 [00:04:00] months to do trial and error. And then the other half of the builders have been using HubSpot already for six months or six years. They just know it's not set up right for their specific builder processes and team structure and industry. They're not getting the ROI they were promised when they originally signed on.

So, there's pretty much every best practice in terms of architecture and automation and solutions from a CRM sales enablement, customer experience, conversion rate optimization perspective is coming out of our team for those builders who are using HubSpot. And so, we're able to get people back on track.

We do a ton of training, a lot of sales training, a lot of executive coaching because, you know, there's a ton more data available, if the data can be trusted. A lot of these executives who have been in this industry for 20 or 30 years, aren't used to having available to them, so they need some interpretation. Here's what I should look at weekly. Here's what I should look at monthly. Here's what I should look at [00:05:00] daily, beyond traffic. We'll have another whole podcast episode about the problems with this industry's obsession with the traffic count.

Greg Bray: Well, today, Josh, what I wanted to drill down into a little bit more is this idea of lead scoring or AI lead scoring. I'm seeing this term pop up a lot. I'm sure other people are, too. But I'm also sure that it means different things to different people. But how do you explain or define the term, lead scoring and AI lead scoring, and then we can drill in a little bit deeper?

Josh Paul: I think we have to start with the traditional sense of lead scoring. There's different types of scoring you can do. You can do deal scoring that helps prioritize which deals your sales team should follow up with first, who's most likely to return your email or pick up the phone. Lead scoring happens at the front of the process. A lot of times it's lead scoring or contact scoring, when a lead first comes in, how much of a fit are they and how engaged are they? There's kind of those two [00:06:00] dimensions in traditional lead scoring.

And then also, you can score your existing contact database based on engagement, and that's where marketing and the website team get to kind of this is the wrong way to say it, lay the traps. That's not what we're trying to do here. But get to put out content and run campaigns with the intention of feeding that lead score to be able to tell who's engaged in your current database. They could have come in as a lead a year ago, and we haven't been able to connect with them. Then all of a sudden, their lead score starts cropping up because you're marketing more consistently, your website has the right structure to it so that you can start seeing them heat up in terms of engagement. Those are the people that deserve a call from your OSC team to try to find out if they're still in the market.

We run a lot of re-engagement campaigns, and there are a ton of people in builder databases that are still in the market, but because they weren't ready to go when they first came in as leads, nobody is talking to [00:07:00] them. And so, lead scoring helps bubble those people up over time who are ready and it allows you to do it at scale. So, that's lead scoring in general.

Kevin Weitzel: It allows you to know what carrots to dangle and when to dangle the carrot.

Josh Paul: Exactly, and who to dangle it to.

Kevin Weitzel: Yes. All right.

Josh Paul: Yeah. So, the AI side is really interesting. It doesn't replace traditional lead scoring. Traditional lead scoring is very metric based. So, we have who's been on the website recently, who's engaging with your marketing emails, are they engaging with specific pages on the website? So, based on their digital behavior, are they opening your sales emails? Are they clicking on the links that you're sending them and opening the materials that you're sending them? That doesn't change. But the AI component does two things. It allows you to take unstructured data and bring that into your score. So, customer sentiment or customer readiness, buyer readiness can be factored into your score. Combined with that unstructured data, [00:08:00] it helps you expand your pattern recognition in your lead score.

For instance, if you have a buyer readiness score, you can tell based on their digital behavior and what webpages they've been to and what forms they've filled out, how they came into your system, you can score them as a tire kicker. You can score them as someone who's just researching, somebody who's ready for a tour, somebody who's been to an appointment, and is researching after having been to an appointment. That's a really good sign.

So, if we can score people along that lines, combining that with your traditional lead scoring just super juices your lead score so that your sales team can prioritize who to reach out to. If a single OSC, is, they had 80, maybe 120 open leads right now that they're following up with, it would be great to be able to identify the 25 of them that are most likely ready for an appointment. It doesn't mean you shouldn't call everybody. It [00:09:00] means that you should prioritize those 25 first.

Greg Bray: Kevin, this may be the first time we've had the term super juiced used on the podcast.

Kevin Weitzel: I was going to say, I was going to say the same thing. I'm like, oh, my goodness, he just broke out with super juiced. Is your lead list super juiced?

Josh Paul: You guys, we're going to get T-shirts next time I see you.

Kevin Weitzel: I love it.

Greg Bray: T-shirts with a trademark symbol on them. Yeah, yeah.

Kevin Weitzel: We need to, man. Super juiced. Yes.

Greg Bray: Well, what does a super juiced lead look like, Josh? How does these scoring kind of appear to folks? You talked about traditional, you talked about AI. You talked about some different categories. Is this a number? Is it just a grouping? Is it super individual? Is it just putting them all in a bucket? What does this look like for the sales team when they're, like, logging in and trying to look at these leads?

Josh Paul: There's a few different ways to do it, but for sales teams, especially in this industry, you have to keep things very, very simple. So, we feed it into the existing lead score or deal score number. We don't want to complicate things, so we don't [00:10:00] say, "All right, you have your lead score, but now you have this buyer readiness score, and then you have this customer sentiment score. Make sense of it, sales team." We want to feed it all into one place and give them as few places in the CRM to go to as possible.

Again, that's another whole episode where sales people, especially in this industry, are not going to fall in love with the technology and with the CRM. So, we need to go into this recognizing that they want to be in and out of the CRM as quickly as possible, tracking what they need to track, but then getting back to what is their superpowers, which is building trust with buyers and ushering them through the process.

So, we want to boil it down to a score. It shouldn't be capped at 100. Like, a lot of people say, "Well, Matt, why is your lead score go up to 140 or 160?" Because there's always new activities happening. It actually mutes hot leads if you try to compress it all into a 100-based scale. But we need to train the sales team that says, "Hey, if [00:11:00] somebody is in the top third, middle third, or bottom third, this is what it means."

Greg Bray: So, Josh, you used the term non-structured or unstructured data a minute ago. What does that data look like when you mean structured versus unstructured, and why is AI better at that than maybe the traditional tools are?

Josh Paul: Let me give you a couple examples. I'm going to give you one on the lead scoring side and then one on the deal scoring side. So, we're going to start with deal scoring. When we're talking about unstructured data, we're talking about emails and notes and call transcripts and SMS messages, and being able to pull out key buying signals and summarize those patterns into a scoring model that you can then use. So, if the buyer is using certain language or asking certain questions or reacting to certain information in a specific way in a call transcript, that is unstructured data that AI can then, if you structure the [00:12:00] instructions right, to then normalize into a score. Now, there are problems with deal scoring. I think we're going to talk about those in a little while with data consistency from the sales team.

But let me talk about the example on the sales side. You guys have been in the website world for a long time. One of the things that we are recommending to builders now, is to add a questions box on every form. You know, first name, last name, email, phone, questions. Now, traditionally, we've been at this for a long time, the name of the game was reduce the number of properties on a form. It was kind of old school advice. Reduce the number of properties because for every property you add or every form field that you add, it reduces your conversion rate X amount.

But we're recommending that people add a question box, multi-line question box at the end of every form, even their short forms, because buyers want to tell their story. They want to give [00:13:00] you some information. They have a specific question or specific pain point that they have right now that they want answers to. So, if they can convey that to you, they don't have to keep it in their head for when the OSC actually reaches out to them and go, "Oh, yeah, that's what my question was." They can ask it in real time.

Now, what that does on the CRM side is allows you to take the specific question that they had, you don't have to wait until the OSC connected with them, but the specific question that they had, and you can use that in the score, but you can also formulate a hyper-personalized follow-up or initial sequence from the OSC team using the exact information that the buyer was looking for. It can be automated, it can be templatized, or it can be prompts to the OSC team, depending on how strong that data is. But that is unstructured data. That question can be three paragraphs long. It can be looking for pricing. We can formulate that into a score, and then also formulate into very personalized initial [00:14:00] outreach.

Greg Bray: Hey, everybody. This is Greg Bray from Blue Tangerine, and I am so excited to let you know that the registration is now open for the 2026 Builder Marketing Summit. We're gonna be in Dallas, Texas this year on September 23rd and 24th, and we are working on an amazing lineup of marketing, OSC, and leadership content for you.

Please check it out at buildermarketingsummit.com and get your registration in today. Remember, there's limited seats available, so don't miss out. Again, buildermarketingsummit.com. Can't wait to see you there.

So, Josh, this reminds me of a conversation I had recently, probably back at the Builder Show, but it was about the idea of balancing how many fields go on the form. Again, this idea, the more work we ask them to do, the less likely they are to finish it all. But actually moving that into a multi-step form. So, maybe the first page is just a few short questions, and then we say, "Oh, once we've captured that, now we've got their name, [00:15:00] it's in the system." But then we go ahead, "Hey, tell us a little bit more about what you're looking for or where you are in the process," and things like that. Have you experimented with those at all, where you see kind of the step one, step two, even step three in a form process?

Josh Paul: Absolutely. For your shorter forms, adding the question to step one, I wouldn't save that until afterwards. Also, you can score based on who filled out the question and who didn't, who actually has an actual comment or question that they have. But then that step two, like you said, step one's already captured in the system because if they don't fill out step two, you can still follow up with them like a traditional lead. But that follow-up information, you don't want to make it too cumbersome because you still want a high conversion rate on that second form, but that's where you ask timeframe and budget.

Because of AI now, you can ask specifics. You can ask things in broader text boxes. You know, what are you looking for in a home? You don't have to break it out into individual fields on the form anymore. You can [00:16:00] ask them to just blurt out their answer. We're also experimenting with audio capture, where on that step two, or sometimes even on the step one, somebody can just talk. Because a lot of times they're doing this from their phone. They click the button, and they can just talk through exactly what they're looking for.

Because there's an AI backend to it, it parses it into the right places in your CRM, and then no matter how they presented it, it formulates in a way that makes sense for the sales team that's going to be following up. But I highly recommend the two-step process. You have nothing to lose because your one-step form is still there, and so let's capture some additional information. But you also need a sales process on the backend that is going to use that information and not ask the same questions that they already took the time to give you.

Kevin Weitzel: Let me ask you this. The listening agents that are out there, you know, like when it's recording to a phone, does it put all the responses in all caps for people that come from New Jersey [00:17:00] because of their accent?

Josh Paul: It does.

Kevin Weitzel: In all reality, so these are all the what-ifs and the channeling that people should do to start adding this scoring to their process. What are mistakes you see builders making in trying to set all this up?

Josh Paul: The biggest weakness in traditional scoring and then it's amplified on the AI scoring side is missing context. So, let's say you want to score people based on, you know, traditional scoring, one of the questions on your form is timeframe. You have 10 forms on your website, and only five of them capture timeframe. You can't reliably use timeframe in your scoring model because you can't knock down people who don't have a timeframe just because you did not offer it. So, you need that consistency.

And so, we're actually seeing AI able to be leveraged better on the marketing side [00:18:00] than the sales side right now because there's less holes in the context window. And so, when somebody visits your website, that's pretty consistent. Everybody who visits your website is going to be captured. Marketing behavior, interaction with emails, form fills, because it's in the buyer's control, that is all very consistent. When you get into the sales side, there's missing notes, there's meetings that are happening that are not logged in the CRM. There are phone calls being made that are not in the CRM. And so, you can't use that data consistently, to follow up with people and to summarize next steps.

And this is where marketing and sales is kind of butting heads right now. Marketing says, "Our hands are tied. We can't move forward with this AI personalization initiative and this customer experience personalization that we want to do because we can't trust the data in the sales pipeline." There's just too many holes. We don't want to reach out to someone and [00:19:00] say, "Hey, it's been a little while since we spoke," or "We'd love to have you in for a tour," when they just came in three days ago for a tour, but it's not actually recorded in the system. And so, if there's significant holes, and if there's even small holes in your context window, it really hamstrings your ability to leverage AI scoring as well as just AI in the customer experience.

Greg Bray: Kevin, I'm dumbfounded. He's saying that there are salespeople that don't record everything in the CRM.

Kevin Weitzel: What?

Greg Bray: That's what I'm hearing.

Kevin Weitzel: No way.

Josh Paul: Are you going to have the breaking news banner across this podcast?

Greg Bray: It is hard to get people to take the time to put in the experience. Is there a secret, Josh, to help get this data collected?

Josh Paul: It's our bread and butter. What you're pointing out is just an age-old adoption challenge in a lot of industries, but especially this industry. But the implications of not having that data in the system have grown 10X in the past two [00:20:00] years. I think that's something that executives at home builders need to understand that this isn't just a, "Oh, our reporting's going to have some holes in it. We'll just report through spreadsheets again." While the industry and while competitors are getting this right and able to take advantage of AI, the builders that have holes in their context window, it's going to be a failure to launch over and over again until they fix that.

Greg Bray: I think your example there, Josh, though, is really eye-opening for someone who may not have really thought about it. It's like here we are, we want to be more personal, right? We're going to use AI to reach out and make these emails feel like they come to someone and we say, "Oh, yes, we know you were here three days ago," or, "You haven't been here for three months," or whatever, and the message will be personalized based on that information.

But if that information's wrong, it's going to actually embarrass us in front of the customer, and actually make it worse, you know, because of the whole, [00:21:00] "Well, you didn't even know I came in last week. What's wrong here?" You know, and it's going to be this total dissonance in the experience that's going to actually do more harm than help, I think. Are you getting people to recognize that and understand that?

Josh Paul: We are, but it's a big problem to solve, and it takes quite a bit of strong leadership at the builder level. What it means right now is there's AI initiatives that are a really good fit and that are on the launch pad that have not gotten the green light because the data's not there. So, people recognize it, but it's going to take time for the training and adoption programs to be in place.

Greg Bray: What about data that's not quite so dependent on the sales team? Is there more that we can gather on the website kind of automatically that maybe hasn't been traditionally done that would also influence and help beyond just the fields on the form? AAny thoughts there on what else we could capture to help?

Josh Paul: [00:22:00] Absolutely. We should and can capture everything, what pages they've been on, how long they've been on those pages, the content that is on those pages. What we're trying to do is get inside the buyer's head so that we can get the right message to the right people at the right time. This isn't necessarily a scoring fit and engagement discussion now, it's a buyer profiling conversation. So, based on the pages, what is on those pages, how long they've been on those pages, what type of media that they're consuming, all on your owned website, you have all of that data.

And so based on that, you can structure your prompts and your AI instructions to be able to profile people. We can tell this is an empty nester, this is a first-time home buyer. So that when they do actually fill out a form, we know so much more about them, based on their behavioral profile than they gave us, and we don't have to rely on us chasing them down for a phone conversation. We can put them in the right bucket based [00:23:00] on their digital profile. Are you seeing that as well? Do you think that's the biggest opportunity?

Greg Bray: I think there's absolutely opportunity there, and as I'm listening to you and kind of thinking this through in real time here, I wonder too, are there benefits to even getting our pages more focused. Instead of having one long page that talks about lots of different things, having separate pages so we can identify what was really their interest in that particular topic. Was it topic A, B, or C? When they're all on one page, we really don't know. But if we kind of break that out. Any thoughts along those lines?

Josh Paul: Yeah, what you're talking about is a shift in how builders have done things before. We're working with a big builder right now and building an AI chatbot based on HubSpot's customer agent feature. What we're finding is that a lot of today's websites are not structured in a way that AI is able to consume and make sense of.

And so, builders in general need to produce more content, not [00:24:00] only for their buyers at every stage in the process, but they kind of need to create this shadow website or this content structure that is able to feed AI. That's short answers, targeted questions, kind of FAQ style elements to the website and specific pages. Not only for what you talked about to be able to profile somebody more efficiently, but also to feed your owned AI agents as well as the large language models. It's a shift in how every industry is structuring their websites.

Greg Bray: So, if a builder's listening to this and they're like, "Man, I think our stuff's a hot mess." What is the starting point? Walk them off the ledge a little bit where they're not going to go, "Oh, this is hopeless." Where do you suggest they start?

Josh Paul: You have to raise it up a level, and they have to get a handle on their funnel metrics to identify where they should start. Is it a getting found challenge that is their top priority? Is it a lead conversion challenge? Is it a [00:25:00] lead-to-appointment conversion rate issue? Is it an appointment to contract issue? So, getting a handle on their full funnel metrics in an accurate way is going to tell them where they need to put their time, energy, and money.

If it is, let's say, a lead generation challenge, then there's specific tactics for converting more website visitors into leads. Some of them are AI-based, some of them can be personalization. I've heard you speak about profiling buyers and rendering websites in a personalized way before the buyer even fills out a form based on that free form fill profile. AI is playing a large role in that.

But let's say it's a lead to appointment challenge. That's where the hyper-personalized outreach can come into play for the OSC team and then for the lead nurturing and re-engagement campaigns. So, in terms of where to start, don't be overwhelmed by everything. We need to take a breath, get back to basics of getting a [00:26:00] handle on our funnel, being able to trust our data, and then identifying which tactics to test at each conversion rate to be able to move the needle.

Greg Bray: Josh, when you've seen builders do this well, who in the organization is kind of leading the charge? Is it coming from sales? Is it coming from marketing? Is it somewhere else? What seems to be the area that is able to make this work?

Josh Paul: That's a really interesting question. I don't think the lines are drawn the way they used to be drawn. The person in the organization who gets it and can articulate it and is clear-eyed about, what is needed from an adoption and data hygiene perspective, and the pace in which things can be tested, that's the person that is going to get things done. Sometimes it's the CEO, COO. Sometimes it is VP of marketing. Sometimes it is the sales leader. But somebody who understands that leveraging AI like this is a different beast. The timelines are [00:27:00] longer, the testing is longer because you're no longer in control of every word that goes out, but someone who recognizes the potential but also is very clear-eyed about the context window that is needed and the integrity of that context window to be able to scale this.

Kevin Weitzel: Should builders keep an eye out for that OSC that's super juiced that uses this platform in such a manner that they need to be considered for somebody that oversees said platform?

Josh Paul: We're not seeing it as much. OSC teams are the most structured teams at builders. We're not seeing as much entrepreneurial, going outside of the lines, leveraging different tools by the OSC team. They're pretty locked down. They're very good at what they do.

Kevin Weitzel: But not super juiced?

Josh Paul: With the right system, they can bring the super juice.

Kevin Weitzel: Boom. You hear that OSCs? Bring the super juice.

Josh Paul: People who are able [00:28:00] to get even a small part of their data accurate, that's what thing you lean into first. So, different builders are starting at different places.

Greg Bray: Well, Josh, it's been a great conversation. I'm sure, like you said, there's some spinoffs we could probably get into at some point in the future. But right now, any kind of last words or thoughts of advice in this area to leave with our listeners today?

Josh Paul: The biggest focus right now is whoever gets their sales data accurate is going to be able to accelerate the way that they take market share. And so, investing in training and adoption programs, investing in simplifying their sales tools and their CRM environment and making sure that this is easy for sales to get in and out of, track the basics, and being able to point to how this is going to make sales' lives easier to actually show the sales team. The last component of that is leadership, is holding people accountable, [00:29:00] educating them on why it's important to track things at the most basic level, because it not only impacts reporting like it used to, it's going to tie the hands of marketing and other customer experience groups if the sales data is not captured easily in the system.

Greg Bray: Well, Josh, if somebody wants to connect with you, what's the best way for them to get in touch?

Josh Paul: You can go to homebuilderops.com. That's homebuilder O-P-S dot com, or you can find me on LinkedIn, Joshua Paul or Josh Paul, Homebuilder Ops. I'm happy to connect with any of you.

Greg Bray: Well, thanks again, Josh, for sharing your thoughts with us today, and thank you everybody for listening to the Builder Marketing Podcast. I'm Greg Bray with Blue Tangerine.

Kevin Weitzel: And I'm Kevin Weitzel with OutHouse. Thank you [00:30:00]