Show Notes
Scott walks through a live build session with Chad Coffman, a land investor whose lead system broke when volume spiked. Instead of stopping production or ignoring the problem, Chad triaged it—kept moving, documented the friction, and circled back when he had time.
The friction: Leads were coming in faster than they could respond. Responses were delayed by weeks. The process was manual, inconsistent, and overwhelming.
Applying SCALE:
S — Scope the Solution: Not the entire sales process. Just the initial lead response. One play, not the whole game.
C — Clarify the Flow:
- Email comes in (trigger)
- AI triages the email
- AI pulls from knowledge base (county rules, property info, zoning)
- AI drafts response
- Human reviews and approves
- Response sent via customer's preferred method (email, text)
- Human uses newfound time to strengthen the knowledge base
A — Automate the Trigger: Four types of triggers: event, time, condition, manual. Manual is the worst—requires memory. In this case, the trigger is an event: lead submission.
L — Leverage the Data: Plan for failure. How do you know if the email didn't arrive? How do you know if AI failed? Build in regular human checks. Start with humans overseeing, then automate the oversight later—that's a separate play.
E — Elevate the Experience: AI should sound like Chad and Cindy, not a robot. Build a voice guide. Create feedback loops so AI improves over time. Make sure error messages are human-readable, not "Signal 19."
Key insights:
- "We're not trying to boil the ocean. We're running one play."
- "The time you gain from automation is shifted—use it to strengthen the knowledge base."
- "If you just threw a person into your business with no training, that's what throwing AI at something looks like."
- "There will be a competitive advantage to dealing with a human instead of a robot."
The 40-minute investment: Planning the framework before building saves you from building the wrong thing.
Your action: Pick one friction. Apply SCALE. Scope it to one play. Clarify the flow before you touch any tools.
Got a business question? Ask Scott here: scotttodd.net/ask
📜 Full Transcript (Click to expand)
Scott Todd (00:01.062)
All right, Chad, how you doing today?
Chad Coffman (00:03.554)
Good, good morning, Scott. How are you doing?
Scott Todd (00:06.585)
Very good, thank you. All right, Chad, so the reason that I encourage you to be on here with me today is to do a build session, okay? Because it came to my attention that back in February, you and your wife who own a company ran into a lead problem that blew up your system, right? So tell me about
Chad Coffman (00:31.256)
Yes, correct.
Chad Coffman (00:36.332)
Yeah, so we had started a land flipping business. We buy and sell land. And we had been advertising through a variety of channels, had some challenges. We started using landmoto.com and had a tremendous amount of success very quickly. And it kind of caught us off guard. We had been kind of preparing for things, but when it all started coming in,
we started realizing all these challenges, bottlenecks and things that we were doing that we would probably prefer not to do to kind of move our, to keep our leads moving, flowing through our systems. yeah, exciting problem to have, but a problem nevertheless.
Scott Todd (01:16.303)
Yeah, and if I hear it correctly, it literally, to me, leads up, your system, right? Like you guys couldn't keep up, you were getting stuff done weeks and weeks later, which is, as you pointed out, is a great problem to have, right?
Chad Coffman (01:21.667)
Yes.
Chad Coffman (01:28.035)
Yep.
Chad Coffman (01:31.394)
Yeah, great problem, but still a problem.
Scott Todd (01:34.191)
So, you what I love about what you guys did is you identified the problem, okay? And you didn't stop and try to fix the problem right then and there. You triaged it, right? Like you went out and you solved the problem. You identified it. You continued to just work with the system that you had. And then as you got time back,
Chad Coffman (01:46.317)
Right.
Chad Coffman (01:57.356)
Yep.
Scott Todd (02:02.785)
it was identified as a key priority for you guys of a vital need, then you're now able to circle back and think about, okay, well, how do we fix this the right way?
Chad Coffman (02:12.96)
Exactly.
Scott Todd (02:15.043)
Yeah. And the one thing I see business owners do a lot of times is number one, they might just stop production right then and there and say, Hey, I got to solve this problem first. Right? Like that deep down inside that sometimes feels like the right thing to do. But the proper way to do it is like what you guys did, which is, okay, noted triage it. We're going to, we're going to limp along. Okay. We're going to limp along. And then when time allows, we're going to circle back and come
choose the problem or attack the problem, right?
Chad Coffman (02:48.384)
Exactly, exactly. And having a way to document the problems as you go is really important, obviously, so you don't lose track of the pain that you felt. When you get past these things, sometimes you can forget and you move on to the next thing. And then if you're constantly trying to solve as you go, sometimes you don't do the highest priority things. You just do what's right in front of you. And you lose that ability to step back and look at the system as a whole and say, what do need to fix next to really make our business better and sort of progress us holistically?
Scott Todd (03:16.696)
Yeah. And you know, I think what you just said is so key. You have to have a way of capturing. I call it capturing the frictions, right? Because if you identify each of these frictions and you have some methodology in which you capture them, which it can be, you know, it can be very analog, which is just on paper. Okay. Just take out a piece of paper and write down all of the frictions that you see in a given day.
Chad Coffman (03:37.816)
Mm-hmm.
Scott Todd (03:44.036)
The other way of doing it is you can get more technical and do things like in a cam-bomb board where you're looking at things and you're like, oh, let's just track all of these things. The way that you do it is really up to you. How do you do it? Tell me how you do it.
Chad Coffman (04:02.702)
So I happen to come from a pretty significant process improvement background. I have Lean and Six Sigma experience. I really love Lean tools. I'm sort of a student of process improvement, and I use technology to do that. So I know what it means to map the process first and then take off after the problem. So we had a very simple swim lane diagram for our processes. When we started running into problems, I created a value stream map, and I started capturing all the specific things that we do and looking at values.
add, non-value add time and the tools we're using and the know just the handoffs in the process. So I happen to have that background so and I have some tool sets at my disposal but the most important thing usually to me is that when something happens you're just capturing the pain so at least you capture the pain and then later you can go back and build out the nice value stream map and things like that and I can do that quickly so it's not a it's not a real problem but I think it's again the most important thing is to capture the pain and whatever your tool set you can go back and figure that out later.
Scott Todd (05:03.491)
Yeah, you one of the things that, because I find the way that you're working interesting. One of the things that I do is I use a tool called Obsidian. It's like a notion. It's, you know, it's a note system. I like Obsidian. But what I'll do in Obsidian is I will throughout my day, because I have a daily report there where I'm making notes or tasks that I've done completed or things that I see. I actually go in there and I create
a friction. I actually write, you know, like an open bracket friction, double colon, that way I can have data views and whatever. You know, I'm not going to go down that route. But I think that the most important part is that you capture it, okay? Because if you're not capturing it, you're going to forget it. And I love what you said. It's really about, hey, let's circle back, find the highest value obstacle that we can remove, and then we remove it, and then we go back and we
Chad Coffman (05:42.658)
Yep. Yeah.
Scott Todd (06:03.552)
we solve that problem once and for all, right? Or at least we think, right, yeah.
Chad Coffman (06:05.758)
Absolutely. Yeah, exactly.
Scott Todd (06:09.898)
So the friction that you identified was kind of the leads process was broken. That's kind of what I'm hearing, the lead process was broken. So that's what I would capture when I circle back to it. I'm looking at this. And then the way that I do it is I actually go through the scale framework and I'll kind of introduce the scale framework to you, but the scale framework is essentially five letters.
They're five acronym. It's one acronym, five letters. And it basically says scale. We're going to start with the scope because Chad, as you know, you could easily over engineer something. You can try to try to, you know, build out the entire system for something. But really the way I is it's a play. It's from here to here. Right. We're not trying to go from one end of the of the football field to the other. We're just trying to run one play.
Okay, so we want to make sure we're scoping it correctly. Then we're going to clarify the flow. And what I mean by that is we're going to map out what triggers it, what comes in, and what the intended goal of this particular play is. The A is we're going to automate the trigger. So in this case, a lead coming in, the trigger is the lead generation, somebody submitting a lead to you. The L
Chad Coffman (07:20.27)
Exactly.
Scott Todd (07:33.298)
is leverage the data. We want to know that the process worked or it didn't work, that our play worked or didn't work. And then E is elevate the experience. And what I mean by that is one, if we're going to have something communicating in a form letter kind of style to a customer, we want to make sure that it doesn't sound robotic. We want to make sure that it sounds human. sounds like chat, right? The other is that if there's an error,
Chad Coffman (07:59.852)
Yeah, exactly.
Scott Todd (08:03.212)
We want to make sure that it's not some system lingo like signal 19 error. What is that? Right? OK. That's the way I'm thinking about it. And what we'll do is we'll just start off here. And we'll start off with the scope, the S. So how do you see the scope?
Chad Coffman (08:10.51)
you
Chad Coffman (08:26.626)
That's interesting. I think it's initial lead. So we had had a totally different workflow and we moved to having leads coming in through the site land moto that you may be familiar with. And so we would see a lead come in and we didn't really have a good way to kind of start our whole, we didn't have a plan for how we reach out to customers and our potential customers leads really, and kind of go through the journey with them. So, as we started reaching out, we got different questions
They would ask, you know, then we realized we need some data that we need to gather for that everybody's asking for let's figure out how we do that and then we had repeated questions that were answering over and over again So the first two or three things we say to somebody is pretty routine There's really no value add and like the the interactions we're having they're just trying to get it kind of get the relationship up and running so and then you know as we got into the process there were you know things we needed to do to
to close a transaction with our customers. They might, you we have to potentially generate a deed or generate a note or, you know, talk just some of the conversations about financing and things like that. But a lot of it was very focused on, you know, we have counties that we work in and they have rules and regulations and so people are asking a lot of routine questions. And that was really a lot of it is helping people understand what they're looking at, if they're looking at buying land and, you know, what some of the limitations
and what opportunities they have when they purchase a particular piece of land.
Scott Todd (10:01.064)
Yeah. So if we go back and we look at something like this, really the scope that I'm hearing you say is it's really about the initial lead response, right? It's not trying to craft an entire sales message. Hey, we get the lead in.
Chad Coffman (10:11.514)
Exactly. Yes.
Scott Todd (10:18.674)
What is the first thing that we want to do? And I think that that's, you know, we're not trying to boil the ocean here. All we're trying to do is this one friction point because then we can daisy chain it into multiple sequence of plays. But for now, what we're going to look at today is that initial lead response, right?
Chad Coffman (10:25.816)
Right.
Chad Coffman (10:40.698)
Yeah, that's exactly right. The real challenge for us was getting somebody qualified to like that we're a good fit for them, they're a good fit for the land that we have to buy and getting them on the phone. And that way we can have, know, kind of do some initial, I guess you'd call it filtering. And then so when we get them on the phone, that it's a really good conversation. That was really the scope of what we were trying to do.
Scott Todd (11:06.484)
okay. So, so when we talk about clarifying the flow here, we know that the, that your initial trigger, if you will, is going to be an email, right? Because it's probably going to come in from an email. And then we want to think about the, that desire, try to see if I can make this here. How can I do this? Let's see how I make this.
look at that. Yeah, there we go. Now, and I can go out a little bit more. Okay, so we have this initial goal of getting the email. And then the end goal is not necessarily to make the sale. I know that's what you want. You want to make the sale. But really the end goal in this case is just a response back to the person who's inquiring, right? It's that initial.
Chad Coffman (12:04.257)
Exactly.
Scott Todd (12:06.527)
So.
you know, whatever it is. And so this email trigger may contain questions, it may contain certain things. So what we're just do here is we're just trying to get it to where we get that response back out to them. And as we start to think about how that happens, there's lots of nuances that could happen because, you know, if you wanted it to be very cold,
Well, then it would just be a robotic answer that says, hi, we have your lead. How can we help you? Or hi, we help you. But I think you want to go further than that, right? Like you want to do something different.
Chad Coffman (12:40.812)
Right.
Chad Coffman (12:46.914)
Yeah, mean, the first response is really important. The timing of it is really important, and that's one of the things that creates urgency and sort of made us drop what we were doing and start to deal with this. But I think the quality of that interaction is incredibly important. It sets the whole tone for how you're working with someone. It lets them know who you are and what their whole experience with us is going to be. So the tone and quality of that is really important.
Scott Todd (13:12.019)
So do you envision like having AI write those based on the initial response?
Chad Coffman (13:18.254)
What's really funny is
One of the things that we started doing was actually using AI to write our responses, but we were taking what they would ask, type it in, get a response, send it, you know, put it back to them. And so as we were actually kind of doing it, was like we're doing, you because we had so many coming in, a lot of times they might ask a specific question and we had a number of crafted responses. And at first it was copy paste some things and then it was, okay, let's use AI to tailor it. And so we were, you know, again, you're moving so fast sometimes and you got a lot of things going on. Somebody asked a very
specific question and we pull in maybe specific county information or other information that we'd already again we ended up loading it to to AI so we were starting to do go down that road but it was still very manual with us in the loop.
Scott Todd (14:04.275)
Yeah. So, you know, where we are with technology today, technology is changing every single day. was not available yesterday is available today. And that's why we never really want to jump to the actual solution because tomorrow you might get tools, shiny object syndrome and want to go chase, never have anything that works. So we try to stay like, hey, we want to know what this flow looks like.
Chad Coffman (14:14.008)
Yeah.
Scott Todd (14:34.493)
And then we can start to layer in how the tools might work, right? And so really when I do a build session like this, I'm just thinking about it like a blueprint, you know, like a blueprint for a builder who's going to go build a home, right? Because we know what the framing looks like. Now what we need to do is we need to actually get some boots on the ground and go, and then figure out like how we're going to put that in there, right? So as we think this through, the email comes in.
And ultimately what we want to do is we want to have, you know, like AI triage that email. Is that how you? Yeah. So we want to have AI, whatever AI model you want or whatever you think would work today. You want to have that triage it. And while it's doing that, it could potentially pull knowledge
Chad Coffman (15:13.93)
Exactly. Yep. Exactly.
Scott Todd (15:31.698)
I'll just put this over here, it could pull knowledge from your internal records on rules of that county, right? Rules of the county, property information, all of that stuff. And I think that the approach that you're doing there where if you stop and say, hey, listen, here's the counties that we're working in, here's the zoning rules,
Here's the rules that we understand, whether, I mean, some properties might be in a property owners association where they have their own deed restrictions. Taking all of that stuff and providing that into a knowledge bank that the AI can pull from, now it can look at that question and then it can turn around and create their response, right? So.
Chad Coffman (16:25.56)
Exactly.
Scott Todd (16:27.101)
Then the question is, would you have a human look at the response or are you just letting it go?
Chad Coffman (16:34.05)
Well, so this is what's interesting. I think that the important thing for us is that the responses are, whether automated or human, that they are very valuable for the folks that are getting them and that they don't feel robotic, that they are, again, truly like what someone would expect and maybe even exceeding their expectations if possible. That's kind of our philosophy across the board. So I think what we like to do is say, if we can use AI to do something and it can exceed expectations,
or at least kind of like, you know, meet them where they're at, you know, where they're at and kind of give them what they need, then that's a plus. Let's move forward with that. And if it can't, then let's figure out how we tune that and optimize as we move forward. And so a lot of times it's, you know, us sort of, you know, walking through things with AI and curating the responses. Then when we see that they get good enough and we've had enough information coming in that the responses can be, you know, very good and very quick, then I think we're ready to release that piece to
to allow AI to do the work. But we're a little conservative about things. Like we wanna make sure that, again, the experience for folks is really good, especially in the early stages. But that's kind of an approach throughout our journey. that was one of the things that was hard for me. My immediate thought was, me go gather a bunch of information. Let me create this knowledge base. Let me do all this stuff. And getting into the solutioning side where I had to just kind of accept that Cindy is using an AI.
Scott Todd (17:39.356)
.
Chad Coffman (18:03.92)
that's, you know, she's using maybe chat GPT and I'm using Claude and we're using different ones and yeah, the data is not all centralized and it's not organized but we're figuring this out as we go and so that was kind of our approach.
Scott Todd (18:16.702)
Yeah, OK. And so this is really where
you you start to think about and I think that everybody should think about because the initial gut feeling that people have is they just want it done, right? And so they want to triage it. They want that AI to draft the response and that produces the response right here. But at the end of the day, really, I think that we have to always remember that we need the human in the loop.
Chad Coffman (18:31.502)
All right.
Scott Todd (18:49.34)
in some way. it's one of the things that drives me crazy is when I hear people that are like, I have no employees. AI or you see in the news where companies are laying off all these employees and they're citing, we don't need them anymore because of AI. Well, maybe you need less people. get that part. Maybe you need less people. But at end of the day, they're
Chad Coffman (19:11.799)
you
Scott Todd (19:19.024)
there will be a competitive advantage to being able to deal with a human as opposed to a robot. I mean, that's already the case today. I mean, if you think about it, if you call up to any major company and you go through their voice response system, you know, like how many times are you like, just give me a human, I want a human, okay? And you you keep hearing, check out our website.
If I could do this on your website, I wouldn't be calling. Trust me. I can the on you. But if I'm calling you, I need you to answer the phone. I think it's the same. There's going to be a premium for that human experience. And it will be a competitive advantage that we don't just want to take and have the bots of the world just communicating everything. And so I like the fact that you're keeping the human there.
Chad Coffman (19:50.446)
Exactly.
Chad Coffman (19:57.965)
Right.
Scott Todd (20:16.859)
So now we keep that human in the loop to make sure that we're getting the response that we want, that it sounds like us, all of that good stuff. And then what? So like the human looks at it and says, yeah, I'm good to go. But now what are we doing? Are we just sending it? Or is there something else that we should be doing here?
Chad Coffman (20:38.734)
I mean, for us, was just, if it's an email response, we would send that to them. If it's a text response, if they wanted to be, use text messaging, which a lot of people do, we would just start texting them. And so we would craft the response, make sure it makes sense, send it out to them. And there was usually a little bit of back and forth. It wasn't usually just one message. It was usually maybe two or three fairly standard messages. And our goal is to get them on the phone. That's like the first thing. But there tends to be, like most people,
want to text or email back and forth a little bit and sort of get a feel for things before they get on the phone.
Scott Todd (21:14.456)
Yeah, I think that that's the other piece too is like, how do you want to communicate? How, you know, like that's, that's one of the early like dances. I think that you have to do with lead generation as you have to figure out, okay, like, Hey, you want to be where they want to be. And what you don't want to do is you don't want to say, well, we only communicate through email, right? Because I mean, you're there's people that communicate in many different ways. You need to show up where they are.
the format that they want to show up with, right? I think that's a good key point. So I'm going to put in here, we're really meeting them where they are in terms of communication method. the method is there, so it's not always an email. But we have that communication method, which leads us to our response, our initial response.
Chad Coffman (21:48.662)
Absolutely.
Chad Coffman (21:58.912)
Mm-hmm. Yep.
Scott Todd (22:10.552)
you know, in doing this where you have the AI triaging something and drafting it based on the knowledge that it has, really the time that you gain in doing this is really just shifted. And this is where I think that people lose sight of the human in the loop is that, you want the human in the loop to figure out where, not only where it's going wrong, yeah, if it's correct, just ship it.
But if it's going wrong, this is where we want to pause and strengthen back up the knowledge base.
Chad Coffman (22:44.684)
Yes, exactly.
Scott Todd (22:47.214)
How many times do people miss that?
Chad Coffman (22:49.526)
Yeah, a lot.
Scott Todd (22:51.928)
answer the AI, but they don't ever stop with the new found time and then go back and say, Hey, here, let's, let's, let's firm this up. Because Chad, mean, if you're working in the same County for years to come, it's really an investment of like that regulation's not necessarily going to change overnight, right? Like it's not, it's something that you and it's going to last you for years. Right. Yeah.
Chad Coffman (23:19.2)
Exactly.
Scott Todd (23:20.858)
put in the institutional knowledge. Okay. So we've clarified the flow. We kind of have an idea of what we want to do here. And then what we want to think about next is how we're going to automate the trigger. And what I mean by that is automate the trigger. And what I mean is that so many times, really there's four triggers in the world. There's four triggers that trigger everything. Number one is
You know, an event, okay. An event is a, is basically something that's, that's happened. could be a calendar event. could be, you know, somebody submits something that's an event. We can have time as a trigger. Okay. Time, time is a trigger. Monday at 8 AM is a trigger. An event is a trigger. A condition like, Hey, it's going to be 90 degrees outside today and it's going to rain. Well then.
Chad Coffman (24:14.371)
Mm-hmm.
Scott Todd (24:20.309)
send me a message that tells me this to bring my umbrella, that's a condition, you know, and the last one is manual. And so many times, it's easier to build those manual triggers, but the reality is that a trigger, a manual trigger is the worst trigger because it requires somebody to remember and we're going to forget, right? So, I think that the approach that you're having there in terms of
Chad Coffman (24:24.502)
Exactly.
Chad Coffman (24:41.816)
Yes. Yep.
Scott Todd (24:49.536)
the event being the submission is, I mean, it's logical in this case, right? Without it, you're relying on an employee to go into their email and check for emails, and you're counting on them to go do it, which they may not do. So clearly, what we're going to do here is the event. The trigger is the event. It's the submission or the receipt of that email is what we're going to do there. All right. Leverage the data.
Chad Coffman (24:53.955)
Yep.
Scott Todd (25:18.137)
This is where we're going to plan failure. And what I mean by that is, where can this process fail? It can fail because you may not get an email, which is your trigger. That's a potential. mean, look, email is email, right? So how do you know that you don't get an email, right? How would you know if you didn't get email? You just don't know. It's just not there. But this is where like a re-
Chad Coffman (25:34.915)
Yeah.
Chad Coffman (25:41.091)
Right.
Scott Todd (25:47.902)
regular check of whatever system sends that email is going to be your friend. Because if you have a regular, know, automated task that says, hey, let me go up there and schedule, let me just go look to make sure that we're getting everything we need. And then we can alert it. You know, how do you know that the AI failed to trigger off of this? And, you know, even humans, if you're going to outsource this or provide it to somebody else.
How do know that the human is not the bottleneck now? And that's the way that we want to think about this. Each step along the way, we want to think about there's data being produced on each individual step. And we want to know if something's not right. We want to know if something's not there. So what are your thinking behind that?
Chad Coffman (26:40.396)
Yeah, so this is really important to me. I think when you start to automate a process, like you said, you have to keep humans in the loop. I think the key that makes this work is that you have a structure for how the humans are involved. So I like to create cues and work tasks and things like that so that people don't have to remember something. They have something that becomes not just a routine, but it becomes something where they can be prompted in some way. So in this case, it would be, instead of just checking Landmoto constantly,
you check it maybe twice a day or once a day just to make sure nobody slipped through the cracks. And then it's scheduled, it's routine, it's very calm, it's very much just something you do. And you don't have to constantly like think, my gosh, did I forget? my gosh, did a new lead come in? So that's kind of what I see in that particular case is once a day you check, you know, Landmoto, make sure you didn't miss anybody, that an email, you know, didn't come out, you know, go astray or something like that. And then, you know, hopefully you can begin to trust that
process over time, again, regular checks are an important thing. And if they're scheduled, they don't take much. And if you schedule a number of tasks, then you've got a really set thing that you can do and your life is very relaxed and you have a very predictable flow to your day.
Scott Todd (27:55.443)
Yeah, you know, and see, this is where it's easy to over automate or over engineer something to catch at because like as you're saying that, the thing that's going through my mind is, yeah, we could have a human go and check this website to see if all the leads are accounted for. Or we could build, you know, an AI automation that logs in and checks to make sure that all the leads are accounted for.
Chad Coffman (28:20.172)
Yeah.
Scott Todd (28:24.416)
But see, that's the risk. Okay, like that is where so many people go down this rabbit hole of automation because the easiest thing to do today is the human in the loop and that's fine. Okay, because as you build this out, in my opinion, what we're doing is we're staying within this framework that we've identified. We're staying true to this scope.
And the fact that we were potentially saying, we want an automation that goes and validates this. That's a separate play. Right. Like that's a separate whole build session. That is a separate, that is a completely separate conversation. That's not for today, but yet it links to today. Right.
Chad Coffman (28:58.87)
Right, exactly.
Chad Coffman (29:13.656)
Right.
Well, I think what's interesting about that too is the concept of plays. you, and you kind of mentioned like, you know, AI is changing jobs, but I think, it's, it's changing the workforce, but it's, really changing. I think in some cases you might downscale, you know, your workforce and, you know, not have to have as many people, but in a lot of cases you want to shift what they're doing. And I think this is an important piece of this. So for example, we talked about the knowledge base and how that can, you know, help AI get better and everything like that. If you have a human that once a day goes in check,
and says, know, did everything go out? Well, that's probably very fast. And generally speaking, it's going to work. You know, it's just a quick check. But if you can also make that a checkpoint to say, let's review a few of the messages and make sure that they're getting proper responses. And let's identify if maybe there's some opportunities to shore up our knowledge base. Then you've got to like a micro play that you do instead of, you know, a macro play. And you can just check your knowledge base, start to tune that up. And you can use that as kind of a trigger for really enriching your knowledge base.
Scott Todd (30:06.87)
Thank you.
Chad Coffman (30:14.608)
as you're sort of checking to make sure that operations are moving smoothly. So I like to kind of combine things and think of these new plays and then again you can go back and automate later. But if you try to do all of it at once it can get very overwhelming and I think your systems can get very convoluted.
Scott Todd (30:21.174)
It.
Scott Todd (30:30.602)
Yeah. You know, it reminds me as you're talking about the micro place. When I started building my business, you know, 12 years ago, and the technology wasn't anywhere close to what it is today, we relied heavily on, you know, humans and VAs. And the problem that I kept seeing people run into when they were trying to build a VA team is that they would just take
all of this work and be like here. what we said is, hey, wait, wait, wait. If you're going to give an assignment to somebody, what's the smallest microstop? I think in microstops, right? Like here, let's go build this one little piece that I want them to go do. And then they go do this over and over and over again. And then it's like, okay, you've done part one.
Chad Coffman (31:01.772)
Right?
Chad Coffman (31:23.758)
Mm-hmm.
Scott Todd (31:26.495)
but yet there's another nine parts to go. So now let's work on two and three now, right? So now you've got them, like they're launching stronger than they go learn two and three, and then they go learn all the rest of the steps. And I slowly would hand the work off to them. And it's the same way with automation because we can't try to automate the entire system. I mean, you could, it's not gonna work. You're not gonna be happy with it.
Chad Coffman (31:54.638)
No, and think one of the things that's really important is that even if you are building a system, you have to design that system. so if you're teaching the same way you would teach people to learn a new process and a new system, and especially with AI, you have that opportunity to use that same thinking, but you still to be methodical and you have to plan that out. And you may not know what the instructions are right in the beginning, and you may not know what's going to work with a particular person or particular AI. So I think you have to be very methodical and have a way that
Again, you roll that out. I love the way you think about that.
Scott Todd (32:27.615)
Yeah. So for leveraging the data, we're going to say, hey, we're going to build in some of those checks, but we're also going to have humans overseeing some of the processes in the beginning before we can later then go back and further automate the humans overseeing it to go do the right. So we're going to have a human oversee it today because that's the micro step. That's the easiest way to go. And then later on, as we look at new frictions that arrive in the business,
The new frictions might be these humans that are overseeing some of this automated work that we can go build the automation, right? And then back down to elevating the experience. And this is the voice. And one of the things I think that you mentioned that you were doing that I liked that you said that you were doing is that you're having AI kind of draft or triage that message, draft the response for the human to go back and look to make sure, is this the messaging that we want? Does this sound like us? So you know, you're...
Chad Coffman (33:03.766)
Exactly.
Scott Todd (33:25.714)
you're kind of doing that, but we all know that, well, I shouldn't say we all know, it is pretty common knowledge that, you can have AI kind of build a voice guide for you, learn from your own writing. That's really two years ago, that would have been wild two years ago, right? know, today it's like, I want it to sound more like me, not like a robot. And we have that part, but what are your thoughts on that?
Chad Coffman (33:54.189)
Ask me that question again. I hear a couple things I wanted to respond to but
Scott Todd (33:59.454)
When you're thinking about elevating the experience and you're going to have AI out there and draft that response, how are you number one, finding or helping AI to find your voice? Okay. Cause you want to part. And then the second part is, is really how, how are you going to elevate the experience or how do you think you're going to elevate the experience for the, for the customer and for you to know that something's wrong?
Chad Coffman (34:10.392)
Yes.
Chad Coffman (34:26.094)
Yeah, I think the key is that...
You know, for an AI to find your voice, you really have to have like a body of responses and things that you've, know, ways that you talk, you respond, information you provide in a given context, things like that. And I think that's kind of a skillset all on its own. It's almost, in many ways, it is a lot like teaching a person. You you want to show them, maybe you have recordings of calls you've made, or you have, you know, recording of something you do so they can see how you
it and you have templates maybe of things that you've responded to in the past you give it to a person that person reads it and says oh okay I get what they're saying I understand the way they respond AI is kind of the same way you've got to gather some information and really have a structured way to provide feedback loops so that it can improve over time and really to me that's the key I think that you know if you just throw AI at something and you don't do at least some you have to have some amount of human in loop to make sure it's
it's working properly. But I think you also have to have those feedback mechanisms either by a person or in place with responses to say this is a good response, this is a bad response, this is how this worked. And I think if you can build that sort of curation, then I think you can get to a point where you can say, okay, that's good enough and I trust that that will stay good enough so that it's responding to people in a way that's in keeping with our voice, our brand, and kind of the way we expect to treat our customers.
Scott Todd (35:57.638)
Yeah, it's a very good point, very good point. So that's the whole part, right? Like we want to think through before we jump into the tools and we start to jump into these, you know, building all of these automations. I want us to slow down and to think through what we're going to do, because when you do it this way, it's a repeatable system. And it all starts with a signal that, hey, something's wrong in our business.
We need to capture that friction point. And then when we go and we sit down, we spend a little bit of time. Look, we've spent like 40 minutes chitchatting about this process, which is not really that long of an investment to go now have some framework. But the number one mistake I see people making when it comes to AI and really automation is that they just want to jump in and start designing. That's fun.
They just want to start connecting all these tools together, but they haven't thought about these pieces. And then this whole system, this whole play becomes way too large. And then they're defeated.
Chad Coffman (37:07.894)
Absolutely.
Chad Coffman (37:12.554)
Exactly. If you just brought a person in off the street and said, here's my business, go get at it. And they have no training. You don't know what they do. You don't know what they're good at. It's kind of the same thing as just throwing AI at something. You'll get some response. You'll get something that's better than what you do right now. But it's a very, very point-like solution. And if you step back and are able to think about the system as a whole, then it kind of helps you make decisions about what's the right thing to do in a given situation and what's the right level of automation you need to go after.
Scott Todd (37:41.446)
Yeah. You're crushing it, man. You got it. You got it. So, congratulations to you guys. I cannot wait for you to put this into practice, this initial response. And just to share with me kind of the timing improvements or anything else because we all want to know.
Chad Coffman (37:44.584)
Yeah.
Chad Coffman (37:53.344)
Yep, I'm excited.
Chad Coffman (38:01.334)
Awesome, well, I'll do that. I appreciate the time, Scott. It's great to have this conversation.
Scott Todd (38:06.194)
All right, good to see you, Chad.
All right, Chad, how you doing today?
Chad Coffman (00:03.554)
Good, good morning, Scott. How are you doing?
Scott Todd (00:06.585)
Very good, thank you. All right, Chad, so the reason that I encourage you to be on here with me today is to do a build session, okay? Because it came to my attention that back in February, you and your wife who own a company ran into a lead problem that blew up your system, right? So tell me about
Chad Coffman (00:31.256)
Yes, correct.
Chad Coffman (00:36.332)
Yeah, so we had started a land flipping business. We buy and sell land. And we had been advertising through a variety of channels, had some challenges. We started using landmoto.com and had a tremendous amount of success very quickly. And it kind of caught us off guard. We had been kind of preparing for things, but when it all started coming in,
we started realizing all these challenges, bottlenecks and things that we were doing that we would probably prefer not to do to kind of move our, to keep our leads moving, flowing through our systems. yeah, exciting problem to have, but a problem nevertheless.
Scott Todd (01:16.303)
Yeah, and if I hear it correctly, it literally, to me, leads up, your system, right? Like you guys couldn't keep up, you were getting stuff done weeks and weeks later, which is, as you pointed out, is a great problem to have, right?
Chad Coffman (01:21.667)
Yes.
Chad Coffman (01:28.035)
Yep.
Chad Coffman (01:31.394)
Yeah, great problem, but still a problem.
Scott Todd (01:34.191)
So, you what I love about what you guys did is you identified the problem, okay? And you didn't stop and try to fix the problem right then and there. You triaged it, right? Like you went out and you solved the problem. You identified it. You continued to just work with the system that you had. And then as you got time back,
Chad Coffman (01:46.317)
Right.
Chad Coffman (01:57.356)
Yep.
Scott Todd (02:02.785)
it was identified as a key priority for you guys of a vital need, then you're now able to circle back and think about, okay, well, how do we fix this the right way?
Chad Coffman (02:12.96)
Exactly.
Scott Todd (02:15.043)
Yeah. And the one thing I see business owners do a lot of times is number one, they might just stop production right then and there and say, Hey, I got to solve this problem first. Right? Like that deep down inside that sometimes feels like the right thing to do. But the proper way to do it is like what you guys did, which is, okay, noted triage it. We're going to, we're going to limp along. Okay. We're going to limp along. And then when time allows, we're going to circle back and come
choose the problem or attack the problem, right?
Chad Coffman (02:48.384)
Exactly, exactly. And having a way to document the problems as you go is really important, obviously, so you don't lose track of the pain that you felt. When you get past these things, sometimes you can forget and you move on to the next thing. And then if you're constantly trying to solve as you go, sometimes you don't do the highest priority things. You just do what's right in front of you. And you lose that ability to step back and look at the system as a whole and say, what do need to fix next to really make our business better and sort of progress us holistically?
Scott Todd (03:16.696)
Yeah. And you know, I think what you just said is so key. You have to have a way of capturing. I call it capturing the frictions, right? Because if you identify each of these frictions and you have some methodology in which you capture them, which it can be, you know, it can be very analog, which is just on paper. Okay. Just take out a piece of paper and write down all of the frictions that you see in a given day.
Chad Coffman (03:37.816)
Mm-hmm.
Scott Todd (03:44.036)
The other way of doing it is you can get more technical and do things like in a cam-bomb board where you're looking at things and you're like, oh, let's just track all of these things. The way that you do it is really up to you. How do you do it? Tell me how you do it.
Chad Coffman (04:02.702)
So I happen to come from a pretty significant process improvement background. I have Lean and Six Sigma experience. I really love Lean tools. I'm sort of a student of process improvement, and I use technology to do that. So I know what it means to map the process first and then take off after the problem. So we had a very simple swim lane diagram for our processes. When we started running into problems, I created a value stream map, and I started capturing all the specific things that we do and looking at values.
add, non-value add time and the tools we're using and the know just the handoffs in the process. So I happen to have that background so and I have some tool sets at my disposal but the most important thing usually to me is that when something happens you're just capturing the pain so at least you capture the pain and then later you can go back and build out the nice value stream map and things like that and I can do that quickly so it's not a it's not a real problem but I think it's again the most important thing is to capture the pain and whatever your tool set you can go back and figure that out later.
Scott Todd (05:03.491)
Yeah, you one of the things that, because I find the way that you're working interesting. One of the things that I do is I use a tool called Obsidian. It's like a notion. It's, you know, it's a note system. I like Obsidian. But what I'll do in Obsidian is I will throughout my day, because I have a daily report there where I'm making notes or tasks that I've done completed or things that I see. I actually go in there and I create
a friction. I actually write, you know, like an open bracket friction, double colon, that way I can have data views and whatever. You know, I'm not going to go down that route. But I think that the most important part is that you capture it, okay? Because if you're not capturing it, you're going to forget it. And I love what you said. It's really about, hey, let's circle back, find the highest value obstacle that we can remove, and then we remove it, and then we go back and we
Chad Coffman (05:42.658)
Yep. Yeah.
Scott Todd (06:03.552)
we solve that problem once and for all, right? Or at least we think, right, yeah.
Chad Coffman (06:05.758)
Absolutely. Yeah, exactly.
Scott Todd (06:09.898)
So the friction that you identified was kind of the leads process was broken. That's kind of what I'm hearing, the lead process was broken. So that's what I would capture when I circle back to it. I'm looking at this. And then the way that I do it is I actually go through the scale framework and I'll kind of introduce the scale framework to you, but the scale framework is essentially five letters.
They're five acronym. It's one acronym, five letters. And it basically says scale. We're going to start with the scope because Chad, as you know, you could easily over engineer something. You can try to try to, you know, build out the entire system for something. But really the way I is it's a play. It's from here to here. Right. We're not trying to go from one end of the of the football field to the other. We're just trying to run one play.
Okay, so we want to make sure we're scoping it correctly. Then we're going to clarify the flow. And what I mean by that is we're going to map out what triggers it, what comes in, and what the intended goal of this particular play is. The A is we're going to automate the trigger. So in this case, a lead coming in, the trigger is the lead generation, somebody submitting a lead to you. The L
Chad Coffman (07:20.27)
Exactly.
Scott Todd (07:33.298)
is leverage the data. We want to know that the process worked or it didn't work, that our play worked or didn't work. And then E is elevate the experience. And what I mean by that is one, if we're going to have something communicating in a form letter kind of style to a customer, we want to make sure that it doesn't sound robotic. We want to make sure that it sounds human. sounds like chat, right? The other is that if there's an error,
Chad Coffman (07:59.852)
Yeah, exactly.
Scott Todd (08:03.212)
We want to make sure that it's not some system lingo like signal 19 error. What is that? Right? OK. That's the way I'm thinking about it. And what we'll do is we'll just start off here. And we'll start off with the scope, the S. So how do you see the scope?
Chad Coffman (08:10.51)
you
Chad Coffman (08:26.626)
That's interesting. I think it's initial lead. So we had had a totally different workflow and we moved to having leads coming in through the site land moto that you may be familiar with. And so we would see a lead come in and we didn't really have a good way to kind of start our whole, we didn't have a plan for how we reach out to customers and our potential customers leads really, and kind of go through the journey with them. So, as we started reaching out, we got different questions
They would ask, you know, then we realized we need some data that we need to gather for that everybody's asking for let's figure out how we do that and then we had repeated questions that were answering over and over again So the first two or three things we say to somebody is pretty routine There's really no value add and like the the interactions we're having they're just trying to get it kind of get the relationship up and running so and then you know as we got into the process there were you know things we needed to do to
to close a transaction with our customers. They might, you we have to potentially generate a deed or generate a note or, you know, talk just some of the conversations about financing and things like that. But a lot of it was very focused on, you know, we have counties that we work in and they have rules and regulations and so people are asking a lot of routine questions. And that was really a lot of it is helping people understand what they're looking at, if they're looking at buying land and, you know, what some of the limitations
and what opportunities they have when they purchase a particular piece of land.
Scott Todd (10:01.064)
Yeah. So if we go back and we look at something like this, really the scope that I'm hearing you say is it's really about the initial lead response, right? It's not trying to craft an entire sales message. Hey, we get the lead in.
Chad Coffman (10:11.514)
Exactly. Yes.
Scott Todd (10:18.674)
What is the first thing that we want to do? And I think that that's, you know, we're not trying to boil the ocean here. All we're trying to do is this one friction point because then we can daisy chain it into multiple sequence of plays. But for now, what we're going to look at today is that initial lead response, right?
Chad Coffman (10:25.816)
Right.
Chad Coffman (10:40.698)
Yeah, that's exactly right. The real challenge for us was getting somebody qualified to like that we're a good fit for them, they're a good fit for the land that we have to buy and getting them on the phone. And that way we can have, know, kind of do some initial, I guess you'd call it filtering. And then so when we get them on the phone, that it's a really good conversation. That was really the scope of what we were trying to do.
Scott Todd (11:06.484)
okay. So, so when we talk about clarifying the flow here, we know that the, that your initial trigger, if you will, is going to be an email, right? Because it's probably going to come in from an email. And then we want to think about the, that desire, try to see if I can make this here. How can I do this? Let's see how I make this.
look at that. Yeah, there we go. Now, and I can go out a little bit more. Okay, so we have this initial goal of getting the email. And then the end goal is not necessarily to make the sale. I know that's what you want. You want to make the sale. But really the end goal in this case is just a response back to the person who's inquiring, right? It's that initial.
Chad Coffman (12:04.257)
Exactly.
Scott Todd (12:06.527)
So.
you know, whatever it is. And so this email trigger may contain questions, it may contain certain things. So what we're just do here is we're just trying to get it to where we get that response back out to them. And as we start to think about how that happens, there's lots of nuances that could happen because, you know, if you wanted it to be very cold,
Well, then it would just be a robotic answer that says, hi, we have your lead. How can we help you? Or hi, we help you. But I think you want to go further than that, right? Like you want to do something different.
Chad Coffman (12:40.812)
Right.
Chad Coffman (12:46.914)
Yeah, mean, the first response is really important. The timing of it is really important, and that's one of the things that creates urgency and sort of made us drop what we were doing and start to deal with this. But I think the quality of that interaction is incredibly important. It sets the whole tone for how you're working with someone. It lets them know who you are and what their whole experience with us is going to be. So the tone and quality of that is really important.
Scott Todd (13:12.019)
So do you envision like having AI write those based on the initial response?
Chad Coffman (13:18.254)
What's really funny is
One of the things that we started doing was actually using AI to write our responses, but we were taking what they would ask, type it in, get a response, send it, you know, put it back to them. And so as we were actually kind of doing it, was like we're doing, you because we had so many coming in, a lot of times they might ask a specific question and we had a number of crafted responses. And at first it was copy paste some things and then it was, okay, let's use AI to tailor it. And so we were, you know, again, you're moving so fast sometimes and you got a lot of things going on. Somebody asked a very
specific question and we pull in maybe specific county information or other information that we'd already again we ended up loading it to to AI so we were starting to do go down that road but it was still very manual with us in the loop.
Scott Todd (14:04.275)
Yeah. So, you know, where we are with technology today, technology is changing every single day. was not available yesterday is available today. And that's why we never really want to jump to the actual solution because tomorrow you might get tools, shiny object syndrome and want to go chase, never have anything that works. So we try to stay like, hey, we want to know what this flow looks like.
Chad Coffman (14:14.008)
Yeah.
Scott Todd (14:34.493)
And then we can start to layer in how the tools might work, right? And so really when I do a build session like this, I'm just thinking about it like a blueprint, you know, like a blueprint for a builder who's going to go build a home, right? Because we know what the framing looks like. Now what we need to do is we need to actually get some boots on the ground and go, and then figure out like how we're going to put that in there, right? So as we think this through, the email comes in.
And ultimately what we want to do is we want to have, you know, like AI triage that email. Is that how you? Yeah. So we want to have AI, whatever AI model you want or whatever you think would work today. You want to have that triage it. And while it's doing that, it could potentially pull knowledge
Chad Coffman (15:13.93)
Exactly. Yep. Exactly.
Scott Todd (15:31.698)
I'll just put this over here, it could pull knowledge from your internal records on rules of that county, right? Rules of the county, property information, all of that stuff. And I think that the approach that you're doing there where if you stop and say, hey, listen, here's the counties that we're working in, here's the zoning rules,
Here's the rules that we understand, whether, I mean, some properties might be in a property owners association where they have their own deed restrictions. Taking all of that stuff and providing that into a knowledge bank that the AI can pull from, now it can look at that question and then it can turn around and create their response, right? So.
Chad Coffman (16:25.56)
Exactly.
Scott Todd (16:27.101)
Then the question is, would you have a human look at the response or are you just letting it go?
Chad Coffman (16:34.05)
Well, so this is what's interesting. I think that the important thing for us is that the responses are, whether automated or human, that they are very valuable for the folks that are getting them and that they don't feel robotic, that they are, again, truly like what someone would expect and maybe even exceeding their expectations if possible. That's kind of our philosophy across the board. So I think what we like to do is say, if we can use AI to do something and it can exceed expectations,
or at least kind of like, you know, meet them where they're at, you know, where they're at and kind of give them what they need, then that's a plus. Let's move forward with that. And if it can't, then let's figure out how we tune that and optimize as we move forward. And so a lot of times it's, you know, us sort of, you know, walking through things with AI and curating the responses. Then when we see that they get good enough and we've had enough information coming in that the responses can be, you know, very good and very quick, then I think we're ready to release that piece to
to allow AI to do the work. But we're a little conservative about things. Like we wanna make sure that, again, the experience for folks is really good, especially in the early stages. But that's kind of an approach throughout our journey. that was one of the things that was hard for me. My immediate thought was, me go gather a bunch of information. Let me create this knowledge base. Let me do all this stuff. And getting into the solutioning side where I had to just kind of accept that Cindy is using an AI.
Scott Todd (17:39.356)
.
Chad Coffman (18:03.92)
that's, you know, she's using maybe chat GPT and I'm using Claude and we're using different ones and yeah, the data is not all centralized and it's not organized but we're figuring this out as we go and so that was kind of our approach.
Scott Todd (18:16.702)
Yeah, OK. And so this is really where
you you start to think about and I think that everybody should think about because the initial gut feeling that people have is they just want it done, right? And so they want to triage it. They want that AI to draft the response and that produces the response right here. But at the end of the day, really, I think that we have to always remember that we need the human in the loop.
Chad Coffman (18:31.502)
All right.
Scott Todd (18:49.34)
in some way. it's one of the things that drives me crazy is when I hear people that are like, I have no employees. AI or you see in the news where companies are laying off all these employees and they're citing, we don't need them anymore because of AI. Well, maybe you need less people. get that part. Maybe you need less people. But at end of the day, they're
Chad Coffman (19:11.799)
you
Scott Todd (19:19.024)
there will be a competitive advantage to being able to deal with a human as opposed to a robot. I mean, that's already the case today. I mean, if you think about it, if you call up to any major company and you go through their voice response system, you know, like how many times are you like, just give me a human, I want a human, okay? And you you keep hearing, check out our website.
If I could do this on your website, I wouldn't be calling. Trust me. I can the on you. But if I'm calling you, I need you to answer the phone. I think it's the same. There's going to be a premium for that human experience. And it will be a competitive advantage that we don't just want to take and have the bots of the world just communicating everything. And so I like the fact that you're keeping the human there.
Chad Coffman (19:50.446)
Exactly.
Chad Coffman (19:57.965)
Right.
Scott Todd (20:16.859)
So now we keep that human in the loop to make sure that we're getting the response that we want, that it sounds like us, all of that good stuff. And then what? So like the human looks at it and says, yeah, I'm good to go. But now what are we doing? Are we just sending it? Or is there something else that we should be doing here?
Chad Coffman (20:38.734)
I mean, for us, was just, if it's an email response, we would send that to them. If it's a text response, if they wanted to be, use text messaging, which a lot of people do, we would just start texting them. And so we would craft the response, make sure it makes sense, send it out to them. And there was usually a little bit of back and forth. It wasn't usually just one message. It was usually maybe two or three fairly standard messages. And our goal is to get them on the phone. That's like the first thing. But there tends to be, like most people,
want to text or email back and forth a little bit and sort of get a feel for things before they get on the phone.
Scott Todd (21:14.456)
Yeah, I think that that's the other piece too is like, how do you want to communicate? How, you know, like that's, that's one of the early like dances. I think that you have to do with lead generation as you have to figure out, okay, like, Hey, you want to be where they want to be. And what you don't want to do is you don't want to say, well, we only communicate through email, right? Because I mean, you're there's people that communicate in many different ways. You need to show up where they are.
the format that they want to show up with, right? I think that's a good key point. So I'm going to put in here, we're really meeting them where they are in terms of communication method. the method is there, so it's not always an email. But we have that communication method, which leads us to our response, our initial response.
Chad Coffman (21:48.662)
Absolutely.
Chad Coffman (21:58.912)
Mm-hmm. Yep.
Scott Todd (22:10.552)
you know, in doing this where you have the AI triaging something and drafting it based on the knowledge that it has, really the time that you gain in doing this is really just shifted. And this is where I think that people lose sight of the human in the loop is that, you want the human in the loop to figure out where, not only where it's going wrong, yeah, if it's correct, just ship it.
But if it's going wrong, this is where we want to pause and strengthen back up the knowledge base.
Chad Coffman (22:44.684)
Yes, exactly.
Scott Todd (22:47.214)
How many times do people miss that?
Chad Coffman (22:49.526)
Yeah, a lot.
Scott Todd (22:51.928)
answer the AI, but they don't ever stop with the new found time and then go back and say, Hey, here, let's, let's, let's firm this up. Because Chad, mean, if you're working in the same County for years to come, it's really an investment of like that regulation's not necessarily going to change overnight, right? Like it's not, it's something that you and it's going to last you for years. Right. Yeah.
Chad Coffman (23:19.2)
Exactly.
Scott Todd (23:20.858)
put in the institutional knowledge. Okay. So we've clarified the flow. We kind of have an idea of what we want to do here. And then what we want to think about next is how we're going to automate the trigger. And what I mean by that is automate the trigger. And what I mean is that so many times, really there's four triggers in the world. There's four triggers that trigger everything. Number one is
You know, an event, okay. An event is a, is basically something that's, that's happened. could be a calendar event. could be, you know, somebody submits something that's an event. We can have time as a trigger. Okay. Time, time is a trigger. Monday at 8 AM is a trigger. An event is a trigger. A condition like, Hey, it's going to be 90 degrees outside today and it's going to rain. Well then.
Chad Coffman (24:14.371)
Mm-hmm.
Scott Todd (24:20.309)
send me a message that tells me this to bring my umbrella, that's a condition, you know, and the last one is manual. And so many times, it's easier to build those manual triggers, but the reality is that a trigger, a manual trigger is the worst trigger because it requires somebody to remember and we're going to forget, right? So, I think that the approach that you're having there in terms of
Chad Coffman (24:24.502)
Exactly.
Chad Coffman (24:41.816)
Yes. Yep.
Scott Todd (24:49.536)
the event being the submission is, I mean, it's logical in this case, right? Without it, you're relying on an employee to go into their email and check for emails, and you're counting on them to go do it, which they may not do. So clearly, what we're going to do here is the event. The trigger is the event. It's the submission or the receipt of that email is what we're going to do there. All right. Leverage the data.
Chad Coffman (24:53.955)
Yep.
Scott Todd (25:18.137)
This is where we're going to plan failure. And what I mean by that is, where can this process fail? It can fail because you may not get an email, which is your trigger. That's a potential. mean, look, email is email, right? So how do you know that you don't get an email, right? How would you know if you didn't get email? You just don't know. It's just not there. But this is where like a re-
Chad Coffman (25:34.915)
Yeah.
Chad Coffman (25:41.091)
Right.
Scott Todd (25:47.902)
regular check of whatever system sends that email is going to be your friend. Because if you have a regular, know, automated task that says, hey, let me go up there and schedule, let me just go look to make sure that we're getting everything we need. And then we can alert it. You know, how do you know that the AI failed to trigger off of this? And, you know, even humans, if you're going to outsource this or provide it to somebody else.
How do know that the human is not the bottleneck now? And that's the way that we want to think about this. Each step along the way, we want to think about there's data being produced on each individual step. And we want to know if something's not right. We want to know if something's not there. So what are your thinking behind that?
Chad Coffman (26:40.396)
Yeah, so this is really important to me. I think when you start to automate a process, like you said, you have to keep humans in the loop. I think the key that makes this work is that you have a structure for how the humans are involved. So I like to create cues and work tasks and things like that so that people don't have to remember something. They have something that becomes not just a routine, but it becomes something where they can be prompted in some way. So in this case, it would be, instead of just checking Landmoto constantly,
you check it maybe twice a day or once a day just to make sure nobody slipped through the cracks. And then it's scheduled, it's routine, it's very calm, it's very much just something you do. And you don't have to constantly like think, my gosh, did I forget? my gosh, did a new lead come in? So that's kind of what I see in that particular case is once a day you check, you know, Landmoto, make sure you didn't miss anybody, that an email, you know, didn't come out, you know, go astray or something like that. And then, you know, hopefully you can begin to trust that
process over time, again, regular checks are an important thing. And if they're scheduled, they don't take much. And if you schedule a number of tasks, then you've got a really set thing that you can do and your life is very relaxed and you have a very predictable flow to your day.
Scott Todd (27:55.443)
Yeah, you know, and see, this is where it's easy to over automate or over engineer something to catch at because like as you're saying that, the thing that's going through my mind is, yeah, we could have a human go and check this website to see if all the leads are accounted for. Or we could build, you know, an AI automation that logs in and checks to make sure that all the leads are accounted for.
Chad Coffman (28:20.172)
Yeah.
Scott Todd (28:24.416)
But see, that's the risk. Okay, like that is where so many people go down this rabbit hole of automation because the easiest thing to do today is the human in the loop and that's fine. Okay, because as you build this out, in my opinion, what we're doing is we're staying within this framework that we've identified. We're staying true to this scope.
And the fact that we were potentially saying, we want an automation that goes and validates this. That's a separate play. Right. Like that's a separate whole build session. That is a separate, that is a completely separate conversation. That's not for today, but yet it links to today. Right.
Chad Coffman (28:58.87)
Right, exactly.
Chad Coffman (29:13.656)
Right.
Well, I think what's interesting about that too is the concept of plays. you, and you kind of mentioned like, you know, AI is changing jobs, but I think, it's, it's changing the workforce, but it's, really changing. I think in some cases you might downscale, you know, your workforce and, you know, not have to have as many people, but in a lot of cases you want to shift what they're doing. And I think this is an important piece of this. So for example, we talked about the knowledge base and how that can, you know, help AI get better and everything like that. If you have a human that once a day goes in check,
and says, know, did everything go out? Well, that's probably very fast. And generally speaking, it's going to work. You know, it's just a quick check. But if you can also make that a checkpoint to say, let's review a few of the messages and make sure that they're getting proper responses. And let's identify if maybe there's some opportunities to shore up our knowledge base. Then you've got to like a micro play that you do instead of, you know, a macro play. And you can just check your knowledge base, start to tune that up. And you can use that as kind of a trigger for really enriching your knowledge base.
Scott Todd (30:06.87)
Thank you.
Chad Coffman (30:14.608)
as you're sort of checking to make sure that operations are moving smoothly. So I like to kind of combine things and think of these new plays and then again you can go back and automate later. But if you try to do all of it at once it can get very overwhelming and I think your systems can get very convoluted.
Scott Todd (30:21.174)
It.
Scott Todd (30:30.602)
Yeah. You know, it reminds me as you're talking about the micro place. When I started building my business, you know, 12 years ago, and the technology wasn't anywhere close to what it is today, we relied heavily on, you know, humans and VAs. And the problem that I kept seeing people run into when they were trying to build a VA team is that they would just take
all of this work and be like here. what we said is, hey, wait, wait, wait. If you're going to give an assignment to somebody, what's the smallest microstop? I think in microstops, right? Like here, let's go build this one little piece that I want them to go do. And then they go do this over and over and over again. And then it's like, okay, you've done part one.
Chad Coffman (31:01.772)
Right?
Chad Coffman (31:23.758)
Mm-hmm.
Scott Todd (31:26.495)
but yet there's another nine parts to go. So now let's work on two and three now, right? So now you've got them, like they're launching stronger than they go learn two and three, and then they go learn all the rest of the steps. And I slowly would hand the work off to them. And it's the same way with automation because we can't try to automate the entire system. I mean, you could, it's not gonna work. You're not gonna be happy with it.
Chad Coffman (31:54.638)
No, and think one of the things that's really important is that even if you are building a system, you have to design that system. so if you're teaching the same way you would teach people to learn a new process and a new system, and especially with AI, you have that opportunity to use that same thinking, but you still to be methodical and you have to plan that out. And you may not know what the instructions are right in the beginning, and you may not know what's going to work with a particular person or particular AI. So I think you have to be very methodical and have a way that
Again, you roll that out. I love the way you think about that.
Scott Todd (32:27.615)
Yeah. So for leveraging the data, we're going to say, hey, we're going to build in some of those checks, but we're also going to have humans overseeing some of the processes in the beginning before we can later then go back and further automate the humans overseeing it to go do the right. So we're going to have a human oversee it today because that's the micro step. That's the easiest way to go. And then later on, as we look at new frictions that arrive in the business,
The new frictions might be these humans that are overseeing some of this automated work that we can go build the automation, right? And then back down to elevating the experience. And this is the voice. And one of the things I think that you mentioned that you were doing that I liked that you said that you were doing is that you're having AI kind of draft or triage that message, draft the response for the human to go back and look to make sure, is this the messaging that we want? Does this sound like us? So you know, you're...
Chad Coffman (33:03.766)
Exactly.
Scott Todd (33:25.714)
you're kind of doing that, but we all know that, well, I shouldn't say we all know, it is pretty common knowledge that, you can have AI kind of build a voice guide for you, learn from your own writing. That's really two years ago, that would have been wild two years ago, right? know, today it's like, I want it to sound more like me, not like a robot. And we have that part, but what are your thoughts on that?
Chad Coffman (33:54.189)
Ask me that question again. I hear a couple things I wanted to respond to but
Scott Todd (33:59.454)
When you're thinking about elevating the experience and you're going to have AI out there and draft that response, how are you number one, finding or helping AI to find your voice? Okay. Cause you want to part. And then the second part is, is really how, how are you going to elevate the experience or how do you think you're going to elevate the experience for the, for the customer and for you to know that something's wrong?
Chad Coffman (34:10.392)
Yes.
Chad Coffman (34:26.094)
Yeah, I think the key is that...
You know, for an AI to find your voice, you really have to have like a body of responses and things that you've, know, ways that you talk, you respond, information you provide in a given context, things like that. And I think that's kind of a skillset all on its own. It's almost, in many ways, it is a lot like teaching a person. You you want to show them, maybe you have recordings of calls you've made, or you have, you know, recording of something you do so they can see how you
it and you have templates maybe of things that you've responded to in the past you give it to a person that person reads it and says oh okay I get what they're saying I understand the way they respond AI is kind of the same way you've got to gather some information and really have a structured way to provide feedback loops so that it can improve over time and really to me that's the key I think that you know if you just throw AI at something and you don't do at least some you have to have some amount of human in loop to make sure it's
it's working properly. But I think you also have to have those feedback mechanisms either by a person or in place with responses to say this is a good response, this is a bad response, this is how this worked. And I think if you can build that sort of curation, then I think you can get to a point where you can say, okay, that's good enough and I trust that that will stay good enough so that it's responding to people in a way that's in keeping with our voice, our brand, and kind of the way we expect to treat our customers.
Scott Todd (35:57.638)
Yeah, it's a very good point, very good point. So that's the whole part, right? Like we want to think through before we jump into the tools and we start to jump into these, you know, building all of these automations. I want us to slow down and to think through what we're going to do, because when you do it this way, it's a repeatable system. And it all starts with a signal that, hey, something's wrong in our business.
We need to capture that friction point. And then when we go and we sit down, we spend a little bit of time. Look, we've spent like 40 minutes chitchatting about this process, which is not really that long of an investment to go now have some framework. But the number one mistake I see people making when it comes to AI and really automation is that they just want to jump in and start designing. That's fun.
They just want to start connecting all these tools together, but they haven't thought about these pieces. And then this whole system, this whole play becomes way too large. And then they're defeated.
Chad Coffman (37:07.894)
Absolutely.
Chad Coffman (37:12.554)
Exactly. If you just brought a person in off the street and said, here's my business, go get at it. And they have no training. You don't know what they do. You don't know what they're good at. It's kind of the same thing as just throwing AI at something. You'll get some response. You'll get something that's better than what you do right now. But it's a very, very point-like solution. And if you step back and are able to think about the system as a whole, then it kind of helps you make decisions about what's the right thing to do in a given situation and what's the right level of automation you need to go after.
Scott Todd (37:41.446)
Yeah. You're crushing it, man. You got it. You got it. So, congratulations to you guys. I cannot wait for you to put this into practice, this initial response. And just to share with me kind of the timing improvements or anything else because we all want to know.
Chad Coffman (37:44.584)
Yeah.
Chad Coffman (37:53.344)
Yep, I'm excited.
Chad Coffman (38:01.334)
Awesome, well, I'll do that. I appreciate the time, Scott. It's great to have this conversation.
Scott Todd (38:06.194)
All right, good to see you, Chad.