Latest / Agents on Main Street / The AI Constitution
Transcript
- David: Before you buy anything or record anything, Mitchell: Before you buy anything or record anything, you need to decide what your business will and won't do with AI. David: you need to decide what your business will and won't do with AI. One page written by you, not by a vendor. Mitchell: One page written by you, not by a vendor. Everybody wants to know which AI tool to buy. That's the wrong first question. David: Everybody wants to know which AI tool to buy. That's the wrong first question. Mitchell: f the first question is what is your company willing to do with AI and what is it not willing to do? David: If you write that down, Mitchell: you write that down, David: you've Mitchell: you've David: got something people in tech will call a constitution. Mitchell: got something people in tech will call a constitution. David: Skip it, every decision after this one's just a coin flip. Mitchell: Skip it. Every decision after this one's just a coin flip. And it starts with one page, you can write it this week. David, I not to to dig into the politics of it, but yesterday top AI leaders come together and agreed on what they're calling a constitution. Now we weren't gonna dive into what they're talking about, but it does bring up the topic for small businesses and an AI constitution, which is something you've really, really been focused on for the last seven or eight months now. Can you tell us what that is? David: really been focused for the last seven or eight months now. Can you tell us what that is? man, I heard about an AI constitution from another podcast probably you know eighteen months ago and Mitchell: And David: probably about a year ago I started considering like well what would our constitution look like at at our our company and Mitchell: year ago I started considering well what would our constitution look like at at our our company and yeah I had no idea. and so I started out I I listened and learned about the pillars of that concept of a constitution. Like what are the parts David: Yeah, I had no idea. and so I started out, I I listened and learned about the pillars of that con s of a constitution. Like what are the parts of a constitution that I actually need? And then I could maybe start thinking about the different, you know, components of that, the subcomponents. And you know, I I Mitchell: of a constitution that I actually need and then I could maybe start thinking about the different, you know, components of that, the subcomponents. And you know, I I introduced those pillars into an L L I answered questions. I had a conversation with my with my AI and told it about my company, what my values were, the culture of my company, my customers, how I interacted with them, how I protected their information, what was sensitive, what was not sensitive. we you know, I David: introduce those pillars into an LLM. I answered questions. I had a conversation with my with my AI and told it about my company, what my values were, the culture of my company, my customers, how I interacted with them, how I protected their information, what was sensitive, what was not sensitive. we you know I Mitchell: I put a lot of thought into David: I put a lot of thought into Mitchell: that and you know it spit out a document and then read that document and I said, you know, hey that looks great, you know, I'm and I'm I tweaked a few things and and I came out with our first iteration which was like February of twenty six and we worked with that for a while. David: that and, you know, it spit out a document and then read that document and I said, you know, hey, that looks great. You know, I'm and I'm I tweaked a few things and and I came up with our first iteration, which was like February of twenty six. And we worked with that for a while. you know, it wasn't a policy per se. It's not it's Mitchell: it wasn't a policy per se. It's not it's it's it's not like a binding thing. It was more of a you know, David: It's it's not like a binding thing. It was more of a you know, just it it was something that was a guideline for our team to to touch to every time they had an AI question. They could say, you know, is it okay for me to use AI here? should I reveal that information to someone, you know, that I used AI here or it just kinda gave us a framework for how to behave. Mitchell: just i it was something that was a guideline for our team to to touch to every time they had an AI question. They could say, you know, is it okay for me to use AI here? should I reveal that information to someone, you know, that I used AI here or it just kinda gave us a framework for how to behave. David: Well, and you you mentioned Mitchell: Well, and you you mentioned, your company values and and so one of the purposes behind the AI constitution, it's not like you said, it's not a policy or procedure document, but it is about defining your corporate goals as they relate to AI. for example, you know, in your case building a better home. can you kinda expand a little bit on on how how do you codify David: values and and so one of the purposes behind the AI constitution, it's not like you said, it's not a policy or procedure document, but it is about defining your corporate goals as they relate to AI. for example, you know, in your case building a better home. can you kinda expand a little bit on on how how do you codify what your company is Mitchell: what your company goals are with AI. You know, I I like I said I started with those pillars and here's the first one we made. is actually several pages, but it's in the outline form, you know, it's not a it's not just like paragraphs. But those p those pillars were our transparency, David: goals are with I I like I said I started with those pillars and here's the first one we made is actually several pages but it's an outline form you know it's not it's not just like paragraphs but those those pillars were our transparency, our ver how we would verify information, Mitchell: our ver how we would verify information, pri how we dealt with privacy, how we dealt with accountability, enforcement and governance, and and just, you know, kind of the doctrine at the end. And just going through that was really instructive itself, you know, just thinking through those problems and how it related to my company. David: pri how we dealt with privacy, how we dealt with accountability, enforcement and governance and and just, the doctrine at the end. And just going through that was really instructive itself, you know, just thinking through those problems and how it related to my company. you know it's also Mitchell: it's also a document that's gonna change. I mean we all we know a constitution is you know quote unquote living document and I think back to like February of last year it just seems so far back and because of the rapid change of David: a document that's gonna change. we all we know a constitution is, a quote unquote living document. And I think back to like February of last year, it just seems so far back. And because of the rapid change of of AI and and AI and business and the perception of AI and how people at that point in time Mitchell: of AI and ha and AI and business and the perception of AI and how people at that point in time were very concerned about was this AI generated information? You know, like, David: were very concerned about was this AI generated information? You see it on socials, right? You see it where, it comes up and says, This is an AI generated infr image, And at first people were having trouble discerning between, was this AI or was this a real image? And so they felt like there was, you know, a need to do that. And Mitchell: You see it on socials, right? You see it where, you know, it comes up and say, This is an AI generated inf image, you know. And at first people were having trouble discerning between, you know, was this AI or was this a real image? And so they felt like there was, you know, a need to do that. And you know, AI th also has been purported to have an accent of sort. And it produces prose in a David: and, you know, AI th also has been purported to have an accent of sort. And it produces pros in a Mitchell: in a certain way that, you know, for those of us who use it regularly we can you can hear it. I know what you did there. David: in a certain way that, for those of us who use it regularly, we can kinda like y you can just see it and you're like, I know what you did there. and I think as that has gone on, people have become more and more comfortable with it. Like the they're just like, yeah, everybody's using AI for all kinds of things, you know, like I think there was I I heard somewhere like there was a New York Times author Mitchell: And I think as that has gone on, people have become more and more comfortable with it. Like they they're just like, yeah, everybody's using AI for all kinds of things, you know, like I think there was I I heard somewhere like there was a New York Times author David: And that somebody was, you know, accusing him of using AI and he was like, Ps, of course I use AI. Everybody uses AI. Like it's not saying AI wrote his piece, but he's using it as a thought Mitchell: and that somebody was, you know, accusing him of using AI and he like, Ps of course I use AI. Everybody uses AI like it's not saying AI wrote his piece, but he's using it as a thought. partner. David: partner. And s Mitchell: And I think the question has been refined, right? It it started was this AI generated? Now I think the the better question is did you read it? Yeah, and I you know do have some ownership David: it started was this AI generated now I think the the better question is did you read it? Yeah and I think do you have some owner Mitchell: in this work product? And that goes into the Constitution, David: And that goes into the constitution. Mitchell: right? Because you you can't have your employees just blindly spitting things out, especially from these modern technologies that you know We don't know David: We don't know. Mitchell: how they behave yet. can you speak a little bit on the the risk of blindly trusting AI and how your constitution can address that? I mean that was one of the primary David: speak a little bit on the the risk of blindly trusting A constitution can address that that was one of the primary concerns in my constitution was dealing with AI slop. I don't want a company with just a bunch of AI generated that's what they call it in the vernacular is s you know slop. And it's just where people produce massive amounts of content Mitchell: concerns in my constitution was dealing with AI slop. Yeah. I don't want a company with just a bunch of AI generated, that's what they call it in the vernacular is you know, slop. Yeah. And it's just where people produce massive amounts of content David: And they never read it. It could be it could have errors in it. It could have what the you know, what we call hallucinations, it could have things in there that are Mitchell: and they never read it. It could be it could have errors in it. It could have what the you know what we call hallucinations, it could have things in there that are David: just completely made up. who knows? And so it requires a hu you know, human to proofread just like you would read anything else an employee made for you. You used to labor over that email before you hit send. Yeah, yeah. Yeah, you should read it. Mitchell: completely made up. who knows? And so it requires a h you know human to proofread just like you would read anything else an employee made for you. Right. You used to labor over that email before you hit send. You should still do that. Yeah you should read it David: And you should make sure that it's saying what you want it to say. And I think that was one of the early things was, verification. Mitchell: and you should make sure that it's saying what you want it to say. And I think that was one of the early things was you know verification. no AI output is final until a qualified human signs off. That was, you know, the mantra in that section. you know, if a reasonable person would ask was this written or generated by AI, disclosure is required. That David: no AI output is final until a qualified human signs off. That was, you know, the mantra in that section. you know, if a reasonable person would ask, was this written or generated by AI, disclosure is required. That Mitchell: these are some of the things that David: These are some of the things that are in the constitution that that give employees something to go back to and and real you know, how to do their job using AI. Because I mean l just taking a step back. Mitchell: are in the constitution that that give employees something to go back to and and real you know how to do their job using AI. Because I mean just taking a step back. David: Everybody is using AI in some way. I as a business owner, I'm going to assume at this point that all of my employees are using AI. How they use AI Mitchell: Everybody is using AI in some way. I mean, I as a business owner, I'm going to assume at this point that all of my employees are using AI. How they use AI David: matters more to me than whether or not they're using it. It's because I want Mitchell: matters more to me than whether or not they're using it. It's because I want them to use it in a responsible fashion. I want them to check what they're producing and make sure that it's what they wanted to produce. And I want it to fit into the culture of our organization so that they're not doing things that none of us would do. David: them to use it in a responsible fashion. I want them to check what they're producing and make sure that it's what they wanted to produce. And I want it to fit into the culture of our organization so that they're not doing things that none of us would do. Mitchell: And kind of along those lines, how do you how do you go about thinking about David: what Mitchell: what the AI can be involved in, what it can't be involved in. are there some sensitive areas that that you've learned to to keep the AI out of? yeah, like you know, I'm you know, we have a we have an agent that works inside and you know all this, but we have an agent that works right in the middle of all of our data. And David: the AI can be involved in, what it can't be involved in. are there some sensitive areas that that you've learned to to keep the AI out of? yeah, like you know I'm you know we have a we have an agent that works inside and you know all this but we have an agent that works right in the middle of all of our data. And We've had to talk about that a lot. do you want your employees to be able to ask your agent who got a raise last week? 'Cause somebody might ask. And if the data is available, the agent either knows or doesn't know to answer it. And you have to have Mitchell: We've had to talk about that a lot. you know, do you want your employees to be able to ask your agent who got a raise last week? 'Cause somebody might ask, and if the data is available, the agent either knows or doesn't know to answer it, and you have to have David: those privacy safeguards in place and that's the kind of thing that's in an AI constitution so that when you give your agent instructions, the agent also follows the constitution. Just like you would have done with an employee. Just like that. Mitchell: those privacy safeguards in place and that's the kind of thing that's in an AI constitution so that when you give your agent instructions, the agent also follows the Constitution. Just like you would have done with an employee, right? There there are topics that, hey, you can't share this with everybody in the company. Don't discuss your pay with your coworkers. And I mean AI is it is an employee. I mean look at David: Discuss your pay cohort. Mm-hmm. And I mean, AI is it is an employee. I mean, I look at it like as a digital employee. And I would say the same thing to anybody managing my HR, like, you know, keep those files or those files are private. Like they have private information in them. They have, you know, Mitchell: a digital employee and I would say the same thing to anybody managing my HR like you know keep those files or those files are private like they have private information in them they have you know David: social security numbers or birth dates or, you know Mitchell: social security numbers or birth dates or whatever, a all kinds of personal identifiable information. And that needs to be protected. Similarly, the finances of the company. You know, y you don't want everybody knowing everything about the finances of the company. I mean unless you you know even publicly traded companies really curate certain information, then yeah it's curated. And David: Whatever, a all kinds of personal identifiable information and that needs to be protected. Similarly, the finances of the company. y you don't want everybody knowing everything about the finances of the company. unless you even publicly traded companies release certain information that yeah, it's curated and Mitchell: and filtered s for the public eye David: filtered s you know for the public eye. Mitchell: because you know, David: Because you know, th different competitive information is in there and it could be detrimental to the company if it's shared outside of a certain circle. So that circle has to be defined. And, you know, there are different ways of doing that. And we've dealt with it, like in some ways we had to we just said that the agent will not have access to any of this data, but then we made another agent that had access to Mitchell: Different competitive information is in there and it could be detrimental to the company if it's shared outside of a certain circle. So that circle has to be defined. And you know, there are different ways of doing that, and we've dealt with it like in some ways we had to we just said that the agent will not have access to any of this data. Yep. But then we made another agent that had access to David: all of that data but only certain people had access to that Mitchell: all of that data, but only certain people had access to that. agent. And so that's how we we are using our constitution to guide us in making those decisions as we, move through the development process. And and so David: agent. So that's how we we are using our constitution to guide us in making those decisions as we, move through the development process. And and so Mitchell: as as we've worked through that, you know we've we've seen kind of both sides of that discussion where you don't want the you don't want the AI sharing inside the company. David: We've seen kinda both sides of that discussion where you don't want the you don't want the AI sharing inside the company, but what happens if you let your AI be involved in external communications? Does it say wild things to your customers? Does it divulge you know I what if it tells your customer like, Hey, cash flow is kinda tough? Can you hurry up and sign this contract? You know, I mean Mitchell: But what happens if you let your AI be involved in external communications? does it say wild things to your customers? Does it divulge you know what if it tells your customer like, hey, cash flow is kinda tight? Can you hurry up and slide this contract? You know, I mean David: I no, I would not want that, right now. I mean, I can see a future where that would be, you know, some th those types of tools could be useful for customers to have access to certain information, that the comp you know, about what their role or their interactions with a company. So Mitchell: can see a future where that would be, some th those types of tools could be useful for customers to have access to certain information that the comp you know about what their role or their interactions with a company. So David: that specific relationship customer to business, you know, the B to C relationship, i in that o in that Mitchell: that specific relationship, customer to business, the B to C relationship, i in that own in that David: one channel I could see a future where there's utility in having Mitchell: one channel I could see a future where there's utility in having David: a bot speak with them about their job or about, you know, their the product they're being delivered. Mitchell: a bot speak with them about their job or about you know their the product they're being delivered. Reminds me of what David's bridle has done where each customer gets a dedicated agent and that helps one it helps with the data leakage right because your agent just never speaks to anybody but you so it can't tell your secrets. David: Each customer gets a dedicated agent and that helps. One, it helps with the data language. Right? Because your agent just never speaks to anybody but you. So it can't tell your secret. Mitchell: but you know I think that kind of brings to mind David: But I think that brings to mind the next problem that you need to address in your constitution and that is accountability and liability for what an agent does. And and does does responsibility mean liability? Mitchell: the next problem that you need to address in your constitution and that is accountability and liability for what an agent does and and does does responsibility mean liability? David: I don't I don't know. And I don't think anybody knows yet. These are this is a completely I mean we call them frontier models, right? We are in the frontier of this industry. We are yeah, I mean I'm not even sure we have a trail. I mean we're just we're just bushwhacking here and by definition Mitchell: Don't I don't know. And I don't think anybody knows yet. These are this is a completely f I mean we call frontier models, right? Right. We are in the frontier of this industry. This is the Oregon Trail. We are yeah, I mean I'm not even sure we have a trail. I mean they're just we're just bushwhacking here and by definition David: the legal Mitchell: the legal profession follows technological innovation because how would you if you think through this with me like you how would you be able to have a legal opinion on something that has not yet been created? Sure. You know, or or conceived. The law can never get in front of it. Yeah, the law should not be in front of this. But typically, at least up until, you know, the last David: profession follows technological innovation because how would you if you think through this with me like you how would you be able to have a legal opinion on something that has not yet been created you know or or conceived. So never get in front Yeah the law should not be in front of this but typically at least up until you know the last three to five years Mitchell: three to five years, technological innovation has happened at a pace that the legal profession could keep up. Yep. relatively keep up. You know, maybe they're a year or two behind, but they were on top of it. Now change is happening so rapidly that the the the challenge of being able to respond legally to I mean the the David: Technological innovation has happened at a pace that the legal profession could keep up. relatively keep up. You know, maybe they're a year or two behind, but they were on top of it. Now change is happening so rapidly that the the the challenge of being able to respond legally to I mean, the the problems are myriad. There are Mitchell: problems are myriad. There are all kinds of legal questions out there. So who is responsible, who is accountable? I think that's just something that you have to do your best good faith effort at defining in your company at this point. You know, and you're gonna have your morals and your ethics and your culture in your company and that all needs to be visible in your constitution because David: all kinds of legal questions out there. So who is responsible, who is accountable? I think that's just something that you have to do your best good faith effort at defining in your company at this point. You know, and you're gonna have your morals and your ethics and your culture in your company and that all needs to be visible in your constitution because it's much better Mitchell: it's much better in in my opinion that if I had a challenge in my company that somebody presented and said as a l a legal question, at least I have a document to fall back on and say, well this is our policy. Right. If I had no policy or a policy of no policy, you know, I'm at the whims of nature. I mean I don't have any idea David: in my opinion that if I had a challenge in my company that somebody presented and said as a l a legal question, at least I have a document to fall back on and say, Well, this is our policy. If I had no policy or a policy of no policy, you know, I'm at the whims of nature. I mean, I don't have any idea Mitchell: what people are doing, why they did it, how they did it, who they did it for. I I'm just letting things happen, and that's irresponsible in my view. I think that's irresponsible. So as a business owner, it is the onus is on you to lead and make sure that your people know what they should and shouldn't do in in the operations of the business. Well, sure. I mean an employee. David: what people are doing, why they did it, how they did it, who they did it for. I I'm just letting things happen and that's irresponsible in my view. I think that's irresponsible. So y as a business owner it is the onus is on you to lead and make sure that your people know what they should and shouldn't do in in the operations of the business. Mitchell: An employee can't be guilty of breaking a rule if the rule was never written. something that that I really think helped as we were putting together the the AI constitution and as we've adapted it over time has been it's a broader than just the Constitution. It is an internal approach to mistakes and problems that can kind of shortly be described as blameless. David: An employee can't be guilty of breaking a rule if the rule is never written. That's right. something that that I really think helped as we were putting together the the AI constitution and as we've adapted it over time has been it's a broader than just the Constitution. It is an internal approach to mistakes and problems that can kind of shortly be described as blameless. Mitchell: And that's where we we discuss the problems, right? That beam is undersized. we forgot to send the the equipment back to the rental company. but especially as as AI is accelerating what your business is doing and it's giving greater autonomy to your employees, can you speak a little bit on how you've David: That's where we we discuss the problems, right? That beam is undersized. we forgot to send the the equipment back to the rental company. but especially as as AI is accelerating what your business is doing and it's giving greater autonomy to your employees, can you speak a little bit on how you've how do you allow your Mitchell: How do you allow your employees and by proxy their agents to make mistakes and and learn things? David: employees and by proxy their agents to make mistakes and and learn things. Yeah, I mean that is a huge component of this. Like you know it's a sandbox environment. You have to have this environment where Mitchell: that is a huge component of this. Like, you know, it's a sandbox environment. You have to have this environment where, you know, there are toys to play with in a sandbox and it there the sandbox protects you in it's a you it is a safe place to dig a hole. It's that safe place to to dig a hole, you know. Or or fill one up or whatever. And David: there are toys to play with in a sandbox and it there the sandbox protects you and it's a you know, it is that it is that safe place to to dig a hole, yeah. Or or fill one up or whatever. And Mitchell: And so if you have that, you you want to be able to ha let people innovate. You this is a new technology. People are exploring it. They're trying to learn it. They you know n none of none of us are compl David: And so if you have that, you you want to be able to ha let people innovate. You this is a new technology. People are exploring it. They're trying to learn it. They n none of none of us are completely Mitchell: completely knowledgeable about what AI can and can't do. We're all learning and some of us have you know gotten a jump on this and others are are still coming along and so you have to David: knowledgeable about what AI can and can't do. We're all learning and some of us have, you know, gotten a jump on this and others are are still coming along. And so you have to Mitchell: provide people that that safe place and the flexibility. I mean again that's the constitutional role is t to provide a safe place to innovate David: provide people that that safe place and the flexibility. again, that's the constitutional role is t to provide a safe place to innovate under certain constraints. Right. Mitchell: under certain constraints. Right. Just like the American Constitution allowed us to run that great experiment. It gave us that sandbox. David: Mm-hmm. It gave us that sandbox. Yeah. Mitchell: kind of tangential, but it it's I think it's a big deal for a lot of small business owners. your team is not a group of technical wizards. They are not geeks, so to speak. David: Team is not a group of technical wizards. They are not geeks, so to speak. And so what have you seen in terms of resistance and adoption with with your employees approaching AI or adapting to AI encroaching in their space? Mitchell: And so what have you seen in terms of resistance and adoption with with your employees approaching AI or adapting to AI encroaching in their space? Well, different David: I've you know, there's certain Mitchell: there's certainly a little bit of, you know, fear and anxiety because people see how powerful it is and and how effective it can be in producing high quality work and especially digital work. David: a little bit of, fear and anxiety because people see how powerful it is and and how effective it can be in producing high quality work and especially digital work. and so I think that gives humans some concerns. I've had everything from Mitchell: And so I think that gives humans some concerns. I've had everything from wow, this is amazing, I love this, this helps me do my job better to you know, outright disgust. and David: Wow, this is amazing. I love this. This helps me do my job better. To you know, outright disgust and hatred of AI and and then seeing those people transform into adopters of AI. So like I've seen people who had tremendous fear of it and and a lack of knowledge. Mitchell: hatred of AI and and then seeing those people transform into adopters of AI. So like I've seen people who had tremendous fear of it and and a lack of knowledge And now they're some of my most diligent users of the technology. So I you know, you have to let people have space to grow. You're gonna have setbacks, you're gonna have blow ups, you're gonna have people that that express themselves in different ways. I mean, it's a very it has a lot of tentacles, right? You know, it gets into every bit of our work and David: And now they're some of my most diligent users of the technology. So I you know, you have to let people have space to grow. You're gonna have setbacks, you're gonna have blow ups, you're gonna have people that that express themselves in different ways. I mean, it's a very it has a lot of tentacles, right? You know, and gets into every bit of our work. And some people Mitchell: Some people they just react differently to it and I think you just have to provide that space to do it. That you know, again I d I wanna beat that horse that that constitution has allowed us to do that because you know one of the things in that we talk about and this wasn't in our original constitution, you know, now we have another one. six months later we updated our constitution. David: They just react differently to it. And I think you just have to provide that space to do it. That again, I d I wanna beat that horse that that constitution has allowed us to do that because, you know, one of the things in that we talk about and and this wasn't in our original constitution, you know, now we have another one. six months later we updated our constitution. But in the the most modern one we have, it talks about AI replacing people and that is not our goal. That is not what we want to do. What we want to do is to improve the productivity of the people we have. And that's been I think that was one of our our best ways to communicate and help reduce the the felt risk is Mitchell: But in the the most modern one we have, it talks about AI replacing people, and that is not our goal. That is not what we want to do. What we want to do is to improve the productivity of the people we have. And that's been, I think that was one of our our best ways to communicate and help reduce the the felt risk is these tools are assistance, right? I I I'm thinking of our project management. David: These tools are assistance, right? I I I'm thinking of our project manager Mitchell: Who once the light bulb came on that it was a CYA move for him to in our case, right, just make a phone call so that it could enter the agent's awareness. that has been really David: who once the light bulb came on that it was a CYA move for him to in our case, right, just make a phone call so that it could enter the agent's awareness. that has been really Mitchell: transformational in the approach. It's it's not an overlord, it's not here to spy on you. It's certainly David: transformational in the approach. It's it's not an overlord. It's not here to spy on you. It's certainly in the case of home building, it's not here to take your job. It's here to help you do your job. Yeah, and be more effective and be more efficient. How it's done that for you and how having these things codified have reduced stress for you. Well, just being able to have meetings Mitchell: in the case of home building it's not here to take your job, it's here to help you do your job. Yeah. Be more effective. Can you talk a little how it's done that for you and how having these things codified have reduced stress for you? Well, I be just being able to have meetings with the team and a whole David: with the team and a whole I mean every all these meetings require you to take copious notes. Right. I mean somebody's not paying attention while they're Yeah, you you've got to have you've got to have a lot of notes. And so we used to have like in the notes or the creation of those notes was done by the office manager. and you know, she's sitting there in the meeting running Mitchell: Every all these meetings require you to take copious notes. Right. somebody's not paying attention while they're writing 'cause we don't multitask. You gotta have you gotta have a lot of notes. And so we used to have like in the notes, or the creation of those notes was done by the office manager. and, you know, she's sitting there in the meeting running a laptop with a screen and David: a laptop with a screen and you know, w running, you know, an agenda and and I'm going through that agenda, you know, one by one with the team. And as we're going through the different jobs and we're getting the debrief on all those jobs and updates, she's sitting there trying to keep up, you know, and Mitchell: we running, an agenda and and I'm going through that agenda, you know, one by one with the team and as we're going through the different jobs and we're getting the debrief on all those jobs and updates, she's sitting there trying to keep up, you know, and David: and there's this to do and that to do, and then by gotta have this by the 15th, and and then the inspections on the eighteenth and blah blah blah blah blah. And it's just like all these things. And she's a human. Yeah. She Mitchell: and there's this to do and that to do and then by gotta have this by the fifteenth and and then the inspections on the eighteenth and blah blah blah blah blah it's just like all these things and she's a human. Yeah. She David: actually, you know, like all humans, makes mistakes. And Mitchell: actually, all humans makes mistakes. And David: it's not even a mistake, it's just the volume of information is just so much. And so when we implemented a recording of our Mitchell: It's not even a mistake, it's just the volume of information is just so much. And so when we implemented a recording of our meetings, you know, a note taker, now I I know people s some there's a whole debate. Yeah. Like should we still take notes or should we let the note taker the digital note taker do it? You know, some people actually and I'm one of if I write something down, you know, it it resonates. David: meetings, a note taker. Now I I know people s some there's a whole debate, should we still take notes, or should we let the note taker the digital note taker do it? You know, some people actually and I'm one of if I write something down, you know, it it resonates Mitchell: Yeah, I never even need to look at the paper again. Yeah, so it just helps me synthesize information to write. But David: with me better. Yeah. Yeah. So like it just helps me synthesize information to write, but Mitchell: you don't have to write everything now. I don't have to write everything down. In fact, I don't have to be anxious about it. I know the digital note taker is getting it. David: I don't have to write everything down. In fact I don't have to be anxious about it. I know the digital note taker is getting it. Mitchell: You write down the thing like, there's there's a key for me, there's a takeaway. And then when I introduce that in that data to our agent. David: And then when I introduce that in that data to our agent, the agent generates the to do lists and distributes it out to the right roles. And so and according to our constitution, you know, it it doesn't share sensitive information that was shared in the meeting. It shares the to do list that came from the meeting. And Mitchell: The agent generates the to do lists and distributes it out to the right roles. Mm-hmm. And so and according to our constitution, you know, it it doesn't share sensitive information that was shared in the meeting. It shares the to do list that came from the meeting and lets the right people know what they need to do. That has been, I mean, a tremendous David: lets the right people know what they need to do. That has been a tremendous Mitchell: gain l so many fewer dropped balls, so many more efficiencies. So what areas David: gain. l so many fewer dropped balls, so many more efficiencies. So what areas have you Mitchell: have you found that you're not ready or AI is not ready? What still needs basically a hundred percent human David: AI is not ready. What still needs basically a hundred percent human? Well, I mean, still has to have a human to swing a hammer. you know, still have to have a human go and see a job site. you know, I still there's still a need for humans to talk to other humans to generate new business. Mitchell: To have a human swing a hammer. you know, still have to have a human go and see a job site. you know, I still there's still a need for humans to talk to other humans to generate new business. you know, people aren't to a point where and I don't if they ever will be, maybe they will, maybe they won't, but David: people aren't to a point where and I don't know if they ever will be, maybe they will, maybe they won't, but people aren't to a point where they wanna talk to a robot to buy something. Sure. you know, they if you're gonna make the largest investment in your life building a custom home, you wanna have a human being to talk to. And and so I think they're you know Mitchell: people aren't to a point where they want to talk to a robot to buy something. Sure. but you know they if you're gonna make the largest investment in your life building a custom home, you wanna have a human being right to talk to. For sure. And so I think they're you know David: not only is the AI not there, I don't think it Mitchell: not only is the AI not there, I don't think it David: I d I can't imagine a world where it would be there. It would it would be really difficult to replace the humanity of sales and and transactions in business. There there I think there is a certain human element that would be really hard to to replace. but those humans can become incredibly more effective at doing their jobs. Mitchell: I can't imagine a world where it would be there. It would it would be really difficult to replace the humanity of sales and and transactions in business. There there I think there is a certain human element that would be really hard to to replace. but those humans can become incredibly more effective at doing their jobs. Definitely and I I I think about commitments made on phone calls, commitments made as you're walking around a project site. Details, right? That's details, like the things it hears that our human ears don't hear. the details, it it's that I think that's been the hu the biggest unlock is just, you know, like going through a a a walkthrough with an inspector and and getting that full David: Details. Right. The details, the things it hears that our human ears don't hear. the details, it it's that's I think that's been the hu the biggest unlock is just you know, like going through a a walkthrough with an inspector and and getting that full checklist of things to do and you know, things that weren't really required. Mitchell: checklist of things to do. And, you know, things that weren't really required to meet inspection but are quality concerns that you want to take care of as a company anyway 'cause you care about quality. So, you know, there it can go above and beyond, you know, what's required. David: to meet inspection, but are quality concerns that you want to take care of as a company anyway because you care about quality. So, there c it can go above and beyond what's required. Mitchell: as a segue into a into the topic of disclosure, David: as a segue into the topic of disclosure Mitchell: right? You mentioned, you know, it's important, especially if something's kind of obviously AI, just front run it and let people know, like, hey, this is an AI thing. Have you run into issues, employees, customers, vendors, government officials where David: Important, especially if something's obviously AI, just front run it and let people know, like, hey, this is an AI thing. Have you run into issues, employees, customers, vendors, government officials, where Mitchell: you have r ha I g I guess obviously you've done the right thing and you disclose to people like, Hey, I'm recording this have you felt pushback David: you have I mean, I I guess obviously you've done the right thing and you disclose to people like, Hey, I'm recording this have you felt pushback? I there were some employees that were a little bit unnerved at first that they're that they were being recorded and then they realized that they found out what they could do with that information and now all of a sudden they want everything recorded, you know, like because th the the data is there. If you're recording voices Mitchell: I there were some employees that were a little bit unnerved at first that they're you know, that they were being recorded and then they realized that they found out what they could do with that information and now all of a sudden they want everything recorded, you know, like because th the the data is there. If you're recording voices David: it that's where that's how we convey information. And once they get comfortable with that, they actually they go, man, I really wish we had this conversation recorded because we could really use this and I I think also just one of the things that was surprising to me recently, I had I have a Mitchell: i that's where that's how we convey information and once they get comfortable with that they actually they go, man, I really wish we had this conversation recorded because we could really use this and you know, I I think also just one of the things that was surprising to me recently, I had I have a David: a new client and they were ask we were just talking before a site visit and they were asking me about some of the things that were going on at our company and I was telling about some of our work with AI and Mitchell: new client and they were ask we were just talking before a site visit and they were asking me about some of the things that were going on in our company and I was telling about some of our work with AI and David: I mean they got so excited. Now this was a younger couple that we're building for, but they were just like, Man, that's so awesome. Your company is thinking like that. You're on, you know, you're forward thinking like that. And I was Mitchell: they got so excited. Now this was a younger couple that we're building for, but they were just like, Man, that's so awesome. Your company is thinking like that. You're on, you know, you're forward thinking like that. And I was very pleasantly surprised at how positive their response was. You know, obviously it could go the other way, but I think that their response is probably more in line. David: Very pleasantly surprised at how positive their response was. obviously it could go the other way, but I think that their response is probably more in line Mitchell: with what people are gonna start expecting because the expectations are going up like very much so you missed I I told you in June that this wasn't David: with what people are gonna start expecting because the expectations are going up. Like, you you missed I I told you in June that this wasn't Mitchell: gonna happen until Yeah how how did you miss this deadline? Do you not have a calendar? In our conversation you're like David: gonna happen until December. How did you miss that one detail in our conversation? You're like, you know, cause I have like I've only got five hundred conversations a week. I don't know. I don't know. I'm a human, I forget things. But now like the expectations are going up, not down, because of AI and because you're capturing so much more information, you've gotta have something to help you deal with all that. You just Mitchell: I've only got two years. I don't know, I'm a human, I forget things. But now like the expectations are going up, not down, because of AI and because you're capturing so much more information, you've gotta have something to help you sure deal with all that. You just the I I recently heard this thing about like a human can David: the h I I recently heard this thing about a human can a human can conceptualize and give thought to about seven ideas at a time max and that's a maximum. I'm not seven at a time. I'm probably I think I'm like maybe two. You know, I don't I Mitchell: A human can conceptualize and give thought to about seven ideas at a time max. And that's a maximum. I'm not seven at a time. I'm probably I'm like maybe two. Yeah. You know, I don't I David: I I can and maybe I don't know. Mitchell: I I can and maybe I don't know. I mean David: there's probably times where, ain't great stressful times. I'm thinking about multiple things at a time, but personal business, that stuff and it's all a melange in my mind. but Mitchell: There's probably times where, ain't great stressful times I'm thinking about multiple things at a time, but you know, personal business, you know, that sort of stuff and it's all a melange in my mind. but Once you get over, you know, five, six, seven things, it's like you're not gonna be able to pay attention anymore. You're gonna need some help. David: once you get over, you know, five, six, seven things, it's like you're not gonna be able to pay attention anymore. You're gonna need some help. And you know, you're either gonna need a s person to follow you around and write things down, or you're gonna need an AI, you know, to help you. And and so as Mitchell: And, you know, you're either gonna need a s person to follow you around and write things down or you're gonna need an AI, you know, to help you. And and so as David: I think as this development matures and you know, this this market for AI matures, customers are gonna expect you to be able to manage information more, not less. Mitchell: think as this development matures and this this market for AI matures, customers are gonna expect you to be able to manage information more, not less. And having that constitution is the first step to that. I wanna go back, you you described the positive feedback that you got of when telling a customer about your AI initiatives. How would you deal with an objection? Say say your customer had a apprehension about AI starting from the point of your constitution, how do you reassure that customer that they're actually that they're better off with you who uses AI than Chuck in a truck who doesn't? Well, because David: Starting from the point of your constitution, how do you reassure that customer that they're actually that they're better off with you who uses AI than Chuck in a truck who doesn't? Well, because for one thing, we keep records. we document everything. Mitchell: For one thing, we keep records. we document everything. we have David: We have I don't know. five or six, maybe up to ten daily logs per day now on each job. Mitchell: Six, maybe up to ten daily logs per day now on each job. David: we we're getting stuff from subcontractors, they're updating things. I'm capture before somebody can forget. Yeah, I mean I'm getting Mitchell: Getting stuff from subcontractors, they're updating things. You can capture before somebody can forget. Yeah. I mean I'm getting David: I I I g I've got a lot of information on that job. The Chuck in a truck, he's Mitchell: Chuck in a truck, he's he is he's trying to keep it on his head. I used to do that. Mm-hmm. I know what that's like. I I've I remember I I could keep about three houses in my head. Okay. I could build about three at a time. So your business growth was limited by your my real your your intelligence. I could not go beyond about three houses. I'd start making mistakes. David: He is he's trying to keep it all in his head. I used to do that. I know what that's like. I I've I remember I I could keep about three houses in my head. Okay. I could build about three at a time. So your business growth was limited by your my mind. Yeah. I my mind. I could not go beyond about three houses. I'd start making mistakes. So that's where your margin starts disappearing. Mitchell: So And that's where your margin starts disappearing is those little mistakes. Yeah, the little stuff you forget, you know, forget, I left a piece of rental equipment on a job site. Yeah. I sh should have remembered that. And that starts happening around three jobs. Yep. And I think David: Yeah, the little stuff you forget, you forget, I left a piece of rental equipment on a job site and I sh should have remembered that. And that starts happening around three jobs. And I think that that go I don't think that's industry specific either. I think it's back to those seven things. You know, if you got you got seven things you can think about at any given time. It doesn't take long your work stuff. It doesn't take long to have those seven things. So Mitchell: That that go I don't think that's industry specific either. I think it's back to those seven things. if you got you got seven things you can think about at any given time. You got your family, your personal it done well being and then your work stuff and it's gone. So I David: I think if I had a customer who was pushing back saying, I don't want Mitchell: think if I had a customer who was pushing back saying, I don't want David: a contractor or I don't want to deal with a business that's using AI, you know, well, you know, that you're making a mistake because people who are using AI are keeping better records, they're documenting things better, they're being more efficient, they're keeping costs down for you. the future. Well, Mitchell: business that's using AI, you know, well, you know, that you're making a mistake because people who are using AI are keeping better records, they're documenting things better, they're being more efficient, they're keeping costs down for you. David: there are two. Mitchell: That's the future. Well there are two types of people using AI. There's the there's the people who are, they're going to chat and they're saying, Hey, help me help me price this project out. And then there's people who are taking a constitutional approach, people who are they aren't just using AI, they are thinking about how they use AI. and and with David: the there's the people who they're going to chat and they're saying, hey, help me help me price this project out. And then there's people who are taking a constitutional approach. People who are they aren't just using AI. They are thinking about how they use AI. And and with with that thinking about how you use AI and how your company uses it. Mitchell: with that thinking about how you use AI and how your company uses it. There's so much out there. There's so many tools. There's so many products. How do you how do you not wind up with a document that's dead in six minutes? and how do you deal with approved tools and updating right? 'Cause we I mean we've I've just heard this morning custom GPTs are going away. Right. You know, that's That's where we started. Right. David: There's so much out there. There's so many tools. There's so many products. How do you how do you not wind up with a document that's dead in six minutes? and how do you do deal with approved tools and updating right? 'Cause we I mean we I've just heard this morning custom GPTs are going away. Right. Yeah. Yeah. That's where we started. Right. I mean yeah that's in our Mitchell: if that's in our constitution, what do we do now when our constitution doesn't align David: Constitution what do we do now when our constitution doesn't align Mitchell: with open AI's produ market goals? David: Right. it's a problem. I mean, the d the rapid development, I think you you have to s focus in on the values that you hold for your business. You know, like your the those ethics and values and and the things that really define the culture of your organization. You focus on that stuff because you know, that's the thing you probably don't want to change. Mitchell: think you you have to s focus in on the values that you hold for your business. y the those ethics and values and and the things that really define the culture of your organization. You focus on that stuff because you know, that's the thing you probably don't want to change. The tools are gonna change all around you. You should assume that. Sure. One day they invented excavators, right? That's right, yeah, and it was a beautiful day. David: The tools are gonna change all around you. You should assume that. Sure. Change tool. That's right. Yeah. And it was a beautiful day. but I mean it was like you're you're gonna you're gonna have changing tools are changing and evolving. Technology's gonna evolve, but you just have to ha be like a core in the middle of all that, that that perseveres. And you know, we've already we were experiencing that. We we had Mitchell: But it was like you're you're gonna you're gonna have changing tools are changing and evolving, technology's gonna evolve, but you just have to ha be like a core in the middle of all that that that perseveres. And you know, we've already we were experiencing that. We we had David: we had this constitution that we wrote back in February and and Mitchell: we had this constitution that we wrote back in February and and it David: it it changed because we were you know, at that point in February people were looking at people were concerned whether things were AI produced. hallucinations was a big thing. We were h we were told in February or in Mitchell: It changed because we were you know, at that point in February people were looking at people were concerned whether things were AI produced. Hallucinations were a much bigger deal. Agents weren't really a thing yet. We were h we were told in February or in David: early twenty six that the summer of twenty six would be the time for the explosion of agents. The agents would just be like the thing. And that was accurate. I mean, agents has really taken off. And that changes what your constitution needs profoundly. We needed amendments very quickly. We needed a new constitution because Mitchell: twenty six that the summer of twenty six would be the time for the explosion of agents. The agents would just be like the thing. And that was accurate. I mean, agents has really taken off. And that changes what your constitution needs to deal with. Right. We needed amendments very quickly. We needed a new constitution because David: We had Mitchell: we had David: a policy in our original constitution that if you prompted the AI, you owned the result. And I was, you know, at that point I was like, well, that makes sense to me. Like, you're the one that asked the question of the AI, Mitchell: A policy in our original constitution that if you prompted the AI, you owned the result. And I was you know, at that point I was like, Well, that makes sense to me. Like, you're the one that asked the question of the AI. David: the answer it gives you, you know, but. Mitchell: The answer it gives you, yeah, but Wow, I'm reading over it. It it was so inflexible. Yeah. Right? It was prompting AI is an act of authorship and responsibility. The ind individual who initiates the prompt is the owner of the output and that makes that person responsible for errors, omissions, and misuse. David: so inflexible. Yeah. Right? It was. prompting AI is an act of authorship and responsibility. The end individual who initiates the prompt is the owner of the output. And that makes that person responsible for errors, omissions, and misuse. And you in in this original constitution, accountability cannot be delegated to Mitchell: And you in in this original constitution, accountability cannot be delegated to the system, the AI, the automation. The standard is no one gets to say the AI did it. Right. Does that make sense? That means in September. So then I like wake up or I'm sitting over, you know, having a beverage on a Saturday afternoon or whatever, and I'm sitting there thinking about things David: system, the AI, the automation. The standard is no one gets to say the AI did it. Right. Does that make sense? That made that and so then I like wake up or I'm sitting over, you know, having a beverage on a Saturday afternoon or whatever, and I'm sitting there thinking about things at my business because you know we work on this all the time. And I'm s I'm sitting here thinking like Mitchell: business because we work on this all the time and I'm s I'm sitting here thinking like I don't think that makes sense anymore. I mean we just agentified our business. We brought in an agent. I can't hold Mitchell responsible David: I don't think that makes sense anymore. we just agentified our business. We brought in an agent. I can't hold Mitchell responsible for something the agent says. The agent is thinking for itself now. And so Mitchell might ask a question or he and and this is a real use case. It really happened to me. So I said, Mitchell: for something the agent says, the agent is thinking for itself now. And so Mitchell might ask a question, or he and and this is a real use case. It really happened to me. So I said to the agent that we have, our internal agent, I was like, go ahead and write an email to I don't know, a customer and tell them David: To the agent that we have, our internal agent, I was like, go ahead and write an email to I don't know, a customer and tell them y X, Y, and Z. In my mind, I was saying draft an email. But the word that came out of my mouth was write an email. The agent interpreted that. Mitchell: X, Y, and Z. In my mind, I was saying draft an email. But the word that came out of my mouth was write an email. The agent interpreted that David: As send an email. Now luckily I was like stop. I caught it and I said stop. And the agent stopped before it got done. Mitchell: as send an email. Now luckily I was like stop. I caught it. Yeah. And I said stop. And the agent stopped before it got done. David: And I said and it said, I did not send anything yet, David. So I was like, okay, good. But we've seen that. We've seen where the agent is able to find a way. And it's goal seeking. It didn't do anything wrong. It interpreted the language I spoke. Mitchell: And I s and it said I did not send anything yet, David. So I was like, Okay, good. But we've seen that. We've seen where the agent is able to find a way and it it's goal seeking. It didn't do anything wrong. It interpreted the language I spoke. It didn't hallucinate. It did what it was told to do. Because it was a what we had was a failure to communicate. David: It didn't hallucinate. It did what it was told to do. Because it was a what we had was a failure to communicate between me and the agent. So that made me think my constitution is not fair. Because I cannot hold an employee responsible for using a digital employee. We need due process. Yeah. Well, I thought, well. Mitchell: between me and the agent. So that made me think my constitution is not fair because I cannot hold an employee responsible for using a digital employee. We need due process. Yeah, and well I thought, well, how does it work pre AI? How did that work? Like what we did is we would say employee David: How does it work pre AI? How did that work? Like what we did is we would say employee A reports to employee B. So the boss tells employee B, I need this done, and employee B tells employee A to do it. And so the supervisor is responsible in part to the performance of the employee they supervise. Mitchell: A reports to employee B. So the boss tells employee B, I need this done, and employee B tells employee A to do it. And so the supervisor is responsible in part to the performance of the employee they supervise. And so there's this chain of command. David: And so there's this chain of command. And if you've never worked for anybody before, you will learn about the chain of command. It's one of the first things you learn when you go to work for people. and you will probably do something wrong, and you will talk to the wrong person, and they will remind you of the chain of command. Well, it occurred to me at that point that digital employees. Mitchell: And if you've never worked for anybody before, you will learn about the chain of command. That's one of the first things you learn when you go to work for people. and you will probably do something wrong, and you will talk to the wrong person, and they will remind you of the chain of command. Well, it occurred to me at that point that digital employees need to be in the chain of command. So David: need to be in the chain of command. So if I'm supervising five digital employees and one of them goes rogue and sends a bad email to somebody that's, you know, mean or wrong or whatever, Mitchell: If I'm supervising five digital employees and one of goes rogue and sends a bad email to somebody that's, you know, mean or wrong or whatever, David: I'm the supervisor. I have to take responsibility for that. I'm not I shouldn't I don't think you know my boss should fire me because one of my agents did wrong. Mitchell: I'm the supervisor. I have to take responsibility for that. I'm not I shouldn't I don't think you know my boss should fire me because one of my agents did wrong. I'm probably not gonna get fired unless all of my reports are doing wrong things all the time. That's a management problem that you know you could address organizationally. But if one agent goes rogue and does something wrong, that's like hey, make the David: I'm probably not gonna get fired unless all of my reports are doing wrong things all the time. That's a management problem that, you could address organizationally. But if one agent goes rogue and does something wrong, that's like hey, make the Mitchell: agents stop. Yep. Just let's let's reprimand, let's review policy, let's let's look at the you know, access control, right? Maybe this you know, you David: agent stop. Just let's let's Superman, let's review policy, let's let's look at the Yeah. access control, right? Maybe this you Mitchell: you gave that employee keys to the project home. That employee's now proven to you that they should not have keys. And so maybe they need to be redesigned or reconfigured or and and I mean that's the age we live in now. David: you gave that employee keys to the project home. That employee's now proven to you that they should not have keys. Yeah. And so maybe they need to be redesigned or reconfigured or something. And and I mean that's the age we live in now. Whether or not your company out there in the in the listener world is doing any of this, it you will be. You will be. You either will be or you will not be. And if you're not doing it, you will not be. And and the Constitution is a good place to start about this. Again, what is your company willing to do AI? Mitchell: Whether or not your company out there in the in the listener world is doing any of this, it you will be. You will be. You either will be or you will not be. And if you're not doing it, you will not be. And and the constitution is a good place to start thinking about this. Again, what is your company willing to do with AI? What is it not willing to do? What risks are you willing to take? We had David: What is it not willing to do? What risks are you willing to take? We had So d just this living document. The four we had four pillars in our or s six pillars in our original one. but in this most modern version, the one it actually I it's I I haven't ratified this one yet. I've again I've I've created this Mitchell: So d just this living document. The four we had four pillars in our or six pillars in our original one. but in this most modern version, the one it w actually I it's I I haven't ratified this one yet. I've again I've I've created this, And it's much longer than the original one. It's much more in depth. David: and it's much longer than the original one. It's much more in depth. Mitchell: And I haven't ratified it yet and shared it with my employees. We're still operating under the old constitution, the version one. And the reason I haven't ratified it is honestly I haven't read it. David: And I haven't ratified it yet and shared it with my employees. We're still operating under the old constitution, the version one. And the reason I haven't ratified it is honestly, I haven't read it. I have not read this yet for accuracy to what I want it to be. So it will not be entered into the corpus of our agents until I, the boss Mitchell: I have not read this yet for accuracy to what I want it to be. So it will not be entered into the corpus of our agents. Right. Until I, the boss, decide that it's going to be the constitution that our company follows. So that's back to our point. Yeah, I was gonna say that's that's full circle. You better read it and make sure because there are And so you can use a David: decide that it's going to be the constitution that our company follows. And so that's back to our point. you gotta read it. You better read it and make sure because there are you can use AI to help write these const of words in here. I mean there are, you know, retrieval and SOPs and Mitchell: Retrieval and SOPs and silence is not approval and the left column is reviewed and the right column is unreviewed. I mean there's so many nuances you better pay attention to what you're putting out there. And but I did notice here that there are four consequences running through everything in this document. David: silence is not approval and the left column is reviewed and the right column is unreviewed. there's so many nuances. You better pay attention to what you're putting out there. And but I did notice here that there are four consequences running through everything in this document. Authority is granted by role, not per action. That's our point with you you're a supervisor, you have the authority to run these digital assistants. Mitchell: Authority is granted by role, not per action. That's our point with you. You're a supervisor, you have the authority to run these digital assistants. If something happens with those digital assistants, your authority is to take care of it. your supervision is proportionate to consequence. if something David: If something happens with those digital assistants, your authority is to take care of it. your s supervision is proportionate to consequence. if something if s if what can hurt someone gets reviewed before it before it issues. So before you go out there and do stuff that's gonna be consequential to people, you gotta Mitchell: If s if what can hurt someone gets reviewed before it before it issues. So before you go out there and do stuff that's gonna be consequential to people, you got you're a supervisor. You gotta make sure it's okay. So as an example, you you you would say r today David: You're a supervisor, you gotta make sure it's okay. So as an example, you you you would say today the agent may be allowed to put together a a a request for bid on a on a load of lumber. It's not allowed to execute the purchase. That's right. That's right. Because if it buys it, I mean that could hurt us. Like maybe we're not ready to buy it. Maybe we don't have a a big Mitchell: the agent may be allowed to put together a a a request for bid on a on a load of lumber. It's not allowed to execute the purchase. That's right. That's right. Because if it buys it, that could hurt us. maybe we're not ready to buy it. Maybe we don't have a a big David: receivable coming in yet that Mitchell: receivable coming in yet that David: we're waiting on to buy that lumber package. So there's a lot of things that are happening in a business that require supervision. And that doesn't go away. it if anything, I think it expands and increases. people become they have more to to supervise. And does seem like adding AI Mitchell: we're waiting on to buy that lumber package. So there's a lot of things that are happening in a business that requires supervision. And that doesn't go away. No. it if anything, I think it expands and increases. You know, people become they have more to to supervise. And It does seem like adding AI to a job, at least today David: a job at least today in the same way you're you described consumer expectations rising really to start implementing AI into your job you have to start taking your job more seriously. You have to to become a better professional. Man, you've got to you've got to spend some time and you've got to give your space give yourself space to do it. Like you have to spend time Mitchell: in the same way you're you described consumer expectations rising, really to start implementing AI into your job, you have to start taking your job more seriously. You have to to become a better professional. Man, you've got to you've got to spend some time and you've got to give your space give yourself space to do it. Like you have to spend time David: thinking about your job. Yeah, everything. And you gotta let your employees have that space too. Like somebody needs to sit down with them and and say, what do you do? Yeah. And Mitchell: thinking. About your job, right? What is my job? What do I do? And you gotta let your employees have that space too. somebody needs to sit down with them and and say, What do you do? Yeah. And David: and we're gonna talk about that late you know at some point, but like you have to give people space to think. You can't just throw this at someone and expect Mitchell: For sure. At some point. But you have to give people space to think. You can't just throw this at someone and expect miracles overnight. This is not it's not a miracle solution. It it's a wonderful solution. It's very powerful. It's a tool right now. People have to have David: miracles overnight. This is not it's not a miracle solution. It it's a wonderful solution. It's very powerful. But but people have to have space and bandwidth to be able to to think about these things. And learn these new topics Yeah and I've got like I've got employees that are be like, I was Mitchell: space and bandwidth to be able to to think about these things. And learn these new topics. And I got like I've got employees that are be like, I was waiting on my husband at the doctor's office and I prompted the AI about a problem I had at work and this is what it said and I thought it was a great idea. What do you think, David? And David: Waiting on my husband at the doctor's office and I prompted the AI about a problem I had at work and this is what it said and I thought it was a great idea. What do you think, David? And I love that. I love that people are, engaging d in downtime with the AI on to help them through things that are giving them trouble at work. Like the and they're finding solutions. That's great. Mitchell: I love that. I love that people are, engaging d in downtime with the AI on to help them through things that are giving them trouble at work. Like the and they're finding solutions. That's great. And that goes back to the idea of that, these are these can be assistance to your employees and and a thought partner. David: These are these can be assistance to your employees and and a thought partner and all of that's necessary as you as you look to get better at your job or if the t if the task is to automate the job, understanding the job really is the first Yeah, that's right. And the a third thing the consequence in this is that trust is earned on a schedule, not assumed. Mitchell: And all of that's necessary as you as you look to get better at your job or if the t if the task is to automate the job, understanding the job really is the first first hurdle. That's right. And the a third thing, the consequence in this is that trust is earned on a schedule, not assumed, David: not withheld indefinitely. It's just like a new employee. Mitchell: not withheld indefinitely. It's just like a new employee, David: when you Mitchell: When you excuse me, when you hire someone new David: Excuse me. When you hire someone new. you you don't give the keys to the building the first day. Right. You know, you you you don't you don't let manage the till on the first day, you know, like you gotta get people you y trust is earned over time. It's the same thing with these agents. Trust is earned to a trap, right? Because the agents present genius and you're like, I can definitely trust this because they're smarter than we Mitchell: you you don't give them the keys to the building the first day. Right. You know, you you you don't you don't let manage the till on the first day. You know, like gotta get people you trust is earned. Okay, exactly. It's the same thing with these agents. And it's easy to fall into a trap, right? Because the agents present as genius and you're like, I can definitely trust this thing. Because they're smarter than we are. David: are. I mean th the But then it comes out Mitchell: But then it comes out and it says, Hey, we should definitely David: Hey we should definitely Mitchell: bring up the hyphenated last name of a customer and you're like, Yeah, I don't know about that. Yeah, that exactly that happened. It's y the the agents I think it's really good to like bring them on David: I don't think we need to bring it up. I mean, just do whatever they said. Like, yeah, that exactly that happened. it's the the agents, I think it's really good to like bring them on in a in a sm slow, steady manner so you kind of test the waters a little bit. You don't just jump right in, you put your toe in, you put your you go knee deep, you can you know, go waist deep. Yes, you Mitchell: Go start with your most well understood, well observed water easily and David: wade into the water easily and and then the final consequence is that the rules must be knowable. And I think, you know Mitchell: And then the final consequence is that the rules must be knowable. And I think, you know, this is an area where we're we're touching on some of this, so we're gonna talk more about this in the future, but you gotta have some SOPs. You've you've gotta have David: This is an area where we're we're touching on some of this, so we're gonna talk more about this in the future, but you gotta have some SOPs. You you've gotta have stuff people who aren't into the acronyms, what's an SOP? What does it what does it actually get you if you have it? It's a standing a standard operating procedure. So it's it's how people are doing their jobs and it's how tasks are being achieved. and goal Mitchell: For people who aren't into the acronyms, what's an SOP? What does it what does it actually get you if you had one? It's a standing a standard operating procedure. So it's it's how people are doing their jobs and it's how tasks are being achieved. and this is how we answer customer calls or Yeah, this is you know we have one for building a house. we actually have the entire process. David: Yeah, this is you know we have one for building a house. we actually have the entire process defined. you know, we have one, like you said, how to answer a call, how you know, how your voicemails should sound, how we how often we pay bills, you know, how we deal with sales process. Mitchell: you know, we have one, like you said, how to answer a call, how you know, how your voicemail should sound, how we how often we pay bills, you know, how we deal with sales process, how we build the selections for a a given house. So like we have all these SOPs that we've written David: How we build the selections for a a given house. So like we have all these SOPs that we've written, Mitchell: And that is part of this brain, this hybriding brain that we can use to to share with other employees. You know, one of the big things of several years ago was like cross training, you know, getting David: and that is part of this brain, this high breathing brain that we can use to to share with other employees. You know, one of the big things of several years ago was like cross training, you know, getting getting one employee to know Mitchell: getting one employee to know getting one employee to know what another employee was doing. Mm-hmm. And and so, you know, being able to if you if the rules are knowable, then everybody can be on the same page at any given moment. And and I think, you know, that's a huge thing, with the constitution. And involving your employees in writing these SOPs I think is a good step David: Getting one employee to know what another employee was doing. And and so, you know, being able to if you if the rules are knowable, then everybody can be on the same page at any given moment. And and I think, you know, that's a huge thing, with the constitution. And involving your employees in writing these SOPs, I think is a good step. towards what we're gonna talk about in the next episode, which is how do we get our employees to take more ownership? How do we get them to really care about the job the same way, for example, you care about the job. Mitchell: towards what we're gonna talk about in our next episode, which is how do we get our employees to take more ownership? How do we get them to really care about the job the same way, for example, you care about the job? David: Yeah, you gotta we gotta figure out how we're gonna I we're sitting here w today's episode we covered the whole thing about how to build a constitution. There's a reason for that. That's where it starts. And the next thing is getting people on board. And so Mitchell: Sitting here with today's episode, we covered the whole thing about how to build a constitution. There's a reason for that. Yep. That's where it starts. And that's and the next thing is getting people on board. And so we'll cover that next week. Yep, definitely. that's awesome, man. Yeah, I love it. David: We'll cover that next week. Yep. That's awesome, man. I I love it.