Why AI Made You a Faster Bottleneck (And the Decision Layer Fix)
About this episode
Many service-based founders hit a growth ceiling with AI, not because they chose the wrong tools, but because they are layering AI onto a system that still requires their judgment at every revenue-critical transition point. Better tools just make the founder a marginally faster bottleneck.
This episode of The Growth Ceiling breaks down the three handoffs where founder dependency lives in every service business: lead to pipeline, pipeline to onboarding, and onboarding to delivery. In most businesses between $1M and $5M, the founder is the decision engine at all three points because no one ever documented the criteria for what "good enough" looks like at each transition.
The distinction that changes everything is the difference between task layer AI and decision layer AI. Task layer AI writes proposals faster and summarizes meetings. Decision layer AI qualifies inbound leads against documented criteria and routes them without the founder touching them. Most businesses have a full stack of task layer tools and nothing in the decision layer. That is why the founder bottleneck stays in place and the growth ceiling does not move.
Nate and Simone walk through how to map your three handoffs, document your actual decision criteria, audit your current AI stack against the decision layer, and build quality thresholds your team can apply without you. This is not a technology project. It is a documentation project that technology accelerates. The episode also covers what AI should never replace: key client relationships, strategic judgment, and genuine novel situations.
If AI has not changed the structural dynamics of your business systems, the problem is not the tool. It is the architecture underneath it. Take the Growth Ceiling Assessment to find where your real constraint lives.
- [00:00] This Week's Growth Ceiling: why a founder's "visibility problem" was actually a positioning problem with two businesses under one brand
- [04:27] The misdiagnosis most founders make about AI adoption (and why "wrong tools" is the wrong answer)
- [06:01] The three revenue-critical handoffs where founder judgment gets embedded: lead to pipeline, pipeline to onboarding, onboarding to delivery
- [16:20] How one founder turned "gut instinct" lead qualification into documented criteria an AI could score against
- [19:57] Why 70 to 80 percent of your client onboarding is identical (and how to find the repeating pattern in your last 10 clients)
- [28:46] Task layer versus decision layer: the one-question audit that shows whether your AI is in the right place
- [40:13] The three moves you can make this week: map your handoffs, document your criteria, audit your AI stack
If this episode made you suspect your real constraint is not what you have been treating it as, click here to take the Growth Ceiling Assessment. It takes about five minutes and shows you where your actual constraint lives in your growth engine.
Subscribe to The Growth Ceiling wherever you listen or watch. And if this episode helped you see something differently, send it to one founder who needs to hear it.
Simone Henry: You added AI to your business this year. Maybe a few tools, maybe a whole stack. And the honest answer is your business does not feel lighter. It might even feel heavier because now you're managing the AI on top of managing everything else. This is not a tools problem. This is an architecture problem. Today We're going to break down something most founders are getting backwards about AI. AI implementation in a service business is not a technology adoption decision. It's a systems architecture decision. And the founders who get it right are not asking, what can AI do? They're asking, where does my judgment still sit in the middle of a revenue critical handoff? And how do I get it out? We are gonna walk through the three handoffs where founder dependency lives in every service business and show you how AI belongs in those handoffs. Not bolted on to the edges of your task list. This is the Growth Sealing Podcast. Welcome. We are Nate and Simone. Thank you for joining us today, and let's get started.
Nathan Grossman: Yeah, so here's the pattern we're talking about. ⁓ founders in that one million to five million range adopting AI, they they they adopt AI enthusiastically. They use it to write their proposals faster, they summarize their meetings, they draft emails, generate content, come up with strategy plans, etc. Useful stuff, right? But six months later, nothing structural has changed really. ⁓ the founder is still the person who decides which leads move forward. And ⁓ s and th they're still the person who decid who designs every onboarding sequence by hand, still the person the team checks with before a client deliverable ships. ⁓ the AI made the edges of the business faster, but it didn't really touch the center, in other words. The reason most founders are misreading this is because AI vendors and AI content sell the tools and not the architecture. So the the founder evaluates the AI tool ⁓ by tool instead of asking where in my revenue system is my personal judgment still required for something to move forward. And that's the bottleneck, not the task list. What's at stake here? Every quarter the business grows. The number of decisions that require the founder's involvement grows with it. Having AI layered on top of that system makes the founder a slightly faster bottleneck. it doesn't necessarily it's it's like what ⁓ what they say hurry up and slow down or slow like that. ⁓ it it does not remove the bottleneck. The the ceiling stays in place, in other words.
Simone Henry: Right. So here's how it shows up operationally. The team has AI tools. They're using them, but the workflow still routes through the founder for approval, for context, and for just that quick gut check. The tools are active, but the dependency is unchanged. It's still, it's still in there. That's the problem.
Nathan Grossman: Yeah. So ⁓ let's get into our first segment here. what is the nature of this misdiagnosis? So you know, using the using the the the language of diagnosing and prescription, right? It sounds a lot like a very clinical, ⁓ you know, doctorish, if you will. But Yeah. But you know, it's it's kind of like that, you know, like this is it's it's an apt ⁓ metaphor, I think, in this case. ⁓ a lot of the diagnoses you know ⁓ revolves around symptoms what are the what are the what are the symptoms what are the those outward facing things that you know ⁓ really bug the the ⁓ the the business owner the founder ⁓ what are the the what is the the the actual nature of the friction there so for example we adopted ai but it's not really moving the needle on growth ⁓ maybe we picked the wrong tools or you know they start they start s second guessing right maybe we picked the wrong tools or we need a better prompts or the team is not using it well enough something ain't right here in other words you know and so that that tends to be the the default diagnosis. So it's a tool selection adoption or excuse me, a tool selection or adoption problem. And the fix in that case then accord you know, which makes sense, it's logical, right? Evaluate more tools. Maybe it's the nature of the tool itself, so we need to a actually ⁓ figure out if we're using the right tool. So we need to go and test other tools, AI tools, or we need to get more in-depth in our training. or hire someone who gets AI, ⁓ you know, ⁓ a consultant, as they as they say, which are a dime a dozen these days. You know.
Simone Henry: Yeah, a lot of people are becoming AI consultants and there's lots of training programs to train people to become AI consultants. and so some are better than others.
Nathan Grossman: Yeah, yeah. Which is yeah, I I can understand why, but anyways. The the re so this diagnosis is wrong and here's here's why. It treats the AI as the variable. The real variable is the system that the AI is being dropped into, right? If the revenue system is built around the founder making all the judgment calls at every critical transition point, having better AI tools just makes the founder marginally faster at becoming the bottleneck, you know? ⁓ So the trick is thinking about it structurally. A service business has three revenue critical handoffs. First one is lead to pipeline. When does an inbound lead become a real opportunity and who decides that? The second one is pipeline to onboarding. When does a closed deal become an active client and who designs that transition? And then onboarding to delivery. When does a new client move from setup into ongoing work and who holds the quality standard in that case? So in in many, you know, ⁓ one million to five million service businesses in that case, the the founder's judgment is embedded in all three of those. ⁓ Not because the founder wants to be there necessarily, but because no one ever designed a system that could take could could make those calls without them. And when you layer AI onto that structure, you get AI writing better outreach emails, which is useful, but it doesn't change who qualifies the lead. You get AI summarizing discovery calls, which is useful again, but it doesn't change who decides that the deal is ready to close. And you get AI generating onboarding checklists, which again is useful, but doesn't change who checks the work before the client sees it. The tools are doing the tasks, they're not replacing the judgment, and that is the misdiagnosis in this case.
Simone Henry: Interesting. ⁓ I always like to say like even even in this we're in this age of of very fast ⁓ and very powerful technology but business systems and business rules still need to apply, right? So There is an operational tell that gives away this misdiagnosis every time. Ask your team a question. If I disappeared for two weeks, which revenue critical processes would stop or stall? Not slow down. Stop. Actually stop. If the answer includes anything in lead qualification, lead progression, or client onboarding, you don't have an AI problem. What you have is a system design problem and a system that's designed with the founder in the middle that can't where everybody comes to that founder for that decision. The AI is just sitting on top of that process with and no tool can fix that.
Nathan Grossman: Yeah, exactly. And to clarify, what what Simone is saying is that ⁓ you know what What we see in most of the businesses that we look at is that the founder doesn't even realize that they are the bottleneck in these handoffs. They think they're staying close to the work or they're maintaining quality, but what they're actually doing is serving as the system's decision engine. Every handoff routes through their judgment because no one ever codified what good enough looks like at each transition. And s yeah.
Simone Henry: Yeah. We had a a a chat about this in in a previous episode where we talked about we talked about ⁓ founders needing to really empower their employees to make some critical decisions. Every decision should not be on the founder's shoulders.
Nathan Grossman: Yeah.
Simone Henry: So, you know, it it the founder isn't necessarily doing tasks, but being but being in the center and making all the decisions means that all the tasks can necess can't necessarily be done without their say-so and without their without their decision making. So ⁓ whether you're using AI or you have employees. they need to be empowered to make some of these decisions so that the founder doesn't become that bottleneck. The founder can ⁓ can release the founder from some of these ⁓ some of these decisions.
Nathan Grossman: Yeah. That's the the key distinction here. That we we want it to get to the place where the founder is not doing the tasks. ⁓ you know, they're making the call. The AI can automate those tasks, you know, much easier, much more efficiently. replacing a judgment call though requires a different t kind of implementation. and what this boils down to is ⁓ a viable problem, right? So we talk about visible, viable, and valuable. Having a business that's visible, viable, and valuable. So making one of these steps in making your business more viable is pipeline management, conversion, and those handoffs between the sales and delivery. All of that is viability mechanics. And when you have a business owner's the judgment, their judgment is embedded in those handoffs, then the business has a viability ceiling that no amount of visibility or value can overcome. So yeah, the founder's judgment is what we're talking about here, and how it's embedded into three those three critical handoffs. So here's here's how it works, mechanically speaking. Every revenue critical handoff has three components. A trigger is number one. What event or data point signals that something needs to move from one stage to the next? So you we that you know marketing gets ⁓ a lead in. That's great and everything. Presumably they have automations that are moving them down, you know, through a down a path, but You know, what what is it along that way? How how is the decision being made as to which path it's going down, right? That's that's something
Simone Henry: how does the system decide which lead is a good lead, which lead is a which lead is a cold lead, which lead just kinda needs to go into long term nurture and you know, maybe and and down the down the path of of ⁓ automated sales or or low ticket sales rather than actually getting on a call with somebody, say.
Nathan Grossman: Right. Yeah. So it's the decision, right? What is the criteria there that determines whether it's ready to move to the next stage and what is that stage? And then, of course, continuing on that path, what happens next once that's made, that decision is made. So if the founder is in a founder-dependent system, the founder is the decision layer in that case. The trigger happens, so a lead comes in, or a deal closes, or a client finishes their onboarding. And the system routes to the founder's inbox for approval, right? They're their Slack or whatever it may be. Maybe they just need a quick review before anything moves. However that, however you want to word that, whatever that may look like in your given circumstance. But what what essentially has happened is, you know, you bolted on AI to the edges and it's handling the maybe handling the trigger better. ⁓ maybe it's able to you you're using AI to create that initial outreach, you know, which is great and everything. You know, that's that that part of it is is faster. So you're you're you're getting faster to capturing that lead, and maybe those ⁓ notifications are automated, which is great, you know, efficiency. or maybe you know your proposals are getting generated faster, ⁓ or your your tasks are being automated, which again great, fantastic, but the final decision that stays with the founder. And So that could be seen as a bottleneck. And that's where ⁓ AI can really if you're if it's properly implemented and architected, that could really ⁓ be useful. ⁓ all right, so let's go through a like a hypothetical handoff here. Let's start off with the first one. That from going from lead into okay, there are prospects on our pipeline. And Most founders many or many founders personally qualify leads. they look at the inquiry and they make a gut call about the fit and decide whether to invest their time in in a discovery conversation. And ⁓ the AI version of this is not a chatbot on the website necessarily. It could look at like scoring and routing systems that codify the founder's qualification criteria into into a decision framework. ⁓ I know, for example, of ⁓ someone who owned a business that was surveying, right? They did land surveying. And Somebody described pricing to me as kind of like a dark art one time, and I love that description. Pricing as something of a dark art. But in this case, the the business owner wanted to be free of the decision making that went into pricing out ⁓ the work involved with surveying. And there was a number of different variables involved, right? As you can probably imagine. I mean you got like elevation concerns, you know, total acreage, I don't know what all goes into that. Are there obstructions, you know, things in the way? Is it easy access? I don't know. ⁓ there there's a ton of other of like minute decisions that go into how hard is this gonna be for us to actually accomplish this, right? And so He was able to kinda work with someone to translate that into an AI workflow where you so so and part part of this was too, speaking of the dark art, right? ⁓ a a lot of a lot of what goes into the pricing in that situation could be considered proprietary. And you don't want just any Joe on your team to understand, fully understand what's going on behind the scenes. Because what if they left? Right? They could go to the competition and voila, all of a sudden they've got your special sauce over here and able to to to to to basically recreate the the whole thing, right? So in this case it was protected by the AI. in in a in a ⁓ GPT where the back the back end was not accessible to the the workers who were inputting the data up front. And the AI would walk them through the information that the owner knew would help in determining the actual pricing on the other side. So this this is this can be done, it's effective. Anyway, so ⁓ another other things to consider here. So s scoring, routing, what industry, what's the revenue range, ⁓ what's the problem description, what are the buying signals involved. So when all that criteria gets defined and it's weighted, the AI can start to make the first pass qualification decision and route accordingly, without the founder touching it necessarily.
Simone Henry: ideal for for ⁓ an automated system where you are you're bringing in leads on on autopilot and having a system that can make that first pass qualification decision and route and route the lead to the right to the right location is is very good to have. ⁓ that's something that will definitely free the founder from from that as at least that initial decision. Right. So, ⁓ when it comes to to installing AI in your com in your in your company, ⁓ This is an operational piece. So oper when most founders miss it when they're just thinking about, ⁓ well what AI tool is gonna make my life easier? Well, you know, th they're s looking at this operational decision point, ⁓ is that they have to think about, like they can't automate a qualification decision that they've never written down. If there's no system for it, if everybody has to come to the founder and the the founder only knows it, they know it in their heads, but nobody else knows it, then you know, it it's not written down and everybody has to come to them. So the founders qualify the leads. They're they're doing it on instinct. They just know, ⁓ this is a person's a good fit. ⁓ Or that person's a a a bad fit when when they see it, right? But and yes, that gut instinct may be real, but it's not a system. There is some mathematical formulas, there is there are some decisions being made that are logical, that can be written down, that when you think about it, ⁓ okay, yeah, we can actually systematize this. I had I had this with with a client of mine You know, she she ⁓ in her onboarding process she had ⁓ she had her her clients fill out an assessment form. And she said and she said, Well yes, I you know, I'm gonna approve them or disapprove them and You know, and I said, Okay, well we can automate that. ⁓ and so the our first thought was send her send her the assessment. When the assessment comes in, notify her she she looks at it and then she can, you know, she ticks a box and says, Okay, this is approved. But then talking to her further, I realize that ⁓ I asked her a question, I said, Well, wait a minute, how many of these do you approve or how many do you disapprove? He said, Well, I don't really disapprove that many. ⁓ really? Okay. So that means most people are approved. So in the few cases where you do disapprove somebody, well, what is that? Like why do you disapprove them? ⁓ well, they have to be this age and they have to be and they have to be a have this ability. They have to be able to do such and such. I said, Well, now that we know that We can then automate it. Right? And now it doesn't have to go through you at all for disapproval or approval. You know? so yeah, you may be going on gut instinct, but that's not a system. So first step Isn't about picking the I the AI tool, it's sitting down and documenting the actual criteria that you're using to decide whether a lead is worth your time or not, right? There are criteria. You know, weight that criteria. That's your foundation. And then the AI comes after. Now, once we have that criteria, now we can say we can give the AI something to look for, right?
Nathan Grossman: Yeah, yep.
Simone Henry: You know, does the person have this kind of a budget or this kind of an income? Well, if they do, they can go on to the next step, right? Or if the person is this age or they have this kind of certification or something, okay, great. They can go on to the next day. Those are things that the that the AI tool can actually can actually see.
Nathan Grossman: I'm gonna drop drop a ⁓ a truth bomb here maybe or give give everyone out there listening or watching a an inside tip. If you have thoroughly thought through your ideal customer profile, this shouldn't be too difficult because many of those points are going to appear here and that is what you're going to wait, you know. So might want to start there if you haven't already. documenting documenting it, have it written down, and then wait it. okay. So once we have taken we've figured out what are what the criteria are that are letting us know that, you know, this is a good lead and we're going to pr you know pursue them. the next step would be when once they've gone through that pipeline, whatever that may be, to they're we're ready to onboard them as a paying customer. And this is where a lot of service businesses lose their momentum. So their deal the deal closes and then everything slows down because the the the founder has to design or approve that onboarding sequence for each new client. the the scope, the timeline, the team assignments, cla and then client communication, what all that's going to look like is you know starting from scratch every time. And The founder's in the middle of it all because every client is different. You know, again, going back to what I was talking about earlier about the ⁓ the surveying company and that whole situation with figuring out the pricing and all that kind of stuff, you know, a lot of different elements go into that. And, you know, they were having to start from scratch every time, and so they wanted to systematize it, and that's exactly what we're talking about here. So here's a way to think about this differently. Every client may not be as different as you think. In most service businesses, 70 to 80% of the onboarding sequence is identical across the clients. And the the remaining 20 to 30% is where ⁓ any actual customization may occur. And AI implementation at this handoff means you're building a templated onboarding system that handles the 70 to 80 percent automatically and surfaces only the count customization decisions to the founder or a senior team member for the approval in that case. So we're not completely eliminating, we're not talking about completely eliminating human intervention here, which would not be wise, okay? We're just saying that there's probably more that could be automated and handled by AI than you actually are implementing now. So yeah.
Simone Henry: This came up when we were talking about ⁓ productizing. and even even in a service-based business where you have where there is lots of customization, there is there are still opportunities to to systematize and productize. So,
Nathan Grossman: Yeah.
Simone Henry: you know, you want your you want your ⁓ your operational system to move your leads and clients through a a set set of steps. Now within those steps there may be some customization, but the system as a whole, you know, you're you're bringing your client and like Nate said, ⁓ knowing your ideal client avatar is very is very crucial here. Everybody is going to come in at at one place and you're and gonna end up at another place. And within those those steps along the path is where that customization lives, right? So ⁓ at an from an operational standpoint, you wanna map out Your last 10 client onboardings side by side, highlight every step that was identical. And ⁓ then think about and highlight where the customization came in. And you'll probably see that there's not as much customization as you think. There may be some decision points, but not so much customization. Right? ⁓ most founders are shocked at how much repetition there really is in their business. And so once you see it, then implementing AI becomes pretty obvious. You can automate those repeating steps and build decision logic along along the path where where clients are diverting. and where clients need to make you know a decision here or here or here, right? But they're still moving along the same path, right? So what most founders really have is a process that requires them to redecide every single time that a new client starts. And that doesn't necessarily have to be the case.
Nathan Grossman: Yep. Okay. And ⁓ Again, here we the c the key thing to remember and one of the things that we are constantly, you know, talking about and preaching is The the time the the the transition from when they are they're they're going from being a prospect to a paying customer is a critical transition. And you get you really need to nail that down and get it right. In order because you only get one chance to make a good first impression, right? And I know that the the technically the first impression was your first meeting, which is, you know, don't get me wrong there, but going from prospect that's the that's the next best right yeah and so you really you really need to get that right because that's when their the trust in you is either confirmed or they start questioning things right so you don't want them to start questioning things in other words. Alright so now once we've done that we're talking about going from that onboarding process into the actual delivery.
Simone Henry: Like the the next first impression, right?
Nathan Grossman: And this is the that third handoff. And this is where the quality of the experience, all that anxiety is living, right? This is where the business owner is 100% convinced that no one can do it like them. So that yeah. The client gets onboarded, the work is underway, and then the the business owner can't stop checking in because they're worried that the team is not going to deliver up to their standard. This happens all the time. I was just talking to somebody here recently, ⁓ owns it owns a business here locally to me. and he's, you know, basically saying that ⁓ you know, just can't find anybody these days that's willing to work and do good work and do it the way I want it to d be done, you know. Like he says when I when I when I when I talk to my workers the first day, basically I want them all to understand that what I say is gospel and should be followed to the letter. You know, like and and if you can't do it that way then you don't belong here, basically. So y This is like, you know, ⁓ the emith, right? That book when it when it gets into that. Where, you know, the this is the the pitfall that entrepreneurs fall into where the they are so invested blood and bone into their business that they're sure that no one else can do it like them, which is probably true. Okay, don't get me wrong. Nobody is gonna look at your business like you do, you know, like like like the ⁓ the business owner.
Simone Henry: Yeah, it's probably true. Exactly. Exactly.
Nathan Grossman: You basically have to be comfortable with accepting like seventy to eighty percent of your actual dedication and input, right?
Simone Henry: Yeah. I know a few coaches that that teach that, no, nobody can do things the way you do it. And what it will probably take is you'll probably end up getting two people to do eighty percent of w of the the quality that you can. Right? You know? ⁓
Nathan Grossman: Yeah. Yeah. Yeah. Two people doing forty percent each, right? Yeah. Yeah.
Simone Henry: Right. And it's very interesting 'cause I'm actually finding that out now, ⁓ with with my team, you know, and I'm thinking, ⁓ my goodness, this is all
Nathan Grossman: Yeah. Nobody's gonna be as invested in your business as you, you know, like yeah.
Simone Henry: This is not that difficult. Why is it so hard? Just do it this way. But but you have to go through those frustrations because otherwise you can't you can't grow. If it all depends on you, you can't grow.
Nathan Grossman: Yeah. Mm-hmm. Yeah. What ends up happening is you know, it's just the the business owner, the founder, ⁓ ends up becoming a bottleneck because they're so convinced that the the experience, the quality of the experience is gonna be low, right? They're very concerned about that and it's understandable, I totally get it. You want to make sure that this client is getting the best possible experience because They are going to tell other people about you and you want them to tell, you know, give good glowing recommendations and the yeah, it's it's totally understandable. But you know, it starts to gum up the works, you know. they're reviewing all the deliverables, they s they're sitting in on the client calls, they're inserting themselves into every feedback loop just to make sure, you know. And the AI opportunity here is not about generating deliverables, it's about building a quality standard into the system so the founder doesn't have to be the quality control layer. And so that means defining what good enough looks like at each milestone. ⁓ Building a review criteria that the team can apply without the founder. whatever that may look like, and using AI to flag the exceptions rather than review everything. You know and so once you have this installed, you start testing it out and you know you're working on tuning the AI and training the AI as to what are the things that are falling through the cracks, let me know immediately and notify me. That's something that's doable. So you don't have to constantly be there.
Simone Henry: Yeah, so this is that ⁓ that quality quality threshold system. ⁓ you define your criteria and at each delivery milestone don't make it vague, don't make it ⁓ high quality, client ready, that kind of thing. Make it measurable. There should be numbers that the AI can actually check and and you know, your AI can then serve as a first pass reviewer. It knows exactly what to look for. So it knows that, okay, does this particular deliverable meet the meet the standard or no? If yes, well then it's gonna move to the client. If no, well it gets flagged and now a human can review it. So this it doesn't mean taking a human out of the process altogether and and letting AI just kinda run everything and it also doesn't mean that AI is not gonna work for you. It can absolutely work, but it's about working the AI and your your human team working together to make sure that your client gets a gets quality deliverables. Right? So now the founder only needs to see exceptions rather than You know, if everything is working nice and smooth, the founder doesn't necessarily need to worry about it. It's only gonna see the r he's only gonna see the recep the the exceptions ⁓ and and things that will land on his desk because it's out of the norm. And maybe then ultimately the founder makes a decision, well, okay, let's add that as part of the criteria. And so ultimately then the founder can
Nathan Grossman: Yeah.
Simone Henry: to see there won't be any exceptions if you're if you're really like continuing to massage and and update your system, right?
Nathan Grossman: Well, I don't know if never, but ⁓ you know ⁓
Simone Henry: Well, not not never, but like but then you can you can have it set up where the founder doesn't necessarily even need to make those types of decisions, even on exceptions because you know, the team is able to do that. Right? Because yeah.
Nathan Grossman: Yeah. So here, yeah. Here's a question that you can answer in 60 seconds that's gonna tell you whether this applies to you. Look at your calendar from the last two weeks, and you're gonna count the number of times you reviewed, approved, or made a judgment call on something that was moving between the different stages that we've talked, we've been discussing today. maybe it was a lead that you qualified, a proposal you reviewed, an onboarding plan that you customized, a deliverable you checked before it went into went to the client. If that number is higher than five per week, your AI implementation is the wrong is in the wrong place. You're using AI to be a faster version of yourself instead of building AI into the system, so the system doesn't need you at those transition points, or at least doesn't need you as much. Okay. So let's think about this. How can we think about this a little bit differently? basically stop asking what can AI do for my business. You know that was it John F. Kennedy, don't ask, ask not what your your country can do for you, but what you can do for your country. Yeah. ⁓ kind of kinda in that same vein here. what can AI do for my business? And start asking so stop asking that. Stop asking what can AI do for my business? And start asking where does my business still require my judgment for something to move forward? those are r really different questions. So what what can AI do for my business? Stop thinking about that. Where does my business still require my judgment for something to move forward? The first one ⁓ leads to tool adoption, and the second one leads to system architecture. where you know maybe the tool is a lot a little less important in that case.
Simone Henry: But ultimately what happens really, when you're thinking about that system's architecture, that actually helps you to dis to determine what tool you need, because all tools are not created equal.
Nathan Grossman: AI implementation. Yeah. Yeah. That's true. That's true. Chicken or the egg, I guess. I got you you know, choose one, get the architecture correct, get it working, and then see how the AI is working. You might need to tune it then, right? That may be but but the thing here is like you can't it's hard to evaluate the tool if what it what it is meant to affect is, you know, kind of Wishy washy and uncertain, you know what I mean? You have to be very literal and very detailed ⁓ with the AI and telling it exactly what you want it to do, giving it exactly the information that it needs ⁓ to process these days. It it's the newer models are becoming very good at this. They're very, very good at it it they used to be where you would give it a role and it would it would pretend to be like a a ⁓ a marking marketing strategist or it would pretend to be like a content strategist or something like that. But the nowadays it's you tell it exactly what you want it to do and you're giving it exactly the tools or the the the context that it needs in order to complete that. And then it's able to execute much better. But you can't know that you know until you have this architecture figured out. ⁓ So You gotta map those three handoffs. You gotta identify where the the business owner judgment is embedded and and necessary and required, or at least thinking that it is. And then start to design AI into the that decision layer of those handoffs. Not the task layer necessarily. That's easy enough to take care of. We're talking about the decision layer here, which is like One la one layer down, right?
Simone Henry: So here are your steps. First, you build the qualification criteria, right? Because we're talking about not just tasks, but using these AI tools to help you make decisions, to take you out of out of all of the of the decision making, so it's not dependent on you, right? So build those qualifications. What's the qualification criteria that you need for your automation, right? Most founders are skipping this. They go straight to, well, what AI tool handles lead scoring? The tool is irrelevant until you have a documented actual decision criteria. ⁓ And so that you know, well, how does what makes a lead fit? What disqualifies them? What qualifies them? What's the urg where's that signals urgency, right? And write that down. Make that a, make that your your step-by-step decision, what do you call it, matrix, right? So then you template your repeating handoffs. So map that the last 10 instances of each handoff and pull out the common steps. Build the autumn, build that template, and then layer the AI onto the template to handle that standard path. This is it's not a technology project, it's really a documentation project that technology then accelerates. And third, You want to define the quality thresholds at every delivery milestone. So you want to make it measurable, right? It shouldn't be like, ⁓ it's good, or this is ⁓ this is not so good, right? Make it specific so that your team actually can evaluate it without you. And then the AI actually becomes that. reviewer against those criteria and you become the exception handler, not quality the quality control department.
Nathan Grossman: Yeah, exactly. So what here's what it does not change. ⁓ a few things ⁓ So ⁓ when when business owners are thinking about this and they hear this, they they start to, you know, ⁓ my gosh, we gotta we gotta do this, right? Your relationships with key clients are not changing. The your presence, the the the business owner's presence, the founder's presence in high-value client relationships is not necessarily a bottleneck. It is an asset. Don't automate that away. Never discount ⁓ that crucial, you know. that crucial something that i you know, the relationship factor there. AI is never gonna be able to account for that. No matter maybe well, maybe eventually I but I I don't know, like maybe robots, but anyway, that's that's a long ways away. ⁓ you're not gonna be able to automate that. Having strategic judgment also does not get delegated to AI. ⁓ A AI handles that repeatable decisions at at the handoff level. If something is quantifiable and is able to be measured, there are certain criteria around it, you know, this, this and this has to be in place in order for this to happen. That's something that AI can handle, no problem. But being strategic, you know, that strategic judgment level, no AI cannot replace humans on that. That requires way too much brain power and insight that AI simply doesn't have. the founder still needs to set the strategy, the overall strategy for the business, right? And they are the ones who are defining the criteria because they are the ones who have the deepest appreciation and understanding for the business and all the things that go on in it. And then of course the business owner or the founder is still handling those ⁓ genuinely novel situations, though the out the outliers. That's what we want to get to, where it's the outliers that they they're able that they come in and and fix or handle as needed. But in the rest of the time, they're able to step back and be strategic, think strategically about the business, spend a lot of time thinking, right? There's so many, so much there's there was a ⁓ a Harvard Business Review article ⁓ not too long ago, relatively recently, within the last couple of years, that was published, and they did a survey of ⁓ top CEOs. And the the outcome was basically like I think the question was what do you what is the most important thing for you to be doing in the business? The number one answer that came back was think strategically. Okay. But do you know the number one reason why they're not doing it?
Simone Henry: Original stuff. ⁓
Nathan Grossman: Exactly. They don't have enough time. That's what they said. They don't have enough time because they're doing all this other stuff. So that's what we're talking about, right? We're gonna offload that all that busy work, all those ⁓ sort of like, you know, intermediary decision thinking layer stuff that could be offloaded because it's you know criteria process detail oriented onto the AI, which ⁓ will free up brain space for you. All right. Yeah, so taking yourself out of that routine stuff, but not the strat not the s the strategic aspects. And your team's need ⁓ for clear direction does not go away. They still need you, they still need to know what it is that they are supposed to do, and that needs to be clear. Having AI in in the system does not replace management s as of yet. I know that there are there were businesses that found this out the hard way, you know, within the last year or two, where they they fired their entire management team layer and thought that they could just implement AI. And I think almost all of them have decided to walk that back. you know, sometimes a disastrous effect, but anyways. the AI is replacing the founder as the decision engine in processes that should not require the founder involvement in the first place. That's the key thing. So we're not replacing management yeah.
Simone Henry: Let the let the AI take over ⁓ decisions that are repetitive and that are very well defined, but not those decisions that are that are more abstract that really take human thinking to do. I feel like a lot of these companies are just jumping the gun. I've I've heard of other companies that have fired like their whole development teams and said, okay, well, we don't need you all anymore. We can just let the AI build out everything. ⁓ and then ended up having to hire most of those most of that development team back because even though AI is much faster at doing development work.
Nathan Grossman: Yeah. Yeah. Yeah.
Simone Henry: There are still mistakes and it and it still needs human oversight. So
Nathan Grossman: Yeah. There was I just saw recently where the original promise of AI actually saving companies money as well, especially at the enterprise level, is sort of fast becoming a pipe dream because of the because of the expense of having to implement all these different layers, right? And so ⁓ this is this is a consideration when we're talking about this that needs you need to take this into account. basically you're sort of you're sort of hiring a mini me, if you will, who's able to kind of ⁓ you're able to offload a certain amount of what you would do as the founder. onto the AI. So you have to take into is it worth the the cost in that case? In my opinion, I think it is because of the value of strategic thinking and being able to take time, step back, think about things, and and kind of steer the ship. You know, I I think that that is very valuable in and of itself. So just my two cents in that regard. But There are businesses who are implementing this and finding that the the compute cost, the the token cost, and that kind of thing is prohibitively expensive. And we're talking like enterprise level. Like we're talking like where they've replaced whole departments with AI, right? Like that becomes excessive. And they're the cost trade-off, ⁓ are they really saving money? You know.
Simone Henry: There's a lot of ⁓ there's a lot of electricity and water usage, you know, to run these.
Nathan Grossman: Yeah, there's all that. Good all that too, yeah, no doubt. No doubt. so anyways, here's here's what let's get into like maybe the compounding effects here once we put this into once we put this into action. Here here's what you may not expect as a business founder. When you're removing yourself from the that decision layer in those three different handoffs, at least two things compound. First, the pipeline velocity increases because the deals are not waiting on you to review, approve, or customize at each of the different stages. Pipeline velocity nowadays nowadays is very important. Right? Speed, speed delete, and then of course you want to move them along that pipeline as fast as possible so they become paid customers. And The system is moving at the speed of the s excuse me. The system moves at the speed of the system and not the speed of the calendar, in other words. ⁓ Second, and maybe this is the one that is most surprising, your strategic clarity improves. When you're not buried in the routine handoff decisions all day, you actually have the bandwidth to see the patterns in your business that matter. Which clients are most profitable, which services have the best margins, where the real growth opportunity is. That clarity is a viability unlocked that most founders never get to because they are too busy being the bottleneck.
Simone Henry: That's that difference between being in the business and working on the business. Right? If your business is able to run you know most of its operations without you, you can spend more time being that strategic thinker rather than j you know the operations person. I told
Nathan Grossman: Yeah.
Simone Henry: I told a client recently that she's a doctor. She's a doctor. And ⁓ she was asking me if I thought that she should ⁓ go ahead and pay more to allow She was ⁓ I think selling her she's selling her products through through Walmart. ⁓ if she should pay more to to just let Walmart do the do the packaging when somebody orders orders her product. I said, Absolutely. You're a doctor.
Nathan Grossman: Mm. Yeah.
Simone Henry: You're you're not being paid to pack boxes. That's a that's a what? A ten, fifteen dollar an hour job.
Nathan Grossman: Yeah. Yeah, yeah, yeah. Right.
Simone Henry: doctor that is not your job to be doing that. Just pay somebody, let them do that. And you do you actually help diagnose and actually help people with their health. Your job is not to be packing boxes.
Nathan Grossman: Yeah. ⁓ to put that into the terms of a business owner and the founder, right? Your job is to diagnose the health the health of your own business. You know, you you shouldn't be out there packing boxes either. so yeah. All right. So three here's three things that ⁓ you can do this week ⁓ to implement this. Put this into put this into good use. First off
Simone Henry: Yeah.
Nathan Grossman: You're going to map your three handoffs. Take 30 minutes and just write down exactly what happens when a lead becomes a pipeline opportunity, a prospect, if you will, whatever word you're using there. When a closed deal becomes an active client, and when a new client transitions from onboarding into ongoing delivery. So you're just going to document what happens now, including every point where you're personally making a decision or give an approval. Don't fudge it either, be honest. You know. ⁓
Simone Henry: Right. and yeah, so you can just do this you can do this with ⁓ a shared doc like a Google Doc. ⁓ create put your three columns in there with your decision points, right? Or your handoffs, lead the pipeline, pipeline to onboarding, onboarding to delivery, and list every step in each of those columns. Put a star next to every step where the founder's input is currently required. And that started list is where your AI implementation comes in. That's the AI implementation roadmap, right? Not the tools necessarily, because but this also this will help you to determine which tools you need. Which tools are best for for these particular ⁓ for these particular decision points, right? So now you have your list.
Nathan Grossman: Mm-hmm. Perfect. And next, the second move is to ⁓ we're going to document your decision criteria. So you're going to pick which of those handoffs has the most requires your input. And then you're going to write down the decision criteria for that handoff. What is your thought process precisely? Write it down. excruciating detail, it can't really be too detailed in this case, I think. This is going to be an important proprietary document. You probably don't want to let you know an and just anybody see this by the way. So keep that in mind. But if if it's lead qualification, you're gonna write the 10 criteria that you actually use to decide whether a lead is worth pursuing. If it's onboarding You're gonna write the standard steps that repeat for each for every client. Now you're gonna have to think this through, so give yourself a good chunk of time to be able to do this. You're gonna get it out of your head and onto the paper. Okay? If you have to, there's a number of different things out tools out there that you can use to help you with this, which you know involve AI. But ⁓ I, you know, I have my on my Google phone, I have ⁓ Google Record app. Which you can you can speak into that and it automatically transcribes what you're saying right then and you can just copy paste that wherever you want. there's other AI apps that will do that for you as well, but you know, if you don't wanna type this out, in other words, there's ri whisper sync people will use on their computers, so they're talking into their mic and it's just typing it out on paper. I think I think Windows, excuse me, ⁓ Microsoft Docs doc they have ⁓ When ⁓ w Microsoft ⁓ yeah, Microsoft Docs. They have ⁓ I I keep getting Google Docs and Microsoft Docs confused. ⁓ Word, Microsoft Word, good grief. where they have a feature built into it that will transcribe what you're speaking as well. So, you know, there's tons of stuff out there. Just get it out of your head onto paper. Yeah.
Simone Henry: Yeah. True. ⁓ and a screen recorder will will do it the same. so right, but right exactly, there's no excuse, and also you also you don't wanna don't try to make it perfect, just get it out of your head, take a few minutes and get it down, and then run three real examples through it to make sure that you have it, that you have all of
Nathan Grossman: Yep. No excuse, in other words. So Ha ha ha.
Simone Henry: criteria down. ⁓ see where it breaks, make sure, and then and refine it as you go, right? It's a working draft. You want to get a working draft down, not necessarily a finished system. So we're we're not in a ⁓ you know, let's let's think about this for for ages and ages and ages, right? Get away from analysis paralysis. ⁓
Nathan Grossman: Yeah.
Simone Henry: It's it's not about being perfect. Just get it down.
Nathan Grossman: Yep. Okay, the third move, you're gonna audit your your AI stack. You're gonna audit your current tools ⁓ against these three handoffs. What what are you currently using for AI? For every one of those AI tools that you're currently paying for or using, you're going to ask one question. Is this tool operating in the decision layer of a revenue critical handoff, or is it operating in the on the task layer around the edges? So if every tool is on the task layer, you know why AI has not changed your ceiling, in other words, not really moving the needle.
Simone Henry: You know, be honest with yourself. AI writes my proposals. That's a task layer. AI qualifies my inbound leads against documented criteria and routes them without me, that's a decision layer. So that distinction matters. Yeah, having AI write your proposals is great, but you you're still needed to to to decide, okay, well who is getting a proposal and who is not. Right? So a lot of founders, they have this full stack of of tech task layer tools, but nothing in that decision layer. And therefore your ceiling, your growth ceiling stays in place. Nothing really changes in your business.
Nathan Grossman: Yep. And that's the basically the bottom line. AI is not going to change your growth ceiling by making you faster at the things you're already doing. It's gonna it's ch it's gonna change when it's gonna change that ceiling when it removes you from the decision layer and the handoffs where your business currently depends on your judgment to move forward. And that is a systems architecture decision. It's not about tool adoption. And it maps directly to the viable pillar of the V3 framework. The mechanics of converting, transitioning, and managing your pipeline without the founder in the middle of everything. Remember Visibility gets you leads, viability gets you out of the weeds, and value sets you free. If this episode made you suspect your real constraint is not what you've been treating it as, take the growth ceiling assessment at ⁓ we'll include the ⁓ the link in the in the show notes there. It's gonna take about five minutes, ⁓ if that, and it's gonna show you where your actual constraint lives in your growth engine. And of course, if you want to talk it through with us directly, book a growth clarity call, forty-five minutes, no pitch. We're gonna map out your specific situation against the V3 framework where you know, and you're gonna leave with a clear view of where to focus next. ⁓
Simone Henry: That call. So ⁓ if you enjoyed what you heard today and you want to hear more, subscribe to the Growth Ceiling Podcast wherever you listen or watch if you're on YouTube. And you know, if we talked about something, if you learned something today, put it in the comments section. ⁓ tell us what your biggest takeaway was. And also send this episode to a founder who needs to hear it. You know somebody who's a bit overloaded, ⁓ who is who never has time to hang out with you as as friends because their business is just sucking the life out of them, they need to hear these episodes. So share this with them. Thanks for for being here. We'll talk to you next time.
Nathan Grossman: Yeah. Bye bye.
Podbean