Before Signing That SaaS Renewal

0:00 Ankit

It definitely caught my pocketbook, right? Very much like affected us. It's very real. We had to deliver something that was more market capable. We had to reduce our pricing by thirty percent just to be competitive. What are we what are we gonna do to do that? We gotta cut overhead. Hey, I have this HubSpot subscription we pay around eighteen, nineteen hundred bucks a month. Not a huge amount for most people, but enough for us, whereas like made us think of like, Okay, well, what are we really getting with this?

That signal made me rethink everything that we were doing.

0:24 Sam Rogers

Welcome to Signals and Subtractions. I'm your host, Sam Rogers. So when the software renewal comes up, how do you know whether you should re-sign it or renegotiate it or rebuild that replacement yourself?

0:51 Sam Rogers

Just a year ago, that third option wasn't really on the table for most serious businesses. But here in Q3 of 2026, it isn't a joke anymore. And the people that are finding that out aren't necessarily the ones with the biggest AI budgets. They're the ones with the invoice and the question. So my guest tonight is Ankit Patel. He's an industrial engineer who married an optometrist and became the operations department. He and his wife run Classic Vision Care in Atlanta.

1:20 Sam Rogers

As well as My Business Care Team, which is the back office for other people's optometry practices. He learned lean methodology at Dell and at the Cleveland Clinic, and now he's running about a billion tokens of AI inference every day, which he'll tell you himself is probably too many. But here's what makes him the right guest for this show. Almost everyone talking about AI right now is describing the decisions that somebody else has to live with.

1:49 Sam Rogers

Ankit signs the invoice and then works in the same building as the result and manages offshore operations. We first met in Nate B. Jones's executive circle, which is an insight-rich community you'll find on Substack. Thanks so much for being here.

2:05 Ankit

And thanks for thanks reaching out and I'm glad to be here. So excited. It should be fun. Yeah.

2:10 Sam Rogers

Really glad to have you. And let's get to it. So Ankit, what is the signal that you've been watching?

2:17 Ankit

Thank you for the intro, Sam. We helped with back office support. So traditionally that's meant supplementing with global resourcing, so offshoring or technology and helping make things more operationally efficient. And so that's traditionally what our operations started with a few years back. Now we're getting clients that started to tell us, you know, I I love your service. You know, you have a nice, good, complete package, but I can get 30% of it now with my other software packages.

They're offering these things. Yeah.

2:46 Ankit

Let's leave aside that whether it's a good idea or not and how well they do it. But the their impression was like, This is good enough. And you know what? My team can take the rest of the seventy percent on. So what they're saying is like I'm gonna just gonna replace not a hundred percent, but thirty percent of what I need, and that should be good enough. And so that was an interesting signal to me. And and it was interesting because I was thinking to myself, like, what could I be doing?

That's also similar to that. Yeah. Using at thirty percent of the level that I could be doing. So yeah, that that that was that it caught my attention. It definitely caught my pocketbook, right?

3:16 Ankit

So it was like, hey, very much like affected us. It's very real. It's like, hey, we live in entrepreneur journey.

3:22 Sam Rogers

Exactly. And that's why I was so excited to have you on because like getting shot in the wallet, you tend to notice that in like a a more visceral way. Anyway, it's much easier than thinking of like some big enterprise that like you've worked in before. And the immediacy of clients finding a solution that they're okay with and then being able to take inspiration from that, shall we say, to dive in. And if you can't beat them join them, it sounds like a little bit.

Is that accurate?

3:49 Ankit

A little bit. I think it was more of a fact that hey we had we had to adapt. Right now we have about like sixty employees. We were a little bit bigger this point last year. And because of losing folks and also we had to deliver something that was more market capable. You can think of it as we had to reduce our pricing by thirty percent just to be competitive. That was roughly what it worked out to be. So it's like, hey, what are we what are we gonna do to do that?

We gotta cut overhead, we gotta cut unnecessary expenses. And so, you know, the usual things come up is like, hey, I have this subscription to HubSpot, you know, we pay what what was it about eighteen, nineteen hundred bucks a month.

4:19 Ankit

Not a huge amount for most people, but it was, you know, enough for us. Whereas like made us think of like, okay, well, what are we really getting with this subscription? And so that that was kind of interesting how that signal made me rethink everything that we were doing because we had to get operationally more fit. And we're in a little bit of a unique position too, because then I also decided to say, okay, we have our brick and mortar stores as well that we also own, right?

The actual optometry locations where you see patients.

4:43 Ankit

And that's a little bit different as well. And so that became a a situation of well, how are consumers using it? We shifted our strategy to to try to target more they call it AEO, I guess, or GEO now, which is you know, search engine optimization. So so looking at it from that perspective, if this is on people's mind, how can we get to where they are with that? And so we did some pivots in our business where we're offering more automation type products now.

We don't call them automation or AI. We just say we can do it for this price and this is what you get. Here's the results, not the how, and this is the price point.

5:11 Ankit

So for instance, social media, website, all those verticals doing that. That's a pretty standard offering. We're in the process of building out operational efficiencies in our office because of that as well. Because we realize like look, people just want the end result. They don't want necessarily the, you know, the people or whatever. They don't really care as long as the end result's there most of the time. What we're doing now is building out some things that make it easier for a person in the optometry office.

So for those of you who don't know, one thing that's a really big pain point is

5:39 Ankit

if you have vision insurance, it's different than medical insurance. Vision insurance, there's a couple of main ones, but ultimately they're kind of line itemized out. And there's like four or five tiers of each level of product. So your pair of glasses that you're wearing right now, Sam, it might have five or six different options in there that they each have like five to six different levels. And so that starts to multiply out quite a bit.

And then when you have different prescriptions, that adds even more complexity in terms of what you need to pick and what not to pick. And so you get

6:07 Ankit

hundreds of thousands of possible combinations with one prescription in one person. Navigating that, pricing it out, and making sure it's the right price is is always a challenge. And so we're doing things like streamlining it to where, hey, you don't really have to think, the program does it for you, right? Not necessarily AI, because that's not an AI program, but AI helps us write the program that helps us automate it.

And so that that's been some really good effects for us in in our offices, along with, you know, trying to figure out, hey, how do we get more efficient?

6:32 Sam Rogers

Using AI to write the program, I think that I just to pause there. That's a really important distinction than like using AI to run your business. Could you say a little bit about that and how you knew like when to stop? How deep down the the data flow does that AI go? Maybe some of the concerns that you have, you know, specific to your industry. So

6:56 Ankit

I'll touch on a few different points there. One, I guess one of the perks of burning a billion tokens a day is you learn a lot, make a lot of mistakes. I kind of learn what AI is good at, what it's not good at. Basically, it's non deterministic, right? My first perspective is is this a deterministic type task or not? Can I map it to make it deterministic? If I can do that, it's always code. It's always a script, some sort of code that can do that.

But there are points where we do insert AI. So for instance, there's new insurances that come out all the time. The AI can take that, learn it.

7:26 Ankit

And say, okay, now we need to create a new script around how to read this and how to map it to our existing product set so that we can price it out properly. So now when you come in, you may have a new insurance. You don't have to go and reprogram the machine. The machine learns the new insurance. And then what it'll do is it'll go ahead and say, Well, the doctor prescribed you know a pair of sunglasses, everyday pair.

Okay, great. This is how it maps to your insurance. This is how much you save.

7:48 Ankit

This is great. Does this work out for you? Great. Let's get you going. Instead of having to sit there for twenty minutes and I you know what? When I used to do it, I'm decent at math, I'll I'll say, because of my engineering background. It still took me a while to to do this. So it probably saves like five, ten minutes on every transaction easily. But more importantly, it prevents the mistakes that happen, which can kill a business, especially a small business.

Well

8:08 Sam Rogers

That makes a lot of sense. And thank you for sharing all of that signal. I'll just share a little bit of mine as well. I went a little too far. That's part of the reason that I was asking the question. This is very fresh for me as of what I was doing over Labor Day weekend. I've got a number of open GitHub repos, or places that you can go, like see open source code or places that I'm hosting various websites and services.

And there are six of them that have scheduled jobs.

8:38 Sam Rogers

That check on like a bi-weekly basis, some of them a nightly basis. They all have a very similar kind of model cascade where I'll have it first check Perplexity for some news and then validate that with X and then go to Claude. They're all doing different kind of levels of research, checking each other's work so that I can keep these things current. Basically, I'm the one that comes in and, you know, resolves questions as opposed to

9:08 Sam Rogers

doing all the research myself. And so for an example, everyailaw.com is one of them. So every AI law in the world is basically broken down to machine readable artifacts that I make open and available freely to everyone, like legal products, stuff like that, that really has to be right. And I didn't realize until Friday night that some of my checks weren't checking effectively.

9:35 Sam Rogers

So basically certain kinds of failures about these research tasks would simply come back empty. And empty isn't an argument. So as long as one of this cascade worked, I would get the signal that the research had been done, until I saw this one thing that was like, that that isn't right, on a Friday night, right? And I spent the next three days basically trying to figure out why workflows were coming in green, everything's fine, but the results were not.

10:04 Sam Rogers

So I ended up checking a lot of stuff by hand and going through like legislative records back to the source and figuring out how to better design that system. And I can tell when anything comes back empty at any stage of the process and like get a a finer grain kind of dashboard of how to see this. So that was a a particular pain point that I was not expecting for my three day weekend.

10:27 Ankit

So that happened to me a couple of months ago. It was something very similar. I was like, hey, you know, you have health checks. You said everything was green. What what the heck, man? Yeah. But then basically what it turned out was like, yeah, it did return something. It returned nothing. And that counts because it did something. And so that that tends to be where I'm that's something actually I'm struggling with. Well, not struggling.

I think we have figured out a solution if you want to geek out about that. Nate B. Jones, our mutual friend, he has a great open source repo with Ringer.

10:56 Ankit

The the shape of Ringer, I think you've actually built on it a little bit with your your harness too. I have. But the one thing that I really found effective was it being inspectable work by the AI. And so this gets into a little bit of a nerdy context of how what we have to watch out for. There has to be a predefined criteria of success of what that it looks like. And ideally there's a code checker. So if you can structure in a way that code can check it, that's ideal.

But if an AI has to do it, you know, there's clear

11:22 Ankit

guidelines and you use a separate AI to check it out, which is probably what you're you're doing. But it was it was complicated because what I ran into there was what's the minimum amount of work that needs to be checked?

11:32 Sam Rogers

Yeah. Great question.

11:34 Ankit

Yeah, and and so a good example, so sales outreach. I was like, Okay, go go do the sales outreach, and this is what I was having problems with. I was like, Why can't you do this properly? Then I came down to it, it's like, you're not generating the context properly. So the first deliverable is not who to contact. The first deliverable is the context document for each one, and there has to be a certain way to ground that document.

So, hey, you have twenty different sources, you need to validate that each one of those sources were taken into consideration and trace those.

12:01 Ankit

And then you have your document. Then you can go on to the next step where you actually write the email. And then you have to use our writing rules. And again, you have to ground the data and there's a third checker. So it was like three steps before it got to out. And I was trying to do all that in one step. And so I don't know if those are things that you've been experiencing in terms of like, hey, how do you get to the geeky and that my background being in lean, I was like, okay, this makes it easier to break down the process.

Because in lean they talk about a transferable work element. It's the smallest unit of work.

12:25 Ankit

Anything that's truly transferable is what you use. And I said, okay, well that's kind of what I'm coming to here with designing AI systems. It's gotta be that level of detail and that level of of small deliverable. It tends to be more effective.

12:36 Sam Rogers

With the example of the AI laws, these are global, right? So there is actually an AI law in Kazakhstan that I cannot read, you know? So, like, how is it that I'm breaking this down? Like legalese in English is enough of a challenge, right? So being able to like do those translations other ways to be able to recheck that work. And a lot of that actually opened up new possibilities of taking the same kind of thing that I do with

13:05 Sam Rogers

blog posts, which I translate locally using Gemma 4, and like taking that same architecture to run on a nightly basis to just do all that translation. That now I don't actually need the frontier models to check in the same way. I just need a fetch, which is a, you know, find where it is, but the deterministic part is bring it down, translate it, the local LLM, then give it back up like the next night, it's fine, you know, figuring out how to do that

13:35 Sam Rogers

comes exactly back to like process mapping. And okay, well what if I just tease these things apart and then start to make it like line up better.

13:44 Ankit

And you know, I'm gonna combine a few different themes that I've seen in my own operating. So managing like sixty people, we gave everyone AI and then we took it all away because it wasn't working. Yeah. And I dove deep into learning theory, which you're way more versed than I am. The biggest thing I took away was like chunking information. How do you chunk it and how do you visualize connections and how do you interpret relationships with things?

And so AI, I noticed isn't great at that, but humans are or can be.

13:48 Sam Rogers

Please?

14:11 Ankit

And if a human being is good at that, they're probably going to be good with AI. And so what I'm finding is that if you can take a concept like how to write a blog post, turn it and chunk it into concepts, and then turn those chunks and concepts into steps is sort of the process. So I did something similar to blog writing about a year ago. It ends up being a 17 step process to create a blog post. And so it does great.

My stuff ranks pretty well. But that was, well, it was because of taking the the first principles and breaking down the concepts.

14:41 Ankit

Because anyone can just ask AI, write a blog, but you're right. What are the steps and how do you get from point A to point B to point C?

14:47 Sam Rogers

A year ago we were still at the stage of like trying to cram all that into one big long prompt. And what I'm hearing you describe and where I think I am myself is is more and more teasing it apart to like distinct steps as opposed to the here's the big long thing, now go do the magic and then come back and then I'll say yes, you know.

15:07 Ankit

Although I think the AI companies are trying to collapse it with Astra, the new model coming out. I think that's what they want to do, but I don't know if I like that necessarily, but we'll see. What that allows me to do though, as a you know, one of the things like I mentioned the original signal of like, hey, people are leaving us. And so how do we become more efficient? Well, it turns out that we have to be able to deliver service differently now.

Yeah. Be able to be more agile.

15:31 Ankit

more lean. How do we deliver more value to the customer without waste? We're starting to get to the point now where it's like, hey, we have to own that stack. We have to own every piece of that. Right. To really control every piece of that. The more we can control, the more leverage we have on the end outcome.

15:44 Sam Rogers

Sounds like you're headed straight into your subtractions. Segue.

15:49 Ankit

Sure. So yeah, exactly. So you know, our communication with our patients and our well, not our patients, our clients. Patients is a different animal. We can talk about HIPAA compliance and patient information. I don't use AI very, very, very little, if any, on patient information. And we can talk about that if you want to. But subtraction made us think like, Look, we're spending all this money on HubSpot. Do we really need it?

And how can we be more effective at communicating and what is the end result we're looking for here? The end result is staying on top of communication, and we also had a ticket management system.

16:18 Ankit

But I was like, there's gotta be a better way and easier way to manage the tickets. And so I had a conversation with HubSpot and I said, look, I don't want to sign a year-long contract. I'll pay you X amount of money a month. You know, I'll have to keep this, but like, no, you can't do that. You gotta go your whole year. I'm like, I don't wanna do that. And so I just started asking AI, is like, what would it look like to open source this platform of everything I do?

And so, you know, there's an API for HubSpot. You can go see what you're doing and how are you using it. And it said, yeah, these are the five things you should probably do and replace it with.

16:46 Ankit

And so I looked at those, researched those, vetted those, used AI. AI has access to my AWS, right? So it can automatically pop stuff up securely. So I said, Hey, okay, I like Twenty CRM. It's open source. Let's use that. Let's use Chatwoot. It's another open source. Here's a GitHub repo. Put it on there. And okay, now put our company SSO on the front of it for login. So I don't have to manage passwords. It actually had the whole thing built for me.

And I was like, Wow, that was way too fast. I was like, is it working? Is it broken?

17:16 Ankit

I planned the transition, but like within within two weeks I was able to transition off of HubSpot.

17:22 Sam Rogers

That's fantastic. And and in fairness to HubSpot and you know, all the SaaS businesses out there who may be panicking with that kind of thing, it's not like you were a heavy user of that service, right? Like you're basically just wanting to manage leads and marketing and email, not like the full functionality of HubSpot.

17:42 Ankit

Right. No, because I have familiarity enough with AI that I didn't have to use their automations, I didn't have to use their marketing, the websites. So I wasn't on that platform to begin with. I also will say that it was really more of a source of truth for us. I'm like, well, if that's all we're using it for, do we really need to spend all this? All these other functions that we're doing, we can deliver them better.

And we do now. So our help desk and our help bots, it's way more automated. It gets people

18:07 Ankit

fast help, people are responding faster. It's actually gotten better since we've replaced it. But again, nothing against HubSpot. It's a great system. It's just a Swiss Army knife when you when you need like a scalpel, right? And so it's just

18:17 Sam Rogers

different. That's great. And I didn't realize that you had made that cut over so quickly too. Was there anything that broke or did it get it right on the first try?

18:26 Ankit

Well, I mean, look, there was there was a few things I was like, did you do this? Did you do this? So the first two weeks was transition. I was planning a month of transition. But after two weeks, I was like, everything already passed all the milestones and markers and everything's working. So I was like, Okay. But like it worked pretty well. It's pretty robust. Then I started thinking in terms of my lean hat of adding more value.

So instead of thinking of things I could subtract, one thing I thought of is what waste can we subtract out of the system? So we have a shared email account, right? With some of our clients. It's like, okay, well,

18:54 Ankit

instead of having to log in, get 2FA, waste all that time, we can set up on Chatwoot. You know, I like we are a virtual company. I've seen a lot of platforms that have like virtual spaces where you can, you know, all be in a space and you have these little avatars. There's an open source version like that. We just, I just again told the AI, go do it. And a couple of hours later it's there and my team is playing with it and using it.

Right. And so when the infrastructure is all there, so now it's been fun playing with customizing open source for our needs and what we need to do. How do we deliver better experience for our

19:24 Ankit

our staff and our teammates and our clients and our end customers, the patients too.

19:30 Sam Rogers

Yeah. Well it sounds like you've been taking this approach once it worked in one little spot and kind of expanding it throughout your business. If you don't mind sharing, like how much savings has that resulted in for your business?

19:43 Ankit

We're so far at twenty three or twenty five hundred bucks a month. Again, not huge, but from our budget standpoint, you know, that's a significant reduction. It's about ten percent reduction in our SaaS budget. Wow. I know people talk about the SaaSpocalypse. I think that is a signal for a lot of, well, for non power users, I would say. For my use case, I think that that is an option that people will start leveraging more.

Yeah.

20:07 Sam Rogers

We've talked about that on this show before. Now, this is coming from someone who did learning management system migrations as like my bread and butter for 20 years. Large systems managing learning HR tech stacks, that whole thing for like ADP or you know, standing up systems at YouTube or, you know, like major stuff. My general rule of thumb is at this point in history, if you have to spend a a ton of time creating a whole migrations project,

20:34 Sam Rogers

you should probably look at some of the other options that exist for you, which it's exactly why I wanted to have you on, Ankit. This is the kind of thing that it does make sense for enterprise as well, but to be able to speak about the small medium businesses like yourself, it makes it so clear to see. Like it really didn't take a whole migration project. Like you were able to actually test and deliver in the span of time that you'd usually spend just scoping, you know?

21:01 Ankit

And I'm more of a build the parachute on the way down type of guy. So I was willing to take the risk if something broke. It's not everyone is. So I understand that. The other thing I'll also say that it's kind of, I didn't say it explicitly, but all like the little building blocks to be able to do this were built out already. So my Google CLI, AWS CLI, the right keys and tokens securely, properly done, hosting sites, all these things were already built out as building blocks.

So I could take a Hermes, Codex, Claude Code and say,

21:31 Ankit

this is what I want to do, go do it. And I can jump surfaces because it's all portable. Yeah. So I think that's an important distinction because the speed didn't come until all the infrastructure was built.

21:39 Sam Rogers

That's a really important point. Yep. You can't go from zero to sixty on a bumpy road. You need the track nice and smooth first. Yes. Yeah. Cool. Well, I'll add my own subtraction here, which is an automation that I just killed. So I have an instance of Open Brain, another one of Nate B. Jones's projects, and that I've contributed to. It's a great project to be able to manage context for both AI

21:48 Ankit

True, that's true.

22:08 Sam Rogers

and humans, like in the same spot. My own kind of derivation I call local brain, which is without keeping it all on the cloud, just keeping it on my own hardware and being able to use it with obsidian as an interface and keep track of all this stuff. And I've been using it I think since January. Like I was building this out before Open Brain was even a thing, and then I kind of adapted it to that. And it's been working pretty well.

But again, I built just a little too far

22:36 Sam Rogers

with it. So I started to create a decision system inside that had a bunch of defaults where unless I took an action to prevent something happening, decisions were basically like lit fuses. And I was giving myself the motive to not just keep track of stuff, but also to let stuff go. And just to say like, if I haven't touched this in two weeks, it's probably okay that I don't touch it anymore. It can go away, which is fine.

23:05 Sam Rogers

Until there's a family emergency and you have to take a week away, and then the cascade starts firing when it wasn't really what I intended. And so I didn't have any safety mechanism or anything that I had built into it. So just yesterday, really, I finished defusing all of the bombs of little fuses that were lit of decisions that I'd kind of pre made. I'm big on pre mortems and doing things to like

23:34 Sam Rogers

pre-decide what would happen if this happens or how things would fail before they fail. So I can take those as leading indicators rather than lagging indicators, all that kind of stuff. Which is a nice fantasy. But when it came to actually doing it and making it work in a real way, it reminded me of David Allen of GTD fame. This is a productivity system from 25 years ago that I was big on. He always said that

24:02 Sam Rogers

if it was possible to make a system that would tell you exactly what you needed to do at Tuesday at two o'clock, by the time Tuesday at two o'clock rolls around, you're gonna decide you wanna do something different. Doesn't matter what it is. And so being able to like be responsive to all of the different incoming inputs, be they, you know, family needs or business needs or whatever, being able to respond to that.

24:27 Sam Rogers

I just realized I needed to take a step back from that decisioning system to keep myself as the focal point there, rather than building it out too far into the system where decisions were starting to engage that I didn't approve, basically. I had pre approved, but the circumstances had changed. So yeah, that was my big subtraction. And that's something that I've been building like over the course of nine months. I'm basically back to like April or May levels at this point.

So

24:54 Ankit

I'm curious, what would you need to see or what would it take for you to build that trust back now that the trust has been broken? Because that's almost like gonna be harder to build that back now, right? Yeah.

25:02 Sam Rogers

Yeah. Well, I started building it back in two ways. I'm big on obsidian, but one of the things that I do is like it all has to go in one of two vaults. There's one vault that I keep that AI has access to, and there's a different vault that like architecturally on my computer AI cannot touch. And it's been that way from the beginning, in part because of like sensitive legal materials, but also because I wanna have autonomy over my own thoughts and my own records.

25:31 Sam Rogers

And if I give AI too much access, I'm just not quite sure. Like I could never be sure. So I've got these two vaults that I've maintained for like a year. I never want to make more than that. But in the effort of building trust, I've been making more than that. So like I partitioned out some certain projects that are their own vaults now. And I'm starting to just take this kind of decisioning approach and try it out like in micro again, rather than the whole entire everything.

26:00 Sam Rogers

That's what I was doing before. Like everything that AI can touch, it was touching. To answer your question about building the trust back, I think it'll be just like building trust with people. I'm always bringing it back with all the AI talk. It's not that different than people. Like, I'll trust you, Ankit, but I might not give you the passcode to my bank account just yet. Like you could hold my wallet for me, you know, while I'm doing something, but I probably wouldn't give you like all the passwords, you know?

Fair enough.

26:28 Sam Rogers

We build trust as humans in smaller ways, and that builds up to something big. And it's always stuff that's simple. We never really build trust from the top down. We always build trust from the bottom up. So that was a mistake that I had made and that I've since subtracted.

26:44 Ankit

Yeah, and I guess if it's from the top it'd be more of authority, right? Because it's coming from top down, it's more authority than it is trust, you know.

26:52 Sam Rogers

True. It's true. And that's why we all love authority so much. That's why no one has any problems with authority ever.

26:57 Ankit

No, yeah. That's super interesting. I'm curious to see how it keeps going. Yeah, keep us posted.

27:04 Sam Rogers

I'm sure I'll be sharing things week to week here. It's always great to have these moments to share with folks like yourself who are in the thick of it and making systems and making decisions and engineering workflows in a similar fashion. It's probably very predictable mistakes that I'm making. But you know, some things you just have to try and walk straight into yourself.

27:24 Ankit

Like jump off a cliff and build a parachute on the way down, right?

27:27 Sam Rogers

Anything else that you wanted to share before we wrap up, Ankit?

27:31 Ankit

I'm seeing folks come across similar challenges, or it feels like there's almost a zeitgeist of AI, and not necessarily from like what's going on, but like what people are experiencing through their journey. Folks in our group are about three to six months, sometimes a year ahead of folks. Some things I'm excited to see is as these models get a little bit better, the harnesses get much, much better. I think it's gonna be even more important to understand the pure concepts and understand how to break it down, because if AI already automatically goes to a certain median or a typical response range, that gives

28:00 Ankit

people more opportunity to find those niches, add different value points. And I'm super excited about that, right? You know, lean manufacturing, my background, it talks a lot about waste removal, but it's really ultimately about value creation. We're talking in our office, how do we create mass customized experiences? So like every individual gets its own experience. And how do you do that? And then you can start doing that now.

We're actually playing with that a little bit without making it uncanny valley or cheesy. That's the trick. I think we're gonna be figuring that out though pretty soon.

28:29 Sam Rogers

Well, that's awesome. What lucky clients you have that get to be part of all of that. And we didn't talk about your business group either, or your podcast. Maybe you could say just a little bit about that and how folks get a hold of you.

28:41 Ankit

Yeah, so we have a few different things like yourself. So My Business Care Team, or mybcat.com, is our back office services company that we do a lot of the AI and leverage, but it's kind of, you know, married into the optometry office as well, Classic Vision Care. And with that, we do have a community for eye doctors. We call it iCare Grow. And we also have a podcast, Optometrists Building Empires, which we talk about the root causes of success for optometrists in the field.

29:06 Sam Rogers

You're doing a fantastic job with it. And likewise, I'd love to hear more about how it continues to develop and what you're able to create on behalf of your clients and patients and business. Thanks so much for being on the show.

29:19 Ankit

Fun, thanks for having me on. I looked at your previous guests, I'm like, Yeah, I would love to be on the show. That'd be amazing. High praise. So thanks for having me. Well

29:26 Sam Rogers

we've definitely got some good ones coming up. If you're listening on audio right now, please open up your podcast, search for Signals and Subtractions, and follow us there so that the next one just magically shows up next Friday. All right, on Sunday is issue 68 of the newsletter. Everything about this show is at sigsub.show. And thanks everyone for listening. One signal worth watching, one subtraction worth making.

Now go find yours.