How to Build an AI-Native Business That Actually Saves Time

  • Home
  • /
  • Blog
  • /
  • How to Build an AI-Native Business That Actually Saves Time

Catch the Full Episode

Overview

Most small businesses think they’re using AI the moment they open ChatGPT. That’s a tool habit. Building your marketing processes around AI is a business model, and the gap between the two shows up in your results long before anyone can explain why.

In this episode of the Duct Tape Marketing Podcast, John Jantsch talks with Jérémy Jarry, CMO of Bookipi, about what happened when he rebuilt his marketing team around AI over the past year. Jarry explains what AI native means in practice, why his team started with one person and one dreaded task, and how they learned to measure AI by hours saved and revenue earned.

This conversation is for small business owners, marketers, and agency leaders who are using AI tools but haven’t seen the payoff yet. You’ll hear how to pick your first process, what human in the loop looks like day to day, when to scale AI back, and how to keep up with constant change without chasing every new model.

Guest Bio

Jérémy Jarry is the Chief Marketing Officer of Bookipi, an international SaaS platform that millions of small businesses use for invoicing, proposals, contracts, websites, and getting paid. Over the past year, he rebuilt Bookipi’s marketing team around AI from the ground up and wrote about the experience for Entrepreneur. His work centers on growth, go-to-market execution, and applying AI across marketing so lean teams can compete with much larger companies.

Key Takeaways

  • Jarry defines AI native as redesigning a process from scratch with AI’s capabilities in mind. He describes it as a mindset change for the whole team.
  • Start with your calendar. Look for the work that eats the most time and the work your team dreads, then see where AI fits.
  • Prove it with one person first. Bookipi started by automating one marketer’s reporting, which freed that person to explain the numbers instead of gathering them.
  • The same thinking applies outside marketing. A plumber with an AI receptionist can keep fixing the leak while the next job gets booked.
  • In a Bookipi survey of more than 2,000 customers, the top reason for skipping AI was a knowledge gap. The team expected cost to top the list.
  • AI still needs your direction. A prompt like “write an email for John” goes nowhere until you define the topic and the outcome you want.
  • Give people time to adopt. Bookipi let slower adopters keep their old processes, and they came around after seeing teammates spend more time on creative work.
  • Human in the loop means real review steps. Bookipi feeds AI a central knowledge base for brand voice, and nothing launches until a person checks every asset.
  • Measure hours saved and revenue. Counting how many agents the team built each week told Bookipi very little.
  • Review time has to be shorter than build time. If checking AI’s work takes longer than doing it yourself, pull the process back and rethink it.

Great Moments (Timestamps)

  • [00:02] – John on the difference between a tool habit and a business model
  • [01:11] – The moment Bookipi decided to rethink its marketing processes
  • [02:04] – What AI native means, and why Jarry calls it a buzzword worth defining
  • [04:12] – Where small teams should start: your calendar
  • [05:08] – Automated reporting and the plumber with an AI receptionist
  • [07:23] – The 2,000-customer survey and the real barrier to AI adoption
  • [10:44] – Where adoption stalled and how the team handled it
  • [12:43] – What human in the loop looks like at Bookipi
  • [14:46] – Measuring agents built versus hours saved
  • [16:35] – Keeping up with the pace of AI without chasing every tool
  • [17:55] – How Bookipi started with a tradie’s invoicing app

Memorable Quotes

“Most small businesses think they’re using AI the moment they open ChatGPT at their desk. And that’s a tool habit, not a business model.” – John Jantsch

“Look at your calendar. Look at what is actually taking a lot of time. Look at what you don’t like doing and see if you can integrate AI within it.” – Jarry

“I don’t know many marketers that like to create reports.” – Jarry

“We need to make sure that the time that the human is actually going to spend to review the campaign is actually less than the time it would take them to actually create a campaign from scratch themselves.” – Jarry

Resources

 

John Jantsch (00:02.021)

So, you know, most small businesses think they're using AI the moment they open Chat GPT at their desk. And that's a tool habit, not a business model. The gap between the two shows up in the results long before anyone can explain why. Hello, and welcome to another episode of the Duct Tape Marketing Podcast. This is John Jantsch, and my guest today is Jeremy Jerry. He's a chief marketing officer at Book.

BookKeepy, a SaaS platform millions of small businesses use for invoicing and day-to-day admin. Over the past year, Jerry built or rebuilt his own marketing team at BookKeepy around AI from the ground up and wrote about what actually looked like for Entrepreneur Magazine. So, Jeremy, welcome to the show.

Jeremy Jarry (00:51.086)

Thank you, John. It's a pleasure to to be with you today.

John Jantsch (00:55.195)

So you as I mentioned in your intro, you spent the last year rewiring your own marketing team around AI. wha what was the moment maybe that that that you kind of took that stand that hey, maybe more tools wasn't the same thing as actually using AI?

Jeremy Jarry (01:11.83)

Yeah, so about a year and a half ago, we saw how AI was developing and all the possibilities that AI could provide to a marketing team. And like a lot of businesses, we started using Chat GPT Claude to proofread our emails, look at the different possibilities that we could use AI for. And we thought, what if we actually rethink

how our processes are actually working and can we rewire completely our marketing team to become more AI native?

John Jantsch (01:54.726)

So let's start with that term first off. I mean, what's the like what's the real difference between using AI tools and being AI native?

Jeremy Jarry (02:04.332)

Yes, so I think AI native is also a buzzword. So it is important for us to really define it. And internally we've had a lot of conversations about what it means and what we actually wanted from AI and how we could use AI to redefine the processes. So the way we see it is instead of just stacking AI tools.

John Jantsch (02:09.285)

Yeah, yeah.

Jeremy Jarry (02:32.226)

And thinking, we're using AI to its fullest. We thought we have some processes. What if today we were redesigning these processes from scratch and using all the capabilities that AI is providing to actually have an AI-native process? So it's becoming a big difference for us as being a mindset change more than just.

Adding AI to the workflows that we had.

John Jantsch (03:04.773)

So how do you get started on that? I'm I'm thinking that a lot of my listeners are four or five person shops. and so the idea of, hey, we barely have processes, certainly barely have them documented, and now you're telling me I have to tear all down. I mean, how did you get started with that? Because especially a business that's been around for a while, maybe that there's a lot of things that they've done for years and they seem to work.

But but you're right, they were never designed with AI in mind. They're they're sort of AI sort of being inserted into them. So so how do you get started on that I that that process itself of of tearing things down and and starting over again?

Jeremy Jarry (03:45.281)

Yeah, so we at Book Epi consider ourselves as a small business. we do have a small marketing team. and so it's definitely something that we're looking for also our customers, our small business customers, where it's not always easy to have all the capabilities within the organization to compete with large companies. And

John Jantsch (03:50.469)

Yeah, yeah.

Jeremy Jarry (04:12.994)

The way I would look at it for a small business is look at your calendar. Look at what is actually taking a lot of time. Look at what you don't like doing and see if you can integrate AI within it. For small teams like ours, the way we we saw it was it's about people and it's about making sure that people are actually embracing AI.

the way we do and the way we would want them to do. And so the way we did it, we started with one individual within the team, and we wanted to show results. And we wanted to show that it's actually giving more time to the marketer to be more creative, to be more strategic. I don't know many marketers that like to create reports, for example. And this is where we started by.

John Jantsch (05:04.699)

Right.

Jeremy Jarry (05:08.214)

Automating and then adding AI to have better reporting. It doesn't mean that the person is not looking at the report and it's just these are the numbers. No, now the person is explaining the report and has more time. Instead of just gathering the data, then the marketer is actually able to build on top of the data all the reasoning behind it. And we see this for small businesses as well.

for our customers. if you let's say you're a plumber, you're repairing a sink and your phone rings, you need to answer that second customer, but you're not. So you're just gonna keep your phone in your pocket, wait until you finish the job to call back the customer. What if you had an AI receptionist that can answer the phone for you and look at your calendar automatically and book your next job?

I don't know again a lot of plumbers who like to book jobs themselves, they prefer just repairing the leak that their customers have in their bathroom. And this is how we suggest our customers to use AI and to incorporate within the way they work and within their operations how they can actually get the fullest of AI for their business.

John Jantsch (06:34.211)

So how does somebody go about, especially a a firm that doesn't maybe have people that feel like they're very tech savvy or, you know, kind of on the cutting edge? somebody's got to redesign some of these processes. Somebody has to show, you know, how this AI is going to integrate in it. Is that should they is that a different person? Like do they need to go out and hire somebody new, hire some third party you know, resources or consultants to actually come and install it? And

So that's half the question. And the other half though is there's a whole lot of people out there selling, you know, run your business with AI. you know, and and so, you know, how does a small business owner, first off, get that outside expertise and how they know that they're not just being sold, you know, the latest, greatest get rich scheme.

Jeremy Jarry (07:23.19)

Yes, so I think that there are really different ways of approaching it and looking at how AI can actually help your business. And it's not necessarily you need to be AI native, you need to have AI agents for your business. It's really about if you don't have the knowledge, and we actually did a survey to

Over 2,000 of our customers and we asked them, Do you use AI in your business? And if not, why don't you use AI in your business? And we thought all the small businesses are are gonna tell us it's because of the cost, right? It there's a lot of cost pressure at the moment. And we thought this is gonna be the biggest answer on why they're not using AI. And it was actually not. It was actually a knowledge gap. And our customers told us it's

John Jantsch (08:06.587)

Mm-hmm.

John Jantsch (08:17.99)

Yeah.

Jeremy Jarry (08:20.202)

It's scary, it's intimidating to see all the changes that are currently happening with AI, and it's hard to keep up to date with it. And we provide some AI-powered tools. So we know a lot of small business owners want to change their website, for example. And so we give them a tool that is powered by AI to actually change their business. But we also understand that for small businesses.

John Jantsch (08:26.298)

Mm-hmm.

Jeremy Jarry (08:49.194)

And small business owners, it's not they don't really care if we use AI or not in that. They care about the results. They care about is it actually saving them time? And again, for a website, we know a lot of small businesses are looking at their website and thinking, I need to change it. I just don't have the time. But AI is actually helping them to change it faster. For your second part of the question, and

John Jantsch (08:57.628)

Mm-hmm.

Jeremy Jarry (09:18.634)

Also, a bit of the the first part. should they hire AI consultants? if it's what they need? Yes, sure. But I think there's a lot of tools out there that are very easy to get a grip on and that can introduce you to AI and where you can understand the different possibilities that AI can offer your business. And it's also very easy to

Write an email now with AI. And it doesn't mean that you shouldn't write what you should say. If I tell ChatGPT or Claude, write an email for John. An email about what? So we still need to tell AI this is what I want. This is what I want you to, the outcome I want. And then I get AI to actually help me to go faster or to

John Jantsch (10:00.891)

Mm-hmm.

John Jantsch (10:14.372)

Okay.

Jeremy Jarry (10:14.688)

write something, for example, that is without spelling mistakes, where you know you're on the go on your phone, it's easy to make spelling mistakes, use AI, and it's a lot easier to actually do it this way.

John Jantsch (10:28.294)

So when you were going through this process and reworking some of the things on your own, where where did it break down? I mean, where did you find people that just reverted back to the kind of the the habit, so to speak, of just using the tools, but not really changing anything?

Jeremy Jarry (10:44.95)

Yes, so we knew when we started that it would be a long process. So first when we started doing it, we gave ourselves and to the team a lot of time to really embrace AI, make changes, and we did it step by step. We also hired the team to help us to teach us AI, but also to help us rewire all our processes.

And we saw that some people were very quick and very happy about using AI. Well, then there were other people that would take more time, and we actually gave them the time and we've let them go back to their process, understand that maybe the results were lower than when they were using AI, or that they were doing things at a lower pace than when they were using AI.

And I think everyone by themselves understood, but if I do this, I can be like my teammate, I can actually spend more time doing things that I like. In the marketing team, we have very talented people who are very creative, but spending half or more than half of their times building a report, writing a slide, it's not really what they want and what they like. And they saw that.

Other people in the company were actually being super creative and they had a lot of time to do that. And so, bit by bits, I think every team member actually got into using more AI. And even today, we can see that we have some team members who are managing a fleet of agents while some other people are more using AI here and there to actually get better results or go faster, but not necessarily.

Building AI agents themselves or managing AI agents.

John Jantsch (12:43.899)

Yeah. So it's it's become pretty common, whether anybody knows what this means or not, it's become pretty common advice to keep a human in the loop. you know, that's actually like you know, you pretty much can't read anything that talks about using AI that doesn't have that as somewhere as part of the advice. What does that actually mean? you know, for for you guys.

Jeremy Jarry (13:09.804)

Yes, so the way we've designed AI and and our AI infrastructure is we have all the data and all the knowledge from the marketing team that is centralized. And so we're asking AI every time it has to complete a task that we're asking to go and ask this central base of knowledge to

make sure that it's not gonna be drifting from our brand guidelines, from our tone of voice. But it will. And there is a risk. And so we always need to have some steps where we mitigate that risk, but also where we check. I wouldn't let AI publish a campaign without the team members actually looking over and understanding.

John Jantsch (13:44.059)

Mm-hmm.

Jeremy Jarry (14:06.048)

making sure every asset is actually pixel perfect on what we launch. And it's the same in any department in Boukepi, where we wouldn't let AI do things that are risky for the business.

John Jantsch (14:22.917)

Did you have any speaking of that, I guess, did you have any processes that you rolled out, you revamped and then you had to kind of take back? Like once you started measuring it and measuring maybe how it impacted potential customer interaction? were there any that you decided, hey, we need to we need to lessen the role of AI in this and maybe re up the the role of the human?

Jeremy Jarry (14:46.306)

Yes, and for us using AI is about results. it's I think at the beginning it was very exciting and we started building a lot of agents. And we were measuring the results by the number of agents that we would build on a specific week. But we saw very rapidly that didn't really equate to the number of hours that were saved for the company. Or

John Jantsch (14:51.087)

Yep.

John Jantsch (14:56.707)

Right. Yeah.

Jeremy Jarry (15:14.326)

that we were seeing some increases in the revenue that we were generating. And so definitely we saw some processes that we had where we were going way too far with AI and we couldn't see actually that it was saving time. Building an agent and having, as we spoke about, a human in the loop. We need to make sure that the time that the human is actually gonna spend.

To review the campaign is actually less than the time it would take them to actually create a campaign from scratch themselves. So definitely we've had these issues and we fell into that. And for some of the processes, we had to tone back and rethink again the process for having humans actually building the whole thing.

John Jantsch (15:51.675)

For sure. Yeah.

John Jantsch (16:10.031)

You know, one of things that that I think is probably a little frustrating for some people is the change, the pace of change in AI. and so we build all these processes and then really the the way the tools work or you know what they integrate with now maybe eliminates the need for other tools, things like that. How do you is there a rhythm for kind of staying on top of the the fact that you're never done?

Jeremy Jarry (16:35.638)

Yes, it it is very daunting. and the pace is is definitely very high. the way we see it is do we always need to be on the complete cutting edge of all the technology? most of the times we do not. we actually do have some automations and some AI agents giving us, you know.

John Jantsch (16:36.677)

Yeah.

Jeremy Jarry (17:04.386)

These are the latest models, these are the latest tools that exist. But that that doesn't mean that we actually need them. And that doesn't mean that we would actually integrating them do a better job. So the way we see it is: are we satisfied with the process? Do we have another process that we actually need to change with AI? And for that new process, would it make sense to add a different tool or

removing one of the tools that we have. But you're right, it's very difficult to keep up the pace with all the different things.

John Jantsch (17:39.279)

Yeah, yeah. So so tell us a little bit, how does bookkeeping work with small business owners? Give us kind of the the overview of of basically what you do and how you provide value to small business owners.

Jeremy Jarry (17:55.044)

Yeah, so we started a few years ago with invoicing. our founder was a tradie himself, and he understood that it was difficult for him to just invoice his customers all the time, at soon as he would finish the job. So he built an invoicing app. And that was actually the beginning of the end of the process where you do have.

John Jantsch (18:03.301)

Yeah, yeah.

Jeremy Jarry (18:24.665)

To invoice your customers, but also you need to get new customers. So, how do I get new customers? I'm I'm a Tyler and it's not my strength to actually acquire new customers. So step by step, we actually built products around invoice. And so we now have a lot of different tools to help small businesses from

Building a website, sending a proposal, creating a contract, sending an NDA, following all the processes within your CRM, sending an invoice and then getting paid.

John Jantsch (19:07.703)

Awesome. Well, again, I appreciate you stopping by Duct Tape Marketing Podcast, Jeremy. Is there some place you'd invite people to find out more about bookkeeping and and connect with you?

Jeremy Jarry (19:19.053)

Yes, they can go on our website at pukeep.com. and they can follow me on Twitter or on LinkedIn and I would encourage them also to listen to the Ductate podcast.

John Jantsch (19:34.523)

I I second that. Jeremy again, I appreciate you stopping by and I know it's you're you're in a different time zone than me, so I pr hopefully hopefully I didn't keep you up too get you up too early or keep you up too late.

Jeremy Jarry (19:37.263)

Yeah.

Jeremy Jarry (19:48.118)

It was a pleasure being with you, John.


Tags


You may also like