If You Can't Teach It, You Don't Own It

Charlie Liu (00:00)

Hi everyone, welcome back to MCC Making Cents with Charlie. I'm really honored and excited to have Justin Merritt with me today. He's a longtime friend and a co-founder and CEO of TwinSix, a leading finance transformation implementation consultancy. He spent over a decade on the practitioner side of finance, using, evaluating, and implementing planning systems from the inside of finance functions before moving to the other, vendor side in leadership roles at Vena, Planful. He founded TwinSix to bring all three perspectives to bear for finance leaders, modernizing off legacy systems, building scalable, high-performing finance functions through consulting, implementation, and change management. Justin, it is a pleasure to have you here with me today.

Justin Merritt (00:46)

Thanks, Charlie. It's good to be here. It's a long time coming. You and I have known each other a long time, and I'm excited. Excited for it. Let's do it.

Charlie Liu (00:53)

It's, it's fine. I mean, we started our business around the same time as well. Very similar journeys. You've been in this industry for a long time. You sat on the other side. You give a whole different perspective. I've only been on the other side a little bit, but I think you've spent, what, five, ten years actually being the user, in a seat of leadership in a finance position, right?

Justin Merritt (01:15)

Yeah, and it was, I think, the perfect time. When I finished university and went out into the working world, I started at IBM, and it was kind of this experimental group that IBM had going on, and we were termed mini CFOs. So we got handed these baskets of loser contracts that were losing money and told...

Charlie Liu (01:36)

Mm-hmm.

Justin Merritt (01:37)

..."You gotta figure out how to make us money on these." And so it was wild, because, you know, right out of school, a lot of knowledge from school, but not a lot of common sense in terms of how things work. And then working with some really experienced veteran IBM delivery folks and things, because these were a lot of big, heavy contracts for, like, managing SAP environments and stuff. And so it was wild, because we get to, I guess, take all those case studies and theories and things I learned in school and get to try applying those, and then the complexity of big corporate enterprise agreements and all the terms and conditions. And then, yeah, moved to Verizon, again kind of right place, right time. It was this massive work migration from one center Verizon had to a different center, right after the MCI purchase of Verizon out of bankruptcy. Sarbanes-Oxley, you know, that whole thing. But got to work in internal audit systems and some intense stuff about how data flows, and really shaped a lot of that view. And then went over to Oshkosh Truck Corporation, and really started there kind of right as the economy fully tanked into the gutter. So you...

Charlie Liu (02:47)

Perfect timing.

Justin Merritt (02:50)

...know, in heavy equipment manufacturing, and yeah, I mean, we had just completed an acquisition. That was a three billion dollar acquisition. It doubled the size of the company overnight. And then the markets disappeared. 98% of the markets gone, basically overnight. And because of my background with systems and things like that, I got to be part of this really cool team that essentially just worked inside the company on how do we move work, how do we insource from suppliers, how do we reactivate, you know, different facilities in different ways, and then how do we take and capitalize on any opportunity there is to basically keep people employed across all of our facilities worldwide? And it was, yeah, an incredible seven years at Oshkosh doing all kinds of crazy stuff and working on crazy projects, everything from climbing inside of garbage trucks to, you know, setting up remanufactured lines to, you know, selling off some businesses that weren't performing. But yeah, in that I got a lot of experience with financial technology, and that's what really led me over and how we met, because I came over to Vena early days and you were there and we worked on the same team together. You then decided to leave, I guess, because you didn't like me or something. And then yeah, here we are. So, you know, a few... I guess almost ten years later, here we are, both with our own consulting and implementation shops. Kind of crazy.

Charlie Liu (04:14)

You came up through a lot of tech. I actually didn't know that. I didn't know you were that heavily involved with systems, processing, tech from the beginning. What I found is a good... I wouldn't say just consulting, because you don't have to just get consultants, but I think a big skill that's needed today is just systems thinking. Being able to actually think about data, process, and technology within... I think it was Paul Barnhurst I was talking to earlier this year, where I think with AI, a lot of things are getting easier. And it's getting emphasized, the better you are at systems thinking, moving data between different processes, different systems, actually deploying processes, skills, checks, QA validation, figuring out how to actually incorporate that into real day-to-day life with actual business processes. It requires a lot of that. It allows you to think about process and systems thinking. I actually think of it as probably the most useful skill these days, because a lot of things, even if you don't know accounting too well, if you don't know how to code in SQL, AI can take care of a lot of that for you. And I know you and I, it's been interesting. We started a consultancy around FP&A systems. You focus on Pigment, I focus on Vena, different systems. I do think they're all great tools and there's a time and location for everything. But a lot has happened. It hasn't even been three years, two and a half years for us, where AI just exploded. How are you seeing this impact how you think about building a consultancy, and how are you seeing it with clients?

Justin Merritt (05:55)

Yeah, it's a great question. I think back to when AI first kind of started popping up in this space, or in the tech space. And, you know, there was a statement, it's been sent on every meme possible, right? But it's like, AI won't replace you, but someone who can use AI will. And I think for the first time, especially in the last couple of months, I'm actually really starting to understand what that means, and that there's a lot of people who are just kind of blindly putting stuff in and you get an output. And that's what AI is designed to do, right? You put something in, it's gonna give you an output. And then you have people who have truly taken and embraced AI to make it an actual multiplier to the volume and even to the quality of work that you can produce. And I think that's an incredible thing that I've seen. I myself, yeah, I've really focused on trying to figure out how to do that well. And it's become very interesting. And one of the things that drives me crazy is, you know, the AI outputs, the format, you can always tell right away, right? You can see, that was AI. And it's like, I wish there was some kind of flag that says, "Hey, but this guy actually put a lot of work into this AI output. He just didn't waste all the time putting, you know, the formatting in the spreadsheet or the slides to get there." Because I think that's the tough part, right? And then from a client perspective, I can only imagine how difficult that is for them. You're now seeing potentially multiple consultants, very similar presentations, very similar formats. Which one is good? Which one actually understands our business or is going to lead to a successful result? And that's what I'm kind of seeing, the direction it's heading. It's like, the differentiation that used to come from yourself and me and our deep knowledge, and how we can talk about and present that to clients, is still there. But now, because we want to do things more efficiently, the way it looks and feels as an output sometimes can be very similar. And then how do you, as a client, tell the difference between two outputs that look and kind of smell the same? And you see some of that too with coding, right? And some of the software programs in the market. You can put them side by side and they're almost identical. If you took the brand off, which software is which? I have no idea. And it's that under-the-cover stuff that really matters, right?

Charlie Liu (08:23)

It's funny, because I went to Orange Theory yesterday, and I noticed they have little apps that you can track your performance, your heart rate, all that stuff. I've been going for probably two years now. It's the only thing I can drag myself to a gym. And all their UI changed after two years, and I can clearly tell this was from Claude. It's the standard Claude...

Justin Merritt (08:43)

Yes.

Charlie Liu (08:44)

...font. For a company that big, I was surprised. Nothing changed for two years, and suddenly you have a brand-new UI, UX across all the apps and everything. And it's all done by Claude. You mentioned a good piece about putting in that work. For us, we know for a fact one of the reasons we do well is just having that deeper understanding of clients. I know it's true for you as well. I actually recently did an analysis of our past sales, and it was the number one reason for us winning our deals. The number two reason was just that people like Excel for Vena. But you mentioned that in the last few months something changed. What happened?

Justin Merritt (09:24)

I think what I've seen is that companies, especially in the last few months, have just hugely put their chips in on AI, but without any plan. And so they've got, like, one of our clients, they're a very large enterprise client. We've worked with them for over two years. They were very resistant to AI, very careful. But now I was in a call last week, and it's like, all of a sudden, they've got all the AI tools, they're doing test user groups. But again, there's no real guidance or plan. It's more just that they're activating accounts across users. And I'm seeing that a lot. And that's, I would say, that big change. I feel like it's almost that bandwagon effect where some of the companies that have been a bit more careful are either getting pressure, you know, from employees internally or from the market, or just the advertising, right? And it's like, "We better do something. We better turn this on for people. Otherwise we're gonna miss it," you know.

Charlie Liu (10:28)

I've seen that as well. I've seen a lot of companies being slow at it last year, primarily because of security and data reasons. But now I'm noticing the same as you, they now have this huge FOMO, fear of missing out, that if they don't do this it's too late. And now, instead of a slow strategy to implement, they're just letting the gates flood open. And I talked to them and I asked, like, what happened? Last year, you did not want the AI companies touching your data. Suddenly you're just letting them touch everything. You're going from nothing to everything. And the constant answer I've gotten is basically, if you read the fine print from Claude, and if you are on the enterprise license, they do not use your data for training. And if that's what they say, we'll agree to it. And suddenly everyone now is extremely comfortable giving up full control. Their data, their database, access, everything, and just deploying everywhere. I... yeah. Yes, yes, exactly.

Justin Merritt (11:27)

Because that's never backfired ever for anyone.

Charlie Liu (11:33)

But at the same time, I haven't really seen a lot of companies. There are some companies. I find the smaller ones, change management is always easier with smaller companies, but a lot of companies actually haven't had great adoption. Forget all the headless agentic, forget all the really complicated, big multiplier stuff. Just general adoption, I found it really poor. Even for us, we were at 10% this time last year, and we thought, we're a tech company, we're a consultancy, we know data, we can get adoption really quickly. I think after nine months, we were at 10% adoption. We're now just at 65%. I see other companies having even more difficult of a time just getting users to actually use it. How are you seeing it? I know you have more SaaS and tech clients. How are you seeing it, both for you at TwinSix and for your clients?

Justin Merritt (12:23)

I would say the same. We actually, at TwinSix, scaled way back on our licensing because of exactly that. Because we saw there were people using it, and using it very well, and then there were people basically using it just for Google stuff. And it's like, okay, well, that's not helpful. Google still exists, it's still very free. And then...

Charlie Liu (12:46)

Yeah. Very free, yes.

Justin Merritt (12:48)

Yeah. And then, for the client side, you kind of see the same thing. We have several clients who are very AI-forward, but it's that same thing, that have kind of those pockets where you have one or two folks who are extreme power users, they're creating the markdown files, they're creating the skills, and then they're propagating those out. And then you have kind of limited or forced adoption on some of those skills across the team. And I haven't heard any resistance statements or "why not?" I think it's just people. It's like anything, right? It's that change management, and how do I integrate that into my workflow? Yeah, the thing that cracks me up is the Google statement. We have one of our clients that is publicly traded, somebody's working through some preliminary data validation, and they're trying to get... They were, they asked me, "Hey, do you know if we have anywhere in the client files this, you know, revenue number?" And I'm like, I share my screen and I Google "revenue for this company, quarter." And I'm like, "Here you go." They're like, "Whoa, you can just Google that." It's like...

Charlie Liu (13:56)

Yes. Yeah.

Justin Merritt (14:01)

...it's like, "Yeah, they're public." And they were trying to ask Claude and all this stuff. And it's like...

Charlie Liu (14:04)

Yeah.

Justin Merritt (14:05)

...yeah, something, yeah, it's that path, right? And I think it speaks to that adoption component. You have some people who just kind of have shifted their paradigm, and it's like, "Now I'm just gonna go to Claude and I'm gonna ask everything to Claude," versus, you know, the toolbox approach that you and I have been fans of and talked about forever. It's a tool in the toolbox and you need to know when to use it. And I think right now it's the tool in the toolbox that people just don't even know what it's even for. What can it do for me?

Charlie Liu (14:32)

We did the exact same thing. We scaled back licenses. We had licenses for everyone. We took them away. Then we tried Gemini, because it was covered just by our Google account. Now we're testing with Claude as well. Similar case. We're now scaling back a little bit. Initially we just let it loose, let everyone get access and see how the team performed. We did actually pick a person to kind of spearhead it, to really own the adoption across the organization. That made things a lot easier for adoption. We started actually pushing training sessions, show and tell. That was actually a game changer for us. And I think it was two months ago, I got this from a client of ours, a CFO explained to me how he got adoption across the team: he made it a limited resource. Just like your budget that you have for...

Justin Merritt (15:25)

Mm-hmm.

Charlie Liu (15:25)

...your department every year, every quarter. It's another budget. We only have so many licenses. You have to show that you're using it, you're getting output from it. It can't just be token usage, because we've seen this, people can fake that easily. And then you basically allocate that resource to the team. So we did the same thing as well. There's a limited number of users, and you only get it if you're actually showing output and usage from it every week. And from there, that's how we were able to push our utilization, or adoption, from 10% to roughly 60%. Anything on your side that's worked?

Justin Merritt (16:05)

Yeah. The biggest thing is very similar. It's that center of excellence approach, and then the training becomes the big piece. And that's the hardest, because that's where you get into the change management. The show and tell has been really good. But then it kind of comes back to that reinforcement of taking that and actually running with it. I think one of the gaps from a product standpoint that we've seen with AI is just that ability to propagate things across. That collaboration, you know, is really difficult. Say you have a client and you create a project for that client. I mean, it's insane to me, but you can share that project, but it's not truly shared. It's not a collaborative workspace for that project. You see your stuff, they see their stuff. And that, to us at least, is still a pretty major holdback, because when you have multiple people kind of trying to collaborate and work on something, it causes them to drift away from actually using AI, even if you've got some good skills and stuff for them to use, because they're still having to manually exchange some information back and forth. And, yeah, I'm curious if you've done anything around that.

Charlie Liu (17:15)

Yeah, that's a great point. You've... so I went into a deep rabbit hole this year. Basically...

Justin Merritt (17:22)

Yeah.

Charlie Liu (17:23)

...for weeks I was trying to figure out how to build a scalable, collaborative Claude setup so that everyone can benefit from your markdowns, your references. I was talking to engineers, I was learning how to use GitHub. I've concluded, as of right now, AI is largely an individualistic tool. It's not great for collaboration. I've long given up. I've put in probably 100 hours plus, sleepless nights, trying to figure this out. It's weird. It makes you really strong individually. But if you try to share your markdowns, your skills, because it doesn't work the way that someone else likes to work, I found adoption is poor, and if they're good at what they do with AI, then they would change the references and skills to suit themselves. And most people just leave it. And then finally, it's not a great collaboration tool, but it makes you yourself two, three, five, six, ten times stronger than if you didn't have it. And you and I have talked about this multiple times. We're seeing the labor market, the resources that we hire, especially the younger... both the younger and the senior folks that we hire, it's a little bit different from what it used to be ten years ago. Some of it may be impacted by AI. I'm seeing a little bit of that as well, but what have you seen?

Justin Merritt (18:42)

Yeah, I think what I've seen, and it's not new, but it's, I would say, kind of this thing that's crept up, is that shift in priorities. Like, to be constructive about it and not say terrible, nasty things about people.

Charlie Liu (18:58)

Yes, yes, yes, yes.

Justin Merritt (19:00)

I think it's a shift in priorities. And while I don't identify with that shift in priorities, I think it's interesting to watch happen, because I don't even think it's necessarily generational. I have people, you know, our age who also, I'm seeing this kind of thing where they just really don't have their heart and soul in work anymore. And I know not everyone finds jobs they love and stuff like that. But generally, if you rewind the clock, you always had... if you use the GE model, right? The Jack Welch ABC player concept. But generally people understood, if you're working at a company, you're gonna show up, you need to perform, or you're probably not gonna have a job. And it's really interesting, because even with the labor market and these horrible stories you hear about people struggling to find jobs and stuff, I'm seeing, not just for us sometimes, but also at clients, where they're just really having a hard time identifying candidates who are going to come in and want to... I think it's, like, want to leave your mark. You know, they want to be a part, they want to learn, they want to help improve the operations, improve the business, leave their mark on what they do there. And I think more and more it's crept up on us, but I think it's very prevalent now cross-generationally. Like, people, it's just a job, right? And so even in that professional sphere, it's kind of like, I think of that American Beauty movie with Kevin, I think Kevin Spacey, and he goes and interviews to work at the burger place. Like, "Yeah, I just basically want to show up, get a paycheck, and have zero responsibility." You know, like, and yeah, I don't know. It's kind of an interesting thing.

Charlie Liu (20:47)

I think there's lots of reasons behind it. We see it as well for sure. I see it with clients as well. Getting strong talent has always been difficult. I think AI, there's a little bit of that. I don't think it's the whole thing. I think from what I've seen, especially since I live in different places, the rising cost of living sometimes makes it so difficult that you don't see a way out. I mean, I was an immigrant. I knew I came in, I knew that if I worked my ass off, with a little bit of luck, things could get easier. And I do hear sometimes from folks that the cost of living, real estate, things like that, it's so high now that even if they put in the hours, they don't see a path out of it. I also think AI made things easier. I don't think people are going through the trenches. You went through the accounting, I think it's more the accounting route earlier, right? Accounting is difficult early on. I went through the data front. What takes five minutes now used to take me a week, two weeks to code and debug, and you kind of go through that grind, for lack of better words, and you build up the grit for it. But now things are getting a lot easier, right? You don't have to struggle, you don't have to... I don't know if "eating glass" is the right word for it, but you kind of... and I think you lose that muscle as well. If things become easier, when the tough times come, in your example, leaving a mark, that takes work. That takes dedication. And it's easy not to put in the work now, especially with AI. It's easy to just be average. And I think what people are not realizing is these tools are available for everyone. The outputs, the expected outputs across the board should be going up. And if you don't change your benchmark and the way that you work, you're going to be left behind. I spent a lot of time thinking about this, for MCC, because I define my job as, we need to make sure we survive the next six months and the next two years. That's how I think about my role. And a big piece of training the team and mentoring them is to try to push them to basically realize that the benchmark's changing. I think for the last 20 years, it's been maybe growing a little bit, right? Maybe when you moved from paper to Excel, there was a big improvement. Database...

Justin Merritt (23:20)

Yeah.

Charlie Liu (23:21)

...macros, things came in. I think that had some improvements, but it's nothing like what we're seeing now. And if you don't put in the work, you're going to be left behind. I found some luck in studying how people mentor and coach. A lot of this actually came from when I was studying how people train sales. Because that is a difficult place to be, right? Especially doing BDR outbound work. That is the worst, right? It's difficult. Right. So it's not something you want to do day in, day out.

Justin Merritt (23:55)

There's two... there's two jobs that I think everyone should have to have. No matter what, everyone should be forced to work in a restaurant and wait tables, you know, and not...

Charlie Liu (24:07)

Yes.

Justin Merritt (24:08)

...bartend, you know, or work in the kitchen, fun stuff. Literally wait tables and...

Charlie Liu (24:11)

Yeah. Yeah. Yeah.

Justin Merritt (24:14)

...get cut after making four dollars, and have...

Charlie Liu (24:16)

Yes.

Justin Merritt (24:17)

...also have a thousand tables all at once and get screamed and yelled at. You know, it does something to your soul that is important. And that is one, and then the other is the BDR function. I even, for me, I have had to do that. I worked in jewelry. That was my first... in university I worked at a jewelry store. Yeah.

Charlie Liu (24:39)

I had no idea you... yes. Yeah.

Justin Merritt (24:42)

And we had to do cold calls. We had these call-down lists for, you know, special events and stuff like that. And so you're working in the store, in between somebody coming in, you literally pick up the phone and you get a call, you get the script, and you gotta try to get people to commit to coming into the event. And you gotta make, you know, anywhere from a hundred, a hundred and fifty, two hundred dials a day while you're also working the floor and selling jewelry. And what a soul-crushing experience, but in a positive way. You come out the other side and you really appreciate how difficult it is to call to action somebody to make...

Charlie Liu (25:20)

Mm-hmm.

Justin Merritt (25:21)

...a purchase that's nice to have, versus something that is a forced, you know, requirement, like from a government regulation or something. It gives you a better perspective, I think, on people in general, those two types of jobs.

Charlie Liu (25:36)

I was a dishwasher for two years.

Justin Merritt (25:38)

Yeah.

Charlie Liu (25:39)

I realized very quickly... I wasn't a great student at that point. After that, I was like, I need to get my shit together.

Justin Merritt (25:44)

Yeah.

Charlie Liu (25:45)

I need to make it to university and college. I can't keep doing this. It's just rough. Sales is difficult as well, right? How long have you done it? What I've noticed is, we have a head of BDR here, Jacob, and I watched him coach his team, which are these fresh grads who are excited coming out of university, college. And now they realize they have to make two, three, four hundred dials a day. And there's a level of coaching there that's been helpful in terms of what I found to push the team to step up, for lack of better words. How long did you do sales for? And have you found anything that's changed how you approach hiring, coaching, mentoring to help them step up?

Justin Merritt (26:23)

Yeah, I think the biggest thing is trying to identify... it helps you learn how to identify people who can have that self-start conversation. Because especially in what we do, in our businesses, being able to raise your hand and ask a client a clarifying question, or anything like that, is so important. And part of that soul-crushing is realizing that you don't know all the answers and you can just ask people questions. It's okay to just pick up the phone and ask people questions. And I think that is one of the biggest things, you know, that I learned. Like, I'm still trying to figure out...

Charlie Liu (27:06)

Yeah.

Justin Merritt (27:06)

...the best way to identify it, but those kind of little tiny skills can really balloon or blossom, right, into a really great employee.

Charlie Liu (27:15)

You bring up the self-starter. I was having a conversation with my team earlier this week, and we were mentioning that there are a lot of things we can coach. But what makes a really good employee... I don't want to just say employee, because I think these are characteristics that can help you for life. If you break down what self-starter actually means, I think what we came down to was, one is curiosity. One is... I gotta double check my notes. But one was curiosity. And the second one is challenging their beliefs. So I think a lot of folks, they just believe, like, it's not possible. It's not possible. And they just won't do it. If you combine those two, and because things are so easy, information is so easily accessible now, if you're curious and you believe it's possible to change how you work... I think it was Bill Gates' favorite word, right? Like, hire the lazy ones. I'm butchering it, but it's fine. People...

Justin Merritt (28:13)

Yeah.

Charlie Liu (28:14)

...willing to put in the work to improve the process. Like, willing to put in four hours so that you can shave five minutes every day. That is a really good characteristic to have for life as well. Do you think this is something that can be mentored and learned, or is it something kind of innate?

Justin Merritt (28:36)

I, you know, in all the years... it makes me feel really old to say it that way. In all the years of managing people, hiring, firing, I don't know that I've ever successfully trained someone in that skill. I've unlocked it in people...

Charlie Liu (28:54)

Yeah.

Justin Merritt (28:54)

...right? Where maybe they didn't realize that they had it, or it wasn't something that they saw the value in. But I don't know that I could say I've ever been able to successfully teach someone that. You know, it's like you said, I mean, from my experience, it's kind of, you have it or you don't. And sometimes, though, you might have it and it needs to be unlocked. But yeah, if you don't have it, no matter how hard I've tried, how hard you've gone after it, with those folks, they just always revert back to their same, you know, bad habits. And it becomes a lost cause.

Charlie Liu (29:33)

It's interesting you said the word "unlocked," because I was just thinking about my history. I have unlocked it in people. I didn't train it, I didn't coach it into them. They had it in some other aspects or different parts of life before, maybe not for work, but they had it. I think combining that self-starter curiosity and self-studying with AI these days, you can be quite powerful. How do you look for it? Because you do a lot of recruiting as well. I mean, we're a consultancy. A big thing we have to...

Justin Merritt (30:02)

Yeah.

Charlie Liu (30:02)

...think about is pipeline for people.

Justin Merritt (30:05)

Yeah, unfortunately. All day, every day.

Charlie Liu (30:06)

Yes.

Justin Merritt (30:07)

Yeah. No, the things I look for are really... and this goes back to Jack Welch, you know, as well.

Charlie Liu (30:15)

Yeah.

Justin Merritt (30:15)

You know, he put it in words for the first time for me. And that is, you know, somebody who you'd want to sit down and have a beer with, right? So like what we're doing right now. You know, somebody who you could see yourself just voluntarily spending time with, and you think you'd have, like, an enriching conversation or dialogue. You know, I think that's first and foremost, because as someone who has those traits, right, who likes those things, you're kind of, I find, naturally drawn toward other people who also have those traits. And so if you feel that kind of personal bond or connection quickly, that's usually a good sign. I think the other one is, you know, especially in today's world, people love to say that they read, right? Everybody's reading. But I'm pretty sure nobody other than, like, me and you is reading anything. An exaggeration. But, like... Yeah.

Charlie Liu (31:04)

I don't read much now. It's a lot of video consumption, but thank you for believing I do.

Justin Merritt (31:11)

Like, our interns last week were showing me this, I'm gonna butcher it. It's like, Learn LM or something, or NotebookLM.

Charlie Liu (31:20)

NotebookLM. NotebookLM, yeah.

Justin Merritt (31:22)

Yeah. And I made fun of them, because they were showing it to me and all this stuff. And I was like, "Well, in all the time that you spent setting this up, you could have just read the article that I sent you, and then we could have a discussion about it. But instead, you have this generated podcast from the content." Like, yeah, I mean, it's cool, but couldn't you just read the article and then we could be talking about it? And that's kind of, I think the second part is, does that person have interests, right, that they can just talk about, and they are consuming information because of those interests, right, outside of work? Those are things that you have to be that self-starter to want to pursue or go after. You know, it could be literally reading...

Charlie Liu (32:08)

Yeah.

Justin Merritt (32:08)

...fiction or something, or reading the Jack Reacher series, right? It doesn't have to be anything crazy, but it's just that dedication to, "I'm really interested in this. And even though nobody is telling me or forcing me to go do it, I'm choosing to go expand and do it. And I can talk to you very coherently and engage in a good conversation about it." Because that means you're retaining it, you're processing it, you're thinking about it. Those two things to me are kind of the two things that, you know, again, it's hard, it's hard to identify, especially in a virtual setting versus sitting down face to face. But I think those two things are the biggest things I've noticed and seen.

Charlie Liu (32:49)

When you talk about that, I hear the word "passion." I think that if you can sense passion in someone, it doesn't matter what it is, they're generally putting time over and over again, they're putting in reps to get better at something, and that translates quite well. And most things aren't easy the whole time if you're getting good at it. We found that hiring athletes helps. They've done something over and over again. They've put in their time. They're able to kind of build the grit around it. They're also constantly looking to improve. You bring up a great point about being able to have a conversation about it as well. They're passionate, meaning they're happy to talk about it. They can nerd out, they can get really into it, regardless of whether the other party cares for it. I think that helps as well. And I think, to your second point, it requires work. Now, I'm a big fan of NotebookLM. So I think it's a great tool. I've... yeah.

Justin Merritt (33:47)

It is a great tool. It is a great tool. There's no doubt about it.

Charlie Liu (33:50)

I think also, for you and I, maybe we're a little bit older, you have to think about the generations being a little bit different. People don't read as much now. People are listening to podcasts, things like that. But I think what you're trying to say is, sometimes you have to just do the work. There's no easy way out. I learned this in the last six months, where it really improved my AI usage. You have to sit there and read the output. And it's some of the driest content you can get. But if you don't read it and you just let AI keep going... the thing is, AI is the fastest way to produce garbage now, the fastest way to produce goop now. If you don't read it, if you don't read what it's planning to do, you're gonna get a lot of garbage. So you have to sit there, right? Do the hard work.

Justin Merritt (34:38)

So what's funny when you say that, and why I'm laughing, is because I get made fun of sometimes because I literally expand out all of the things, and because I have the coding background and stuff as well, even when it's writing the Python scripts and stuff, I read all of that as it's going. Because to me... and I'll tell you, I mean, maybe 60%, 70% of the time, I end up hitting the stop button, providing a correction before it's even gone to the output, because it's misinterpreted something I've asked, or it's misinterpreted a piece of data in a model that I've put into it. And so that's why I'm laughing, because I couldn't agree more. It is a huge accelerator and a huge value-add component. But reading that output, and even reading the process it's going through, can help you produce such better results in the first pass, you know.

Charlie Liu (35:39)

Day and night. No one told me this. That a big part of good AI usage is just reading. So now I print it out, actually. I do a lot of planning mode, do the whole thing. And I'll print it. Something's like 10 pages. I'll grab coffee, basically leave my computer so that I don't get distracted, and then just sit there and read the whole thing, write my notes out by hand, then I go back and type it out. It's made such a big difference. And I think that goes back to what you were saying, right? You have to do the hard work. There's no easy way out. I mean, AI's gonna... yeah, AI has made it easier than ever, but you still have to put in the work.

Justin Merritt (36:16)

Yep. Yep. And I can tell you, man, I'm a handwriting person through and through. I always have a notebook, but, shameless plug, the reMarkable products are awesome. I got, for Christmas, the little one that's, like, the size of a Moleskine notebook or whatever. Man, it's great. You can load in that document and then you can hand-mark it up, and their handwriting... my handwriting is terrible. Even I can't read it most of the time. It has gotten so good, it can translate that, boom, right into text, and then you can take and actually work with it directly. That, to me, is where I get really excited about AI, you know, because...

Charlie Liu (36:57)

I was looking... I couldn't... I was trying to justify the price. I started actually looking at the mini.

Justin Merritt (37:03)

I don't... it's, to me, it's...

Charlie Liu (37:03)

I have to check... I have to check it out again. Yeah.

Justin Merritt (37:06)

...the, it's that segregation. I'm always an early adopter, and I got the iPad when it first came out with the pencil, and I was the guy in the meetings, you know, taking notes, and in manufacturing everybody's making fun of me. They're like, "What is this guy doing?" And...

Charlie Liu (37:18)

Yes, yes, yeah.

Justin Merritt (37:20)

...but it's the distraction component. That's what reMarkable has solved, just like the Kindle, right? It's like, why do you read so much more on a Kindle versus on your iPad? Well, because there's no distraction. Because it's literally just for reading. The reMarkable, same thing. Why do I now naturally pull out the reMarkable to take notes? Because it's for note-taking. And so it has truly supplemented that corner, you know, of my existence. And the fact they're searchable. I mean, I have stacks of notebooks that are just filled with notes. And you could...

Charlie Liu (37:53)

Mm-hmm.

Justin Merritt (37:53)

...you know, you can never go back to them and find anything. And then I can never let them go, because I have such an emotional attachment to them. Now that is all digitized and it's searchable, which is unbelievable, you know, in my own handwriting. So, shameless plug. Do it. Just pull the trigger.

Charlie Liu (38:10)

I'm gonna look into it, because that was the primary reason I started doing this research a few weeks ago, because I have all these notebooks with notes everywhere and I never go through them and can't find anything. So I'll look into it.

Justin Merritt (38:22)

Yeah. Yeah.

Charlie Liu (38:25)

We have a few minutes left. Justin, this is really great. It's such a changing time right now, and I know we're also seeing it with clients as well. What I've seen with a lot of clients is that many of them don't realize that getting good at AI is actually becoming like mini dev shops. You have to think like a product manager. What are your requirements? If you do sprints, if you're good at it, you start producing tools that serve specific niche needs of yours. But at the end of the day, you're kind of building products sometimes, and that requires maintenance. And people don't think about that. And I've noticed a lot of clients in finance, especially with what you mentioned earlier, there's a big push to use AI, AI, AI. And I've noticed a lot of folks kind of pause traditional strategies, data warehousing, building infrastructure, getting tools to help them with better reporting, budgeting, consolidation, and things like that. I'm actually a firm believer that a lot of it's going to slow down soon, because I've very rarely actually seen clients get results from it. And they're realizing there's still a place for your standard SaaS tools, but they're better with AI. Have you seen anything like this? I know in Pigment there's a big kind of product released in the last six, nine months, a year around AI. Have you seen anything like this?

Justin Merritt (39:48)

Yeah. I mean, we've seen... I would say that's the typical curve, right? There's the excitement, you know, at launch, and then the use. And it's like, okay, we're doing some stuff, but then it tapers off, just like we talked about with the internal usage of various tools. And I think, like you said, it's understanding that it's probably the hardest change management that an organization will ever face, is how to actually integrate AI into what they're doing. Because you're not just having a new tool, you're truly trying to change or break habit structures that people have. And that's the hardest thing to ever do. And then especially in the realm that you and I are in, where you're mostly working with finance and accounting people, there's a natural tendency to not trust. And I think rightfully so. And so when you have AI that, I mean, literally doesn't understand debit versus credit, and just like a lot of people say, "Credit is a negative number." Well, it's not actually a negative number. That's just how it looks in a database, because you have to assign a negative to it. And so that, you know, creates that mistrust. And I think those pieces, right, are where that slowdown you know, I couldn't agree more. Like, I think there's a lot of hype. I think it's a lot like what we saw with cloud. When cloud first became a thing, it was this huge, "Are we in the cloud? Are we getting to the cloud?" And then now, you know, and then it really slowed down. People took a step back. Like, "What do we need to have in the cloud?" Right. And I think that's where we're gonna get to, but I think much quicker than with cloud, because, like you said, the pace is so much faster with AI, and you can do so much rapid prototyping. I think people are getting to that conclusion much faster.

Charlie Liu (41:39)

I love what you said about prototyping as well. I look at it as we're able to implement them a lot faster now. Not because... well, the build partially is getting faster for sure. Any sort of data transformation, things like that. But I think the biggest thing is the prototyping. And I wanna emphasize the word "prototype," because literally last week I was building a model in Excel, and if I did not read the formulas one by one, I did not catch that it decided to just hard-code some figures. Because I asked for a model, I didn't have the greatest documentation. I should have, it's not me, but it just started hard-coding certain drivers for me. And if you didn't do the work, if you didn't actually read the formulas one by one, you would not notice that the model doesn't fully work. And some of the drivers' assumptions I made were just incorrect. Which goes back to what you're saying. I think it's a great way to prototype what you're looking for. But you still need someone that's gone through the trenches, that knows how to actually build it and architect it properly, to really set it up so that it's scalable, reusable. I think that's the other word as well. A lot of these prototypes look great and they might work right now, but they're not designed to be built and reused for the long term. And maybe that's not where we're going with the future, because it's so fast to build. But a lot of accountants and finance people still work that way. They like a model, they like a report, they like data moving from A to B the same way day in, day out. And I think when I talk to a lot of CFOs, there's a big question now: what should be AI-driven, what should be mechanical? I think people need to think about it. Certain things still need to be mechanical. I need A plus B to equal C every day. I can't have it just vibe out its feelings and think of a different...

Justin Merritt (43:24)

Yeah.

Charlie Liu (43:24)

...math every time.

Justin Merritt (43:25)

Yeah. It's a really interesting thing when you get into stuff that is strictly rule-based, you know, and...

Charlie Liu (43:33)

Mm-hmm.

Justin Merritt (43:33)

...and sure, we could debate, like, IFRS versus GAAP, but what I'm talking about is the core, you know, of business, which is always still debits equal credits, assets equal liabilities plus equity, and then all the permutations that come off of that, you know, from the various rules and situations and depreciation schedules and, you know, you get into those things where there isn't flexibility, right? And to your point about hard-coded numbers, I've seen some of those same things where, you know, the AI makes certain assumptions or builds those things in. And then if you don't review, now, hey, next time you go to add a new asset to this model, it's gonna actually calculate the depreciation wrong, because it hard-coded 60 and it didn't create a table with the asset types and the acceptable depreciation life, because it doesn't know, right? And that's the thing. It's almost like it's the dumbest, smartest, hardest, best, fastest, sloppiest...

Charlie Liu (44:34)

Yes.

Justin Merritt (44:35)

...worker that you have. And you have to constantly be molding it like clay to get to what you actually need.

Charlie Liu (44:44)

And for now, there are limitations on context windows, right? Because we're designing systems that need to work for different modules, different parts of the business. Right now it's really good at building individual components really quickly. Like you said, the smartest, dumbest, fastest method to do it, but it's oftentimes doesn't have the whole picture in mind.

Justin Merritt (45:06)

Yeah. Yep. Yeah. And that's the challenge. And I think, to your point about the trenches, one of the things I've seen is that because that trench component has become smaller and smaller, especially with folks who have less work experience, they don't have that ability as much to contextualize the full picture. Because they never had to sit and do horrific month-end journal entries, you know, and stuff...

Charlie Liu (45:32)

Yes.

Justin Merritt (45:32)

...all the way up through building models, thinking about context, how will this roll forward year to year? You lose that contextualization skill. And, you know, I don't know how you teach that to AI or to a person. It's really hard.

Charlie Liu (45:48)

I don't have an answer for that. I know a lot of our audience are younger folks. I get a lot of folks actually in accounting asking how they switch to finance, and I don't have an answer for that. What I often say is that those trenches still exist today. Take advantage of it. I know it's the difficult part. I'm telling you to go enjoy the worst part, because it pays dividends later, because at some point it's going to disappear.

Justin Merritt (46:15)

Yep. Yeah. We're... yeah. Exactly. Cool.

Charlie Liu (46:17)

Justin, we're at the end of this. This is a lot of fun. Definitely excited to have you back again. We like to end the session basically just asking what your two cents advice is for the audience. What guidance or two cents advice do you have for the audience?

Justin Merritt (46:33)

I always like to tell people that if you can't teach it, then you don't own it. And what I mean by...

Charlie Liu (46:38)

Mm.

Justin Merritt (46:38)

...that is, you know, and perfect for what we're talking about right now, right? How many times have you worked with a client, or for those listening, how many times have you gone in to try to, you know, take something over from somebody or learn something from somebody, and they can't explain it to you? You ask a question, "Well, how does that formula work?" or "How does that happen?" and they can't explain it? And I think that everything we've talked about today kind of actually goes right into that. And it's, you know, if you can't teach it, you don't own it, because that means you truly don't actually understand it. And if you don't understand it, how can you own it?

Charlie Liu (47:14)

Absolutely. Absolutely. You need to understand it inside and out. You need to be able to explain it to someone else. You need to be able to teach it. A big piece that we spoke about is AI. If you don't know what it's actually doing, if you don't actually go through what it's actually built, you won't understand it enough to teach it as well. Well, Justin, this was a pleasure. Thank you so much for being on with me. Really...

Justin Merritt (47:36)

Yeah, thanks, Charlie.

Charlie Liu (47:38)

...really grateful for this.

Justin Merritt (47:40)

Yeah, it's been a lot of fun and looking forward to chatting soon. Thanks again for having me.

Charlie Liu (47:45)

Appreciate it.

Justin Merritt (47:46)

Take care. See ya.

Charlie Liu (47:47)

Thanks.