Episode 236 – From AI Assistant to Digital Staffer: Scaling Boutique Firms with Agentic AI – A Member Case with Alex Bratton

Most boutique firms are experimenting with AI assistants, but few have operationalized AI in a way that drives measurable business value. Member Alex Bratton, founder of LexTech, shares how boutique firms can evolve from basic AI assistants to digital staffers; autonomous agents that operate across systems like real team members.


What you’ll get from this episode:

• A framework for identifying AI opportunities for your firm
• A decision model for when to build vs buy
• A method for evaluating AI ROI based on business outcomes

Why it matters:

• Misaligned AI efforts waste time, money, and strategic focus
• Not leveraging AI will put you out of business in the future

TRANSCRIPT

Greg Alexander: Hey everybody, this is Greg Alexander. You’re listening to the Pro Serv Podcast, brought to you by Collective 54. If you’re new to this podcast. This is for founders of boutique professional services firms, so if you’re in the expertise business, this is for you. On this show, we aim to help you do three things. Make more money, make scaling easier, and make an exit achievable. And on today’s episode, we’re going to dive deeper into AI agents. This is an evolving field. We have a member with us, Alex Bratton, who is an expert in this area, and we’re gonna get into the 3 different types of agents, and he’s got some interesting vocabulary for us that we’re gonna talk about. But before we dive into that, Alex, it’s good to see you again. Would you please introduce yourself to the audience?

Alex Bratton: Absolutely. Greg, thanks, it’s great to be here. So again, I am the founder, and most importantly, chief geek of LexTec. I am an applied technologist, because technology is supposed to help people thrive at work. That’s my why and our why.

Greg Alexander: Alright, sounds great. Alright, listen, so you’ve identified 3 types of AI most firms are encountering, and I want to start there, so can you walk us through those and explain why that distinction matters?

Alex Bratton: Absolutely. So let’s start with the one that everybody has been hearing about for, I’ll say, two-ish years now. And that is the reactive assistant. Think ChatGPT, think Claude when I say that. And reactive means it’s my thinking partner. It’s helping me do things as I sit down, I ask it a question, it executes on work, and there’s a ton we can do with that. The systems that you jump into that create graphics for us, or music for us, or execute on things. They work, but they’re reactive. We, the human, have to poke them to really drive the value, and then have a back and forth. Everyone has to have that, but that’s different than when we start using that word agent. This is the reactive side.

Greg Alexander: Go ahead. Yeah, that’s the first one. What’s the second type?

Alex Bratton: So from there, we move to… and here I’m gonna… this is the only place I’m gonna use the word agent… a task agent, because agent means different things to everybody, that’s our challenge. A task agent. And I’m very specifically linking those two together, because it’s an agent that can walk through a series of steps. So, in the old days, you might call that programming. Nowadays, it’s, okay, here’s… do these five things in order. It executes a task. Maybe it moves data from point A to point B, but it’s… it’s moving through a series of steps, which is very different than the third type. And this is where folks like to talk about agentic behavior. My terminology here is a digital staffer. They’re a digital person on your team. They’re goal-based instead of process-based. They know they have a job, they’ve got tools to do it, and they execute it. And those three are very, very different, and they all add value to organizations.

Greg Alexander: Okay, that’s very helpful. So, in our prep for this podcast, you had sent me a note. And you gave me some practical examples, and I want to talk about these, so…The first one was Clario, which preps you before meetings. Then you have TESS that mines transcripts. So, and then these tools are doing that like a human EA, or like a CRM, or like they can’t, I guess. I don’t really understand the difference between Clario and Test. Can you help me understand this?

Alex Bratton: Absolutely. So what you’re talking about here is actually my personal support team. So I have a digital EA that’s actually a team.

Greg Alexander: Okay. So Clario is a part of that.

Alex Bratton: So Clario is a digital staffer, and his job is to give me a daily briefing of who am I meeting with today, and what do I need to know before I walk into each meeting. Okay. And Clario produces an email report to me of, hey, for each conversation, here’s what you need to know, and the first time I knew that there was something really useful here is when Clario was briefing me and telling me, and here’s how you met this person.

Greg Alexander: So, I mean, so many folks, I can’t remember who made an intro, how did we connect, what did we talk about last time?

Alex Bratton: So Clario is surfacing that, and the key here is he pulls the context from my email, my CRM, and my calendar.

Greg Alexander: Okay.

Alex Bratton: So grabbing all of that to crunch it down. So that’s a daily briefing. Tess takes a transcript of a conversation and extracts all kinds of useful information. Such as, what’s all the technology we talked about? Who are all of the people that were mentioned outside of the conversation? What organizational goals or organizational structure was mentioned? So, all that stuff I might want in the future that you don’t get just from a Zoom recording I need more than the to-do list, I need lots of contextual, so Tess is doing that. I’ve got a few others, Folia partners with Clario. Folia….does research on every single person I meet from outside the company, gives me a two-page dossier. Automatically, that’s her job. She just knows, she looks at my calendar every morning puts the briefing together, crunches someone’s LinkedIn down, and from that, I get, what was their career trajectory? You know, where did they go? What important accomplishments did they have? And assuming they’re a prolific poster, oh, such as yourself. What’s important to them? What do they post about? How might I best connect with that person based on what they’re interested in? So it’s all of these things supporting each other.

Greg Alexander: And these names, these are the names you’ve given your digital staff members. These are not names of technology tools, correct?

Alex Bratton: Exactly. This is the step to humanizing things, and we actually have pictures, we have headshots for every one of these. Okay. Because the more human we make the tech, the easier it is to work with.

Greg Alexander: Yep. So, when I was learning from you about this, one thing that jumped out at me is that you know, when you think about the three types of AI that we started the conversation with. This digital staff member seems to make the most sense when they’re pulling from multiple systems.

Alex Bratton: Exactly.

Greg Alexander: Okay, so tell me why that is.

Alex Bratton: So the rule in general I see for AI when you’re looking at it is, if AI is only pulling from a single system to extract some insights or otherwise. That should probably just be a feature of that system. Now, it doesn’t have to be, and I’ll come back to that, but the vendor that sold you that CRM, that email system, that whatever platform, they have to build AI, and everybody does. So if you’re only tapping one system, it’s valuable, but unless it’s proprietary to you. If you’re trying to put your own proprietary insights on top of that system, then totally do it. Then that does make sense. But if it’s something that, you know, you’re just pulling something the CRM should already have. That’s their feature. You shouldn’t build that.

Greg Alexander: Okay, so that’s kind of a litmus test to determine when to build something for yourself and when not to. Is that a fair way to think about that?

Alex Bratton: That’s a great way to put it, and you mentioned something earlier, the… and Clario, for me, is a good example of it. Some of what we’re talking about goes beyond what a human EA might be able to do. Clario looks at all of my emails and CRM notes for the last 6 months, mashes them together, and tells me what I need to know. A human EA might do some quick flipping and might remember that, but we can go a little bit beyond and build a lot more context.

Greg Alexander: Yep. Alex, let me get your opinion on something. We’ve had some members in Collective 54 that were early adopters, maybe even pioneers. And they jumped on this thing really quickly, which I love about them, and I dig that about them, and they built this stuff, and… You know, and a year later, you know. ChatGPT4 or 5 comes out, and what they built is, like, now obsolete. So, they’re now gun-shy, right? And they’re saying, well, you know, should I really invest my time and money in this thing, only for it to be rendered obsolete, you know, very quickly? So, how do you determine, like, what’s future-proof?

Alex Bratton: That’s a great… and future-proof is the magic word there.

Greg Alexander: One of the keys.

Alex Bratton: Pretty much everything we’re talking about today, and even for 2 years ago, was prompt-based. So you were putting some kind of an instruction set together, but the tech that the prompt runs inside of, whether that’s a new version of ChatGPT or the Agentic platform it’s running on, those things keep changing. So, when you’re building it, making sure that you’re creating it in a way that you can move it from platform A to Platform B, that’s really important, because none of this is one and done. All of this is going to require ongoing investment and care and feeding. That’s just the reality of where we are. Couple years from now, things will be a little more stable, but that is the reality. Which does mean, when you’re looking at what you build for yourself, or what you acquire from a source on the outside that can provide you that service, is it worth the investment for you to own it? For the pieces that I’m talking about, the Clario and Folia and other team members for me, those are things we’re offering to other folks to help them, so that’s actually becoming some of our go-to-market. If there was something that was very specific to an organization, then it might make sense to build that out yourself.

Greg Alexander: Yeah, okay, that’s a good way to think about that. Okay, so, because you’re a member, you know our members and who they are, and they’re, you know, entrepreneurs, and they want to go quick, and they’re trying to do big things, and they’re not burdened with a bunch of, sacred cows and legacy thinking. How should they think about when it’s time to move from the… the assistants, I guess, as we’re calling them, to what you’re teaching us about today, which is digital staff members. I’m looking specifically for when.

Alex Bratton: Now. But not just from an EA perspective. I would actually take a half a step back, look in the organization, and look at, is there anything going on that, again, a digital staffer that… and what we’re talking about here is if you gave it access to the same systems could pull that information, again, maybe from a more broader context, tapping into your project management system, tapping into CRM, and generate the insights that you’re looking for. I’m not saying everybody should go out and start putting in 20 of these staffers tomorrow. Where is it going to drive business value? What’s holding you back today? If you have an incredible EA, go with it. Don’t make a change there. If you’re feeling that you’re not getting… I… again, I didn’t have this, Claro saves me a half hour a day. That’s priceless. If you’re feeling the pain of, I can’t get this information I need or synthesized in this way. Then getting a digital staffer to give you exactly what you need makes perfect sense.

Greg Alexander: Yeah, okay. And then in terms of talent, I don’t think any of our members have this talent in-house. They probably need to go outside to hire for this. Is this affordable?

Alex Bratton: Yes and no. Finding someone who can build something in the tech, the Zapier, the N8N, the relevance script, and put the pieces together, I’ll call that more the core engineering pieces, those exist. The challenge comes in the part above that, which is what should we be doing? And really having someone who’s good at applying to a business problem which tech or which processes are going to solve that. That may be something… those are individuals hard to find, you might have to tap a partner to help you do that kind of stuff, but that’s where it starts to get a lot more expensive.

Greg Alexander: You know, one thing that has personally frustrated me on this subject is… I, I think that AI is… is… As advanced as it’s becoming, it’s still really new. And you could end up spending a lot of money on something and not getting a great result. So I wish that all the AI firms that built this stuff would work strictly on outcome-based pricing. So, you know, if I deliver something.

Alex Bratton: I would expect nothing less from you, Greg.

Greg Alexander: Yeah, right. If you’re gonna build me a digital staff member, and it’s gonna make me money somehow, then I’ll give you a percentage of whatever that is. But, you know, to charge me a consulting fee, or a retainer, or a development fee, it just doesn’t sit well with me. Now, I know you’re in this business.

Greg Alexander: And I’m assuming that you have a pricing model, so I want to give you the opportunity to speak on behalf of everybody who disagrees with me.

Alex Bratton: Which, there’s many, right?

Greg Alexander: So why… so tell me why people that are looking to acquire this service, why they shouldn’t force their vendors to work on performance.

Alex Bratton: See, I don’t overly disagree with you, because.

Greg Alexander: I won’t focus…

Alex Bratton: if it’s not delivering business value, don’t do it. So whether you’re buying it externally or building it yourself, if you can’t tie AI and all of this work back to a time savings, a revenue boost, an efficiency, whatever the business metric is, if you can’t tie it to that. That’s, for those that haven’t heard, MIT released a study recently. 95% of AI projects are failing. They’re failing because they didn’t have business goals that folks were focusing on specifically. These have to be about focus on that. For me, everything I just talked about was time savings for me, the executive leader, or the leadership team. That’s worth it. So it… being able to articulate those values, absolutely. But as you were talking about, there’s folks that are gonna charge hourly. What does that even mean anymore now?

Greg Alexander: It’s yet another…

Alex Bratton: coder. That’s hard. That’s really, really hard.

Greg Alexander: Yeah, yeah. All right, well, very good. Well, Alex, we’re going to leave it there, because I want to save the rest for the private member Q&A, but I always appreciate your point of view. You’ve been very generous with your time in contributing to our community, so on behalf of everybody in Collective 54, thanks for being here today.

Alex Bratton: Thanks, Greg, this is great.

Greg Alexander: All right. All right, a couple calls to action for listeners. So, if you are a member and you want to talk to Alex look for the meeting invitation for the private member Q&A. You can ask your questions directly from him. By the time we do that, it’s probably going to be even more changes, so I encourage all of you to attend. If you’re not a member, and after listening to this, you think you might want to become one, go to Collective54.com, fill out an application, and we’ll get in contact with you. But until next time, I wish you all the best of luck as you try to grow, scale, and someday exit your firm.

Note: This transcript was generated by Zoom.