How AI in Family Law Cut Our QA Cost to $600/Month

AI in family law isn't a tool you buy in the US—it's a system you build, and most firms are stuck in first gear.

Because of that system, Sterling replaced a $2,500-per-person offshore QA team with an AI system scoring 200 calls a day for only $600/month.

If you’re still figuring out how to adopt AI effectively, this episode is where your path forward starts to get clearer.

—

📲 Subscribe Now: https://www.youtube.com/@karls.anthony

                             https://www.youtube.com/@TylerxDolph

📚 Order The Waterfall Method Book here: http://rocketclicks.com/the-waterfall-method/ 

📝 Schedule a FREE Family Law Firm Audit: https://rocketclicks.com/schedule-a-family-law-quick-audit/

---

📄 CHAPTERS  

0:00 - AI in Family Law: Why It's a Factory You Build, Not a Tool You Buy

2:21 - The State of AI Adoption in Family Law Firms (Only 20% Have Moved)

4:27 - Two Villains: Owner-Dependent Operations and Duct-Taped AI Tools

5:52 - Enterprise vs. Consumer AI Accounts and Legal Data Retention Rules

10:36 - Building the AI Factory: From Prompting to Looping Skills

14:10 - The $600 AI Intake QA System That Scores Every Call

19:40 - Why AI Adoption Threatens the Hourly Billing Model

Curious to discover your personalized roadmap to scaling a law firm, no matter where you are in the business? 

Follow these steps:  

1. BOOK A FREE 30-MINUTE AUDIT WITH US:  https://rocketclicks.com/schedule-a-family-law-quick-audit/ 

2. CONNECT WITH US: 

LinkedIn: https://www.linkedin.com/in/anthonykarls/  

                     https://www.linkedin.com/in/tylerdolph/ 

                 https://www.linkedin.com/company/rocket-clicks       

Facebook: http://facebook.com/2311.karls.anthony  

                    https://www.facebook.com/profile.php?id=61577369996484

                        http://facebook.com/RocketClicks   

Instagram: http://instagram.com/karls.anthony  

                     https://www.instagram.com/tylerxdolph/ 

                         https://www.instagram.com/...

Transcript

Copy

0:00

Everything you've heard about AI for family law is probably focused on the wrong thing, but keep telling you to buy tools and learn tricks that do one thing, but are divorced from your firm's operation. AI is not something you just buy. It's a factory that you build. Here is where that journey starts. Welcome to the Sterling Family Law show. The show for family law firm owners who want to grow their profits, multiply their cases, and set a clear roadmap on what they need to do to build the firm of their dreams. Listen, this episode is an important one. It breaks down proven business growth, digital marketing, and revenue for strategies to help you scale smarter and not just be busier. I am Tyler Dolph, the CEO of Rocket Clicks, which is a family law only business growth consultancy that helps family law firms increase profits through full service development. Today we have the co-founder of our sister business, Sterling Lawyers. Tony Karl's with us to talk about the state of AI in family law right now and the two villains, which is to do everything trap and the duct taped AI trap, and how to cut through all of the hype around AI by envisioning it beyond a tool you use. But as a factory you build, All right, Tony, welcome back to the show. Big episode today. There's been so many, so many podcasts on AI. You know, obviously it's a very hot topic right now. And our approach to this, we want to take a bit of a different approach. We're not going to talk to you about what AI is or what a chatbot does. We want to get into the weeds a little bit and talk about how we are using this at Sterling, and how family law firms should think about implementing AI in their firms. The first thing we want to talk about is the state of AI in family law. So give us a little bit of like just background on what you're seeing in the space and kind of how firms are adopting it today. So I feel like we're we have a pretty low adoption rate right now. If you look at like adoption rate curves that you typically see, we're kind of fast past the first like early adopter phase, which is usually 6% of the population. We have about 20% of family law firms that have adopted, adopted using AI in their firm, and those that are really holding a real edge, they're seeing they're seeing productivity go up. They're seeing their their overall revenue increase. They're able to produce more work in, in their firm for the same amount of time that they're spending because they're they're delegating some of the work, some of the admin work correctly to AI. Austin Sterling, we've we've implemented a whole bunch of different things to give us different edges. A lot of it's a lot of it's pointed in the sales arena currently in terms of how we do our sets and do our shows and do our follow up processes so that we can generate more for the same spend that we're currently spending on marketing place. So we're in the early phases, and those that are adopting are winning, and those that are kind of holding, holding to the old world are experiencing a lot of what I would imagine the people that were in, you know, the industrial revolution in the early 1900s, they were feeling when, you know, the model T came out and like that, you know, the assembly line factory manufacturing came out and just flawed at how fast things were happening, comparatively to doing it the old way. And I think we're seeing that's that's being seen across a lot of industries. So there's people that are kind of putting their heads in the sand or they're pushing on the brakes and they're losing the opportunity to gain edges. When everything's new. And same things happening in the in the family law space, which isn't a surprise. It's how, you know, this industry is like every other industry. Yes. And I think it's understandable right in the law. Space to be skeptical. You know, there's there's not specific laws in every single county about how to use AI or what AI can do for a firm. But I think we've kind of boiled down to two main kind of issues or, or enemies that we're seeing in firms. And the first is the firm's operations still revolve around the owner. So they're doing all of the things, and they care deeply about their clients, but they don't have the time to invest. And then the second is they misunderstand kind of what AI actually is. Talk to us about that a little bit. Yeah. I mean, I would say the on the, on the first one, there's because the, the work isn't over. There's not an emphasis on delegation and elevation. In a lot of firms the adoption is slower because everybody's on their own little island and nobody's working together to kind of grow up out of that. So what ends up happening is if you can start leveraging AI and building skills for particular use cases, that owner or the main linchpin within the firm, they're delegating their internal knowledge on how they do things into essentially an SOP for an AI to run. And it doesn't have to be legal And I think some of the fear is the lack, the lack of clear guidance available in every state. But there is good guidance from the ABA, and there are different decisions that have come out in different states. Illinois has one that's really strong. Texas does, Florida does, New York does, California does. And they're all they're all similar in different ways. They obviously have different edges for their particular jurisdictions. But. Getting in the weeds and understanding them and then understanding how you can leverage AI is really important. And that's going to give you an opportunity that most, most firms aren't taking advantage of. At this point, it's still 8,080% still or not. I think a lot of that has to do with the misunderstanding of how it works. There is we're going to talk about this in our upcoming webinar, but there are very specific. Rules as it pertains to data retention in, in, in the legal space generally. And if you're not using the right types of accounts, and we've seen this in our own firm, and I know some of our clients are operating without full, full scope of investigation in terms of how they're adopting AI in their own firms is they're really not digging into kind of the data retention policies with the products that they're choosing to use through AI. There's really, really falls into two, two lanes. There's consumer grade and then there's commercial. And those are two very different terms of service. And if you're not understanding what those are, you're going to you're going to put yourself and your clients in a position where you might be crossing violating ethics, ethics rules. that's an important point Yeah, it's like one of the most it's what's unlocked. The unlocked AI usage for our team is are we fully invested several months into investigating that and it and then teaching our team about it and how these how these models work and what data is used and what data is not used in both scenarios. So the consumer grade account, you know, if you look at, you know, OpenAI or ChatGPT is it's mostly known or anthropic, also known as Claude. There's there's two types of accounts. So you have your consumer grade account, which is you can go and you can sign up with just your email and use it for free. You can get a plus version or a max version and get a little more usage out of it. All of that's consumer grade, and it's pretty ambiguous on some of the paid versions, but it's the same terms of service where they're using that data to train their models. So if you're using that, you're like, you're very likely crossing ethics rules. There's there's likely an ethics rule violation that you may be in infringing on, on the enterprise grade accounts specifically to anthropic. Also with the OpenAI, if you enter a enterprise account in terms of service, that that data is now yours. That's their two very different terms of service. They're two very different contracts that you're signing. And they dictate how the data that you're the data that you are putting into the AI models, how it can be used. With the company. So can they use it for data training and modeling or can they not? So it's really important. And like that's what's misunderstood. And just digging into the digging into that is really important. And usually the and what we've seen is that's the unlock is like attorneys learning how how that's being used and kind of digging into it for themselves and understanding how that comports or doesn't comport with their specific jurisdictions. So. to do is if you're using AI, your law firm, ensure that it's on an enterprise account. Yeah. Make sure you are using a business, a business account. So because there the terms of service are written differently for you have more controls, more access. And it's built for enterprises. In these frontier model companies know that if they're going to have business and, you know, make their stock dreams come true, they're going to need to give the ability for companies to use these models with their own data retention policies being in place, versus being just open to everything being used for model for model enhancement. Okay. So we've covered the basics. You need to use an enterprise account. Now I know a lot of firms are just using the chat function, right. Maybe they're using it to like edit documents or check the the content of an email. I think where we've really been successful at Sterling and at rocket Clicks as well, is building what we're calling the AI factory. Tell us about the difference between just using the chat function of AI versus building in something like VS code and creating an actual factory. That's that's creating processes and procedures. So most people that are using AI are just using it to prompt in a browser. And like you're definitely getting productivity out of that, but you're really scratching the surface. You're using like 5 to 10% of its potential. It's like you're driving a. Super high end sports car and you'd never switch it out of first gear. It's just always stuck in first gear. You never get. You never actually get the benefit. Like you're you got to or you're driving a golf cart and it's just got like somewhere regulator on it where it only goes five miles an hour. That's basically how most people are using AI today, and that's through just general prompting and that's useful. It's just a different way to interact with the internet. It's a smarter way to interact with the internet. You get more out of it. So it's not bad. It's just it's just it's like the super basic version. You can progress up this chain of thinking in terms of how you leverage it. So the step up from just prompting is using skills. So a skill is essentially an SOP. If you can teach a human how to do something, you can teach AI how to do something. And it's just a protocol of here's here's a scenario. When I tell you this scenario, this is how I want you to interact with this scenario. And here's what I want back. It's very it's a skill. If this then that it's a it's a basic skill. Then there's a version where you have a little more advanced version of that skill, which is a skill plus tool connections. So you can give it access to different repositories of information like a Google Drive or like an email, like an email inbox or whatever, Salesforce, law, matic, whatever tool. Like there's tons that they're called. There's tons of ways to connect to these different. Data, data repositories where you likely have important information that if AI had access to and it had a skill to use it, it could do even more cool things for you. So that would be like the second or the third level. And then fourth would be you. You start start using it really specifically for particular projects. So this is really pertinent in the family law space because all of our clients are projects in and of themselves. Every client is its own project, and you're going to use a lot of skills and the same tools for that project. But now every time you interact with that project, it's retaining the context of the whole conversation, and the next thing that you're asking it to do has the whole history of what you've already discussed and how how you're how you're thinking about that particular client and so on and so forth. So it's a bigger learning. So and it continues to go up until you get to what I would call a factory, which is it is a whole bunch of skills that can loop on themselves. So what a loop essentially is, is it allows you to create one prompt and then it'll it'll orchestrate a whole bunch of different skills for you, one after the other, until you get to till you get an end result. So, you know, we've we've done a lot with a bunch of different things to enable us to like, fully leverage AI. So an example that we have is we used to have roles at our firm that shade did quality assurance on all of our on on a subset of calls for our intake agents. So we had about 1314 intake agents on our team. They all answered between 10 and 2025 phone calls a day, make 50 to 70 outbound phone calls a day, and then our QA team would score per week, ten inbound, ten outbound per agent, and then we'd use that for coaching. Now we get to score every single phone call because we have AI in the loop. So what it what it can do is we built an AI essentially factory for this, for this system where when a call ends that that gets passed to an AI tool that does the transcription and tone analysis, and then it gets classified as a certain type of call, that call then gets scored against a one of 43 different rubrics, and then that gets logged into a database. And then every morning we get a new report for all calls over the last seven days for every agent. So we can see how are they performing on on their total totality of calls. So and we found some very interesting things in terms of like complying with our own rules where where different agents are struggling with different types of interactions. And it's enhanced our ability to improve our team's performance, because now we don't have those two QA agents. We now have a dedicated person that's coaching, and they have tons of information, more information than they could ever want to coach. Our intake teams or intake team is getting better at a faster rate than they were in the past, and this program costs us about $600 per month to run. And it costs us at, you know, at least $2,500 per person for our offshore QA team in the past. So we're we're net positive cost. We have way more data and it's way more meaningful and impactful. And the speed of the speed of response is faster than anything we've had in the past. Like that's an example of something we've been able to build that's added, that's add value, it's reduced costs and it's it's produced higher results. And I think if I, if I think about just that one process right. The one on one version is hire people to listen to these calls and score them. 201 version is I'm going to take the recording and I'm going to put it into AI, and I'm going to tell AI, hey, here's the rubric score against that. And then the master's level version is what we're doing is the process is built, it happens automatically. And we get a report automatically generated and delivered to the coach so that they can get the data and deliver the coaching. Yeah. So that gets delivered at 7 a.m. every morning. And Sydney preps what she's going to talk about and huddle with the intake team at 8 a.m. based on what we got in the report. So because she she she does training 15 minutes of training every single day with our intake team. So now they have it's real stuff, real time real fast. It's so great. I mean, like what an evolution. And it happened so fast. And I think the validation that we got was this recent report from Clio stating that AI adopting firms are growing at about four times the rate of firms that are not adopting AI. And so it's like everyone's excited about this thing. But I think it's important to realize that this isn't going away and that you have to get on board and learn how to build this factory. Right. And I know you've attended conferences and they're saying the same things. Is that correct? Yeah yeah for sure. Those that are adopting it where they're seeing significant enhancements and like our whole premise is we're not going to use AI to replace people. We're going to use AI to enhance what we're already doing. So, you know, the people that we moved off our QA team, they're now on our intake team. And now nobody's doing nobody's doing the laborious work of sitting there listening to a 20 minute call, checking the boxes in in a spreadsheet and doing it super slow. Now everybody's getting the scores. We're doing the same thing with all the admin work that our intake team has to do, and there's tons of it because we get, you know, three 300 ish text messages per day. We get 200 phone calls per day, we get our public office inbox, email gets, I don't know, 70 to 80 inquiries a day of just random. You know, all of it's not quality, but all of that has to be sifted. So we used to have two people that did the sifting and just to hand it off to somebody else to do something with it. Now that's just happening in real time automatically and like it's happening at a higher rate with better quality. There's still exceptions. You still got to manage it. But now the team can focus on the conversation and the relationships with the clients, which is what we're in business to do anyway. It's like, how do we empower them with these tools so that they can have more meaningful interactions and better interactions. And like, they're they're improved from an ability perspective. And that's kind of how we're thinking about leveraging. Tony I love like the I love the productivity hack. But I think the one call out that maybe people are afraid to talk about is the fact that as AI adoption continues to increase, especially with consumers, they're going to start to question this hourly billing model. So going to be like, wait, why did you take three hours to produce this document? I could have done it myself in five minutes or whatever. So and for our audience who may not know this, Sterling operates on a flat fee model. And so we're incentivized to be as productive and efficient as possible. Right. So AI was a natural evolution for us. For the hourly firms. This is going to be a bit more difficult. What's your perspective on how that's all going to play out. Yeah I would I would. This is my this is my speculation. But I believe it's I believe it's going to be true, that I do think AI is going to very much disrupt the hourly billing model because not because it has to, but because it's going consumers are going to start making that challenge. They're going to everybody's going to get used to knowing AI is involved with business. And then the question of what am I paying for is going to get more intense. And the benefit of so we don't do flat fee, we do fixed fee. So the slight variance there, the way we look at how we do fixed fee is the best corollary is like an underwriting process. It's not underwriting, but it is. There's some similarities and obvious obvious differences. But when a client comes in, we do a huge vetting process over a whole bunch of different variables that might exist in their case. And then we tell them, based on an uncontested route, negotiated a route and a route that gets contested that progression progressive path over three stages, how much it's going to cost them in totality. And then we have to prove that we did the work that we said we were going to do, but that's how we communicate with the client. Here's all the things that we're going to do for you based on your scenario. And then we go to work. We don't have that hourly billing challenge. It doesn't matter how long it takes us. That's always been our premise is like we're going to get we are going to put ourselves out there to say, here's how much this particular thing is going to cost. And then our job is to effectively do it, which is in the direct opposition of the hourly model. So because we don't get to we don't get to skimp on quality, because if we if we do, we're going to get malpractice claims like this is the perfect industry to do this in, because if you screw around and don't do a good job, you're going to lose your law license. So a fixed fee model really puts the onus on the firm to identify where the opportunities for me to delegate and elevate and use my team as best as possible, and to create operational efficiencies throughout the process of a divorce, for instance, so that we can make profit. AI is just going to enhance that for us. All it's going to do is make that make that whole premise better. We're my perspective. Our perspective at Sterling is like, this is going to really inhibit those in an hourly basis. Because now, now the question of what am I paying for? Why is this taking so long? All of those things that's really going to become more front and center, because everybody's going to start expecting AI to be involved, and that's going to create this like weird scenario where why aren't you? Why aren't you using AI, and why am I spending all this extra money? Or it's going to create all these uncomfortable conversations that are going to be hard for you to honestly answer with authenticity versus just a nice good response. So I do believe it's going to be important. We're not there yet because overall adoption of AI is pretty limited. You know, I think the joke currently is most people are using AI to create funny pictures of their mother in law or father in law for their kids to color. But the reality is like this is coming as a significant impact on and like those that are, they're seeing the benefits already. It's definitely having a force. Force multiplication impact. think about it. And you know I, I, I felt like, you know, today we talked a little bit high level. We got in the weeds a little bit. We are going to be doing a whole series on AI for family law firms. Our next episode, we're going to tackle AI policies and show you how to protect your license, your reputation and your client data while still adopting AI. We touched on that a little bit today, but we're going to go deep into that topic in our next episode. Tony, always beneficial. Appreciate your value and your time, and we'll look forward to seeing you again very soon.