Some guy from out of town.

Last week’s question: partners and budgeting

The Sidebar question asked about partners and budgeting, and the headline that came out of the replies surprised all three hosts: AI is forcing hardware upgrades firms did not plan for. Rob’s read is that AI exposes old infrastructure, so CPU and server upgrades jump the queue ahead of what the management committee planned, especially for firms still running on premises.

Hardware for agents, not for chat

Debbie has never had a client tell her that AI forced a hardware decision. She can see it for a firm building its own AI or running a local language model, or for firms whose legacy servers in the closet still turn on but do not connect to anything new. Amanda’s firm had the server space, but it is choosing to run its agents on the server rather than on desktops, including a server set up for Claude Cowork, so people are not burning their own machines’ processing power.

By the numbers: people, tech, and health insurance

In the replies, 78 percent said people costs drive most of their budget increase, 67 percent said AI and tech costs are rising sharply, and 56 percent named health insurance as their top line item. Amanda, back from the Managing Partner Forum Women’s Symposium, says firms across the country are seeing the same thing: the highest renewal quote she has received in 22 years. The money comes out of either employees’ pockets or the firm’s, and she would not be surprised if rates keep rising.

Why marketing is always the easy cut

The replies show marketing absorbing some of the increase. Debbie’s explanation: firms are bad at measuring whether the money they spend turns into business, like the golf tournament sponsored at 25,000 dollars a year with no idea whether it is worth it. That is not a failure of the CMO, who is trying to get lawyers to report what happened at events. Rob adds that firms often hire one junior marketer and expect strategy and execution, and Amanda points out that the ROI problem applies just as well to AI, tech, and people.

From the symposium: recruiting, profitability, comp

Amanda brings back the topics managing partners kept returning to: recruiting and retention, with large firms using what one attendee called predatory recruiting to pull talent out of mid-size firms; profitability squeezed by tech spending and compensation; and the comp model changes smaller firms need to stay competitive. She spoke on panels about profitability, compensation models, and trends, including how AI is changing which first-year attorneys firms need.

Who should own AI inside the firm

One member reported getting half an FTE focused on AI work from their IT provider. Debbie says that role is hard to hire for and easy to lose, since good people get offers elsewhere. Firms are leaning on MSPs that have to earn their keep every 30 days, and an innovation hire brought in house slowly loses the outside view of what other firms are doing well and badly. As the saying she credits to Mark Twain goes, an expert is some guy from out of town.

Five archetypes every AI project needs

Debbie’s firm thinks about AI work through five archetypes: an advisor who can command the room, a legal process expert who knows how the work gets done, a translator who turns that into a tool or a build, a builder who works in the interface, and a deep builder comfortable with development and connectors. A project needs at least the first four, though one person can play more than one role. She compares it to building a house, with an architect, an electrician, and an interior designer.

Protect the person driving change

Rob encourages firms to hire younger people who solve problems with AI quickly, paired with someone who knows the firm’s process, and to let them move. Whoever drives the change, an administrator, COO, or partner, needs protecting because they will be disruptive. Debbie warns that at many firms someone owns AI by default without being the right person. Rob suggests click-through mock-ups as a way to see around corners and get buy-in before anyone builds, and Amanda closes with a recruiter who rewrote a candidate’s resume into the job description with ChatGPT.

Transcript

October 8th, and where in the world is Debbie

Rob: Welcome to Sidebar Live, October 8th edition.

Amanda: This isn’t our first October, you know.

Rob: No, it’s not. I think last week was our second. Still, I’m getting adjusted to the fall, because in Texas we just dropped out of the hundreds like a week ago. It’s been wild. It’s just starting to feel like fall.

Amanda: Yeah, yeah.

Rob: Amanda, you’re at home. You’ve been in Chicago, I know. Debbie, where are you?

Debbie: I’m in Atlanta. It’s nice. Not too far from home.

Amanda: Debbie could have a Where in the World is Debbie Foster, and it would be an entire show.

Rob: It would be. I ask her where she is. That’s my icebreaker every time.

Debbie: And I’m always somewhere different.

Rob: You are. You are. And many times good places. But, okay, they don’t care about this, so we can save all this for our private conversation. Let’s jump into Sidebar this week. Amanda, do you remember what the question was last week?

Last week’s Sidebar question: partners and budgeting

Amanda: Partners and budgeting.

Rob: It was. Okay, so the headline that came out, I’m going to read some information. This headline shocked me, so I cannot wait to hear this.

AI is forcing hardware upgrades

Amanda: Well, I hope it’s the same one.

Rob: AI is forcing hardware upgrades firms didn’t plan for.

Amanda: Is that the same one? That’s the face I made too, Debbie. And I was hoping that Rob, as our tech genius, could tell me why people are having to buy hardware for AI. I don’t even know what that means.

Rob: AI is exposing old infrastructure, and CPU and server upgrades are jumping the queue ahead of anything the management committee planned. So I think it relates to firms that are maybe on-prem. Maybe, I don’t know, they’re being forced off the servers onto newer technology, or what they’re using now just isn’t compatible with where they want to go. And I think maybe it aligns with what Debbie’s always saying: before you put in any AI initiative, you’ve got to make sure your data is in order. And if you’ve got a bunch of stuff on old servers, you’ve got to make sure you’re in the right setup, maybe. I don’t know. What do you think, Debbie?

Nobody has ever said that to Debbie

Debbie: I’m sitting here thinking about whether anyone has ever said those words to me, that because of AI they have to make new hardware decisions. And I can’t think of anyone who has ever said that to me.

Debbie: Sorry, we had a whole big meeting here, and a bunch of people just walked by. Like ten people.

Rob: We’re all human.

Debbie: So I suppose if you’re talking about a law firm that wants to do something along the lines of what Kirkland and Ellis was doing, like they’re going to start building their own AI, or this local language model thing that I like, maybe then they have to. But also it might just be that firms have been using this old legacy stuff for so long, and they’ve just been letting the things in that closet over there exist because they still turn on and they still run. And all of a sudden they have to, I don’t know how connected those things are, right?

Running agents off the server

Amanda: Yeah. Well, one of the things we’ve had to find: it did not require us to upgrade, because we had plenty of server space, but we are going to run our agents off the server versus running our agents locally. So we do have to find some space for agents to run. And so if people are tapped out, or they have old servers, I could see needing it, if they’re being recommended to run agents on a server instead of locally. We’ll have to have somewhere to store all of it, versus having it stored on local desktops, and we need the processing power. For example, we have a server set up for Claude Cowork to go through, so that people aren’t using their own computers and their own processing.

Rob’s new laptop

Rob: Got it. So that’s possible. That’s what I was maybe thinking. We’d love to have somebody jump in the chat and explain this. But on a more basic level, I just upgraded laptops.

Amanda: I was like, don’t you have three laptops running right now?

Rob: I have three screens, two laptops.

Amanda: Yeah, so why did you upgrade your laptop?

Rob: I needed more power.

Debbie: Maybe a bigger ego, too?

Rob: Like, I wanted, no.

Debbie: He wanted it bigger.

Amanda: He wanted it bigger.

Rob: And I wanted some separation of church and state. So I got a new laptop. I tried to buy a Mac mini first, and you know, that’s the hot thing, but unfortunately those are hard to get. So I had to go the laptop route, which was interesting, but I digress. So let’s keep going down this. I think it’s interesting.

By the numbers: people, tech, health insurance

Rob: By the numbers: 78 percent say people costs drive most of their budget increase. 67 percent say AI and tech is arriving sharply. And the third, 56 percent name health insurance as their top line item.

The steepest health insurance renewal in 22 years

Amanda: Listen, I spent this past weekend at the Managing Partner Forum Women’s Symposium, and we were all comparing what health insurance is doing to mid-size firms across the country. It is staggering. I’ve been doing this 22 years. I’ve never received a health insurance renewal quote as high as the one that’s sitting in my inbox right now. Staggering.

Rob: Any reason why?

Amanda: State of health care in the United States in general. It’s all I can tell you. But yes, it’s high. So it’s going to drive so many decisions, because it’s such a huge percentage increase that the money has to come from somewhere else. And it either has to come out of employees’ pockets or it has to come out of firms’ pockets, and a lot of firms like ours value our employees and don’t want to hit their personal resources if we can help it at all. I don’t know that it’s going to set us back on tech projects, and I don’t think that’s what I heard from the group. But I wouldn’t be surprised if rates keep going up, because you’ve got to get it somewhere.

Marketing absorbs the cost

Rob: There was an interesting, I was going to ask this question, that came from that. There was some data in here that marketing is actually absorbing some of the cost increase. So that’s where it’s coming from.

Amanda: Like, reducing marketing?

Rob: Yes. Is that something?

Amanda: Yeah. It’s always the low-hanging fruit, unfortunately, of law firms. We could get a bunch of CMOs on here to tell you how sad they are about that.

Nobody can measure the golf tournament

Debbie: Yeah. Well, and I think that part of the reason why that’s the low-hanging fruit, and it’s easy for people to go there, is because generally speaking, law firms are really bad at measuring if the money they’re spending is actually resulting in increased business, or new relationships, or things that will eventually lead to business down the road. So we’ve sponsored that golf tournament for 25,000 dollars a year, every single year. We have no idea if it’s worth it or not worth it, but we keep doing it, right?

Debbie: I think those are the conversations that make us stop and say, we’re not saying we don’t want to spend 25,000 dollars. We’re saying, does the 25,000 dollars we spend have an ROI? And if not, can we not spend that? I think that’s why marketing is such an easy target. And that’s not, for any CMOs listening, a referendum on the CMO, right? Because the CMO is trying their darndest to get the lawyers to contribute information about marketing activities. Who attended? What conversations did you have? Was it worth it? They’re trying to get that information, but that’s really hard information to get from the lawyers who attend the events.

Firms hire marketing wrong

Rob: I think, too, a hot take: I think a lot of law firms hire marketing wrong as well. They don’t see the value in it, so they bring in a lower-level person who’s less experienced, and they expect them to do the whole world, right? They want strategy, execution, and everything in between. And it’s like going to most attorneys and saying, hey, I want you to practice every different type of case at the same time, right? That’s hard to do, and they don’t necessarily understand that.

Swap marketing for AI, tech, or people

Amanda: But can I also say, Debbie, you could have stopped that sentence where you said law firms are bad at determining the ROI of, and you could have cut off marketing and added AI, or tech, or people. You could have added any other noun at the end of that. Because that’s also what I heard this week at this managing partner group. I had one person sitting right next to me who goes, yeah, I just bought a ton of AI products, and now no one’s using them. And I’m like, but I can tell you why.

Rob: Amanda, I have a list of questions, but you keep referring to this.

Amanda: I’m sorry. Go through your questions. I’ll shut up.

Rob: No, no, no. I’m saying, what else did you learn?

Recruiting and retention at the symposium

Amanda: That recruitment and retention are as hot topics as any other. That small and mid-size firms are feeling the weight of pressure from large firms, and that’s everything from coming straight out of school all the way to, as one woman put it, almost predatory recruiting strategies to pull talent out of mid-size firms. That was a really hot topic. Profitability, and the squeeze on profitability that’s coming from tech spending and compensation.

Rob: I mean, on the recruiting thing, that’s not new. Or is it?

Amanda: It’s not. But it just seems to be harder right now.

Rob: Yeah. Interesting. Sorry, keep going. What were the others?

Compensation models and three panels

Amanda: Comp models. I mean, we’re always all talking about comp models, but the necessary changes to comp models that small and mid-size firms are facing to be competitive with large firms.

Rob: What are they doing differently? I’m trying to figure out if somebody else is doing it differently.

Amanda: I spoke on three panels: a panel on profitability, a panel on comp models, and then a panel on trends in law firms. One of the trends I talked about was hiring, and how AI is affecting hiring in the market and the needs of firms, and the first-year attorneys that we may or may not need, and what that looks like.

Half an FTE for AI from the IT provider

Rob: Are you hiring different people? I hear you on the attorney side, but from a staff side. You know, somebody in here, and I don’t want to look through, but they’re essentially getting a 0.5 FTE that focuses on AI work within the firm from their IT provider. So they’re going to have somebody doing something like that. Are you hearing more firms doing that? Either of you? Debbie, what do you think?

Amanda: I mean, Debbie’s probably heard some of that, too.

A hard role to hire and easy to lose

Debbie: Yeah. I mean, I think that is a hard position to hire for. And if you get someone really good, they will be in high demand, and there will always be an opportunity for them to leave and go make more money somewhere else, and lots of people are motivated by money. So from a consulting perspective, for us, that’s a huge opportunity, right? Now the issue just shifts from them to me. How do I make sure that we have the resources that can deliver those services?

Debbie: But some of the value, even when we work with firms thinking about what their strategy should be around Amanda’s favorite topic, how do you staff IT, right? It’s a hard challenge. So many firms are leaning into a strong MSP, someone who has to earn their keep every 30 days, that is one step removed from being Joe who’s worked here for 174 years kind of thing. So they can really think about the kind of service they want from their MSP.

Why the outside view matters

Debbie: But then you start to think about the innovation side of things, and you can find an industry person who seems to be really dialed in and plugged in on innovation, and you hire them at your firm. And the way they learn, and the things they get involved in, once they get brought into the four walls of your firm, their view of the world changes a bit, because they become part of your universe instead of the big picture ecosystem. Bringing in someone who is watching what other firms are doing, what they’re doing well, what they’re doing wrong, where they’ve made mistakes, where products have really paid off, where they haven’t paid off, what real adoption actually looks like, it’s pretty valuable to have that outside perspective. That’s not to say you don’t have someone in-house, but I think you have to have a mix of all of the things.

An expert is some guy from out of town

Debbie: And I never remember to go check this, but I always have heard that Mark Twain is the person who said, what’s the definition of an expert? Some guy from out of town.

Rob: I quote Debbie Foster as saying that, because I don’t remember where it actually came from. So whenever I say it, it’s like in The Office, where it’s, Wayne Gretzky says, Michael Scott says. I always just say, Debbie Foster says.

Debbie: Somebody says. Yeah, it’s like, what’s an expert? Some guy from out of town. And I feel like we’ve all seen this. I’ve seen it at my own company, where somebody comes in and they’re like, you guys should, and people are like, we should totally do that. And I’m like, seriously, are you kidding me? But I think the way law firms are going to use experts in every aspect of how the operations of a law firm work is going to change.

Listening in at the coffee shop

Rob: You know, it’s been really interesting to me. So, for those listening, I live right next to SMU, so I’m right next to a campus. There’s a coffee shop called Foxtrot right by the school, and I love to go there and just sit and listen in.

Debbie: Kind of creepy.

Amanda: Wait a minute. You like to go to a college campus and listen to people? Okay, say more.

Rob: I’m going the other direction. We’re getting to a better place here. Okay, so I like to listen in on what problems they’re trying to solve with AI. And the reason I say this is, I always think about hiring young people, or somebody with less experience. I like somebody that has a little experience, but if you match them with the right operational person who understands the process of the firm, or a consultant, the way some of these young folks are solving these problems is pretty incredible, and they can hit the ground running a lot faster now.

Let the people who move fast move

Rob: Now, the flip side of that, and to what you were saying: law firms tend to move a little slower. Somebody that’s working in AI and enjoys that is going to lean toward wanting to move fast. So if you want to retain them, you’ve got to be willing to let them go and move and do things.

Five archetypes for AI work

Debbie: One of the things we have been talking about, when we think about AI services to our clients, is what we call five archetypes. There’s an advisor, which we define as someone who can command the room, who can get the stakeholders in a room and they will listen, and they feel like they’ve got someone in the room who can really help them figure out what’s next. And then there is a legal process expert, someone who really understands how the work gets done.

Rob: Yes.

Debbie: And then there is a translator, someone who can take what the legal process expert says and turn it into what tool can help do that, or what do I need someone to build, which then takes us to a builder. And we have actually broken out a builder from a deep builder. A builder is someone who can build in the UI, someone who can build without necessarily vibe coding and understanding how it all connects together, where a deep builder is someone who’s more comfortable with development and really understands connectors and the consequences of what you’re doing.

Debbie: So if you think about a successful AI adoption project, you might not need a deep builder, depending on what you’re doing. But if you don’t have at least the first four, an advisor, someone who can really command the room and set the direction, a legal process expert, someone who really understands how the work gets done, a translator, someone who can figure out how to take that information and translate it to a tool, and a builder, someone who can make the magic happen. And in some instances the translator and the builder are the same person. In some instances, the advisor and the legal process expert are the same person. You can play more than one role, but all four or five of those archetypes, depending on what you’re trying to do, have to be represented to get something across the finish line.

Builders, deep builders, and Claude Cowork

Rob: That’s interesting. I’m listening to your definitions and thinking a builder goes all the way through, like, Claude Cowork, right? They can set up records, they can set up recurring things in the UI. But then when you want to build a standalone application, that’s the deep builder, if I’m understanding correctly.

Build AI like you would build a house

Debbie: Yeah. Well, I’m even thinking you could take that same subset and think about building your own home, right? You need an architect who can make sure it actually all fits together correctly, then a builder, then an electrician who knows that specific component, and an interior designer who understands how it all looks pretty. So if you think about how you’re going to build AI, think about how you would build a house. And this is top of mind because I literally have people hanging drywall four rooms over from me right now.

Entry-level work is going away

Rob: Yeah, I think that makes total sense. It’ll be interesting to see how some of these roles evolve. But I highly encourage people to tap into that younger demographic that’s coming out of school. They don’t know where they can be hired these days, just because entry-level work is going away. I think there are some new opportunities.

We tried that thirteen years ago

Debbie: Well, and how do you think about bringing in those young people who don’t live and die by all the constraints that are in their way of doing something amazing, and they show up in a law firm and they’re like, how about we try this? And everyone in the law firm is like, no, we could never do that. We tried that one time thirteen years ago and it didn’t work.

Protect the person driving the change

Rob: You’ve heard it all before. We’ve talked about it, Debbie. It’s the same thing as I think we talked about a few weeks ago: when you’re changing software, or bringing AI in, you have to really protect that person, whether that’s somebody like Amanda, or the COO, the executive director, or a partner.

Debbie: That’s a fair point, because they are going to be pretty disruptive. I’ve said this to Amanda before, and this is not just patting Amanda on the back, but not everyone has the executive business leader chops that Amanda has. And there are many people inside of a law firm that are not.

Amanda: Grab that soundbite.

Seeing around corners

Debbie: Don’t worry about it, Amanda. But seriously, I work with a lot of firms, and there are other firms that have people who really understand it like Amanda. But there are so many more firms that don’t. Going back to your build-a-house analogy, a really good architect will say something like, are you sure you don’t want to put at least a powder room on that floor? Not everybody can see around corners like that. And listen, even those of us who are pretty good at it screw it up, because it’s new and we’re all kind of figuring it out.

When someone owns AI by default

Debbie: But there are some firms where, by default, someone owns AI, and they’re just not the right person to drive an initiative forward. They feel like a failure. The firm thinks, how hard can this be?

Artifacts and click-through mock-ups

Rob: I think there’s a lot of why things are failing. I think Amanda, and people who think like Amanda in Amanda’s role, have the best chance of it being successful, because they are thinking, we probably should put a bathroom on this floor. And a change I’ve seen at Centerbase is that instead of sending documents around, people send artifacts these days. It’s all moved to artifacts. Something useful for me in seeing around those corners has been doing click-through mock-ups. That’s foreign to people in law firms, but if you type in, hey, I’m trying to solve this problem, I want to build something, don’t code anything beyond a click-through mock-up, then you can click through it, show a few people in the firm, get buy-in, and change it from there. Then you can go to that deep builder, or to a consultant like Debbie who could build it for you. Maybe that helps with some of the change management.

The resume that matched the job description

Amanda: Can I give you an example? Oh, sorry, do you have another topic?

Rob: No, go ahead.

Amanda: It’s completely off topic: how not to use AI to get a job these days.

Rob: Please.

Amanda: It’s totally off topic, but this just happened and I need to share it with you two. We’re hiring a specific role, and we sent the job description to a recruiter, and she sent back a resume. We took a look at the resume and somebody goes, oh my God, this person is perfect, all the experience. I open the resume, and of course, as ChatGPT does, certain words are randomly in bold, which is one of the biggest tells. And I start reading it, and I’m like, this is my job description in a resume.

Amanda: So I get on the phone with the candidate and I said, look, not to be a pain, which I am, obviously, but would you send me your resume of what you’ve actually done, so I know what you’ve actually done? And she’s like, what are you talking about? Well, of course, the recruiter didn’t get her permission to run it through AI and then send it to me. And when she sent me back the resume, it was exactly what I was looking for, but in her own words. So we’re going to make her an offer.

Debbie: Oh, my Lord.

Amanda: Listen, a lot of recruiters modify the resumes of the people sent to them, and they get that permission, supposedly. But in this case, they said, take this job description, take this resume, and make them the same. And then didn’t even take the time to remove the random bold sentences in the bullet points.

Rob: Oh my gosh, that’s crazy.

Debbie: Make sure in your contract you have a good refund policy, or replacement policy, with that recruiter.

Amanda: True. But luckily I knew they were doing that, and I asked her to send me her own resume, and then I met with her, and she’s lovely, and we’re going to make her an offer on the job. But don’t use AI if it’s obvious. It’s AI 101 to pull out the tells.

Rob: Wow. So it says my connection’s unstable. Am I okay?

Debbie: Yeah, you are unstable.

Join Sidebar, and see you next week

Rob: I know I’m not always okay. But I put the join Sidebar link below. We’re already at 30 minutes. Make sure you join Sidebar. The community continues to grow, we continue to get some great insights, and we’re going to continue to release more. So go over there and join. It just takes two minutes: click the LinkedIn button, put your information in, answer a few questions, and Amanda might let you in.

Amanda: So I might. Go test your luck.

Rob: And we’ll be back next week, same time, for more talking and random storytelling, and sometimes we talk about the results from the week before.

Debbie: I love it. I really enjoy it.

Rob: Catch you guys later. All right, bye.