Latest / The Chief Customer Officer Podcast / EP019: The AI Native Services Playbook w/ Jay Nathan
Transcript
- Jeff | GrowthCurve: Hello, welcome back to another episode of the Chief Customer Officer Podcast. I'm Jay Nathan, your host. Jeff's out sick today, so you've got me solo, and ⁓ I'm gonna be digging into something that's been on my mind, and I'm seeing more and more of ⁓ in articles and content out there ⁓ over the past few weeks. ⁓ there's an article from Emergence Capital. They're an enterprise focused SaaS ⁓ VC firm, enterprise SaaS focused VC firm. ⁓ and the article they wrote is called the AI Native Services Playbook. And it's a pretty good ⁓ article and a lot of SaaS companies are taking it seriously right now, so I thought we'd talk about it. ⁓ I want to walk you through the main ideas from the article, tell you where I think they got it right and push back on a few things. ⁓ and I'm also gonna share an end-end example of what an AI native services company looks like in practice, at least from my perspective. The article actually didn't provide an example, and I think ⁓ the the playbook here might get a little abstract and hand wavy without it. So here's what I'll walk through. I'm gonna first define what AI native services actually means. Then I'll walk through my example of it. ⁓ hopefully to make the the concept a little more concrete. ⁓ from there, I'll give you some perspective on where I think emergence gets it right and where I think they fell short. ⁓ we'll talk about the vertical ecosystem angle that the article touches on, but doesn't really develop enough. And then I'll connect all of this to customer success because I think that's where th this concept is sort of rooted in in customer success. ⁓ at least, you know, the low what I call lowercase custom C S customer success. ⁓ So we'll come back to the example and rooted on the AI side of things, and then I'll I'll close ⁓ real quickly after that. So let's hop right in. ⁓ what is an AI native services company? Well, according to Emergence, it's a business that collapses software and services into one integrated system, delivering the full outcome the customer wants. The company owns the full stack, the technology, including the AI, the processes, and the results, the outcomes. The customer never touches a digital product. They just get work done or get the work done. ⁓ So in SaaS companies, you had to sell somebody a tool and then teach them how to use it. But in an AI native services company, you sell someone a result and then your company produces it. So that's not a new concept, right? Business process outsourcing agencies have been around for a long time where they do work on behalf of their customers. ⁓ but I think the difference here according to Emergence is that the service delivery is largely automated using AI. It's a logical evolution of services of a managed services business model, or any services business model for that fact, for that matter. And if you're running a managed services company today and you're not trying to run with AI, what do you or automate with AI, what are you doing, right? ⁓ we all have these tools available. to us now to improve efficiency and gross margin. And so especially in the services world, we should be using them. So ⁓ but here's my issue with how the article frames it up. Obviously Emergence is a VC and they spend a lot of this time a lot of time in the article applying the AI native services model to businesses that VCs can invest in. ⁓ And I think that actually distorts the way these companies really need to operate and run and I'll explain what I mean as we go along here. Before I get in the playbook itself, I want to give you an example. The article doesn't really give one, ⁓ it's sort of an end-to-end example. And I think it helps to make everything a little bit more concrete. So years ago I worked for an HR tech company. So this is sort of a natural example for me. ⁓ we had a recruiting platform. We gave customers tools to recruit their own employees. ⁓ but imagine you're building in the recruiting space today. You might do it a little differently. The the conventional SaaS play is to build recruiting software. Maybe you have recruiting software already and you're adding AI agents to that software that can screen resumes and schedule interviews and generate job descriptions. This is exactly what a lot of existing companies are doing today. They're adding AI features to an existing SaaS platform that they are still expecting the customer to use. But in an AI native ⁓ services firm, it sort of flips it on its head. So instead of selling recruiting software, you're gonna become a recruiting firm, an AI native recruiting firm. You're not selling the tools anymore, you're taking on the recruiting work for your customer. So, what does that look like in practice? ⁓ a company hires you or your company to run part or all of their full cycle recruiting operations. And your company uses AI to source candidates, ⁓ to screen resumes, to score them, flag the top candidates, draft outreach, emails, personalize, ⁓ set screenings. A human on your team would normally have had to handle all of those activities. ⁓ and the humans on your team in the services company will still handle the final evaluation, relationship management with the candidate, coordination with the client, but The client doesn't actually ever log into the platform other than maybe a dashboard. ⁓ they might tell you, hey, I need to hire three senior engineers in Austin, Texas by the end of Q three, and you will come back with ⁓ interview reggae candidates and ideally people who have accepted jobs after the screening process and and and ⁓ sourcing process has taken place. So obviously your pricing model is not a per se pricing model anymore. It's per placement or per outcome. per qualified candidate delivered, per role closed. ⁓ and as the technology that you use to build that process and deliver those outcomes gets better, then that's where the margin expansion comes from for your services company. So just for for you know for reference, most professional services orgs run between 40 and 60% gross margin. ⁓ Most software companies run between 75 and 90% gross margin. So we've enjoyed those high gross margins in SaaS for a very long time. ⁓ so now if we're going to be delivering services, we have to find a way, services instead of or outcomes instead of software, we really have to find a way to ⁓ drive the gross margin way up in these businesses. And the, you know, auto automation capability with AI is is really the way to do that. So what does this require on the operation side? ⁓ first you got to have a team that understands recruiting deeply, right? Not necessarily AI deeply, but recruiting. So this is domain expertise, followed closely by data and AI based process automation expertise. So you actually do need both, okay? ⁓ you need proprietary data from every engagement that you've had with a prospect feeding back into the model. You need workflows that separate what AI does from what humans do, right? Human in the loop. And you need to be very clear about which is which on any given day and continuously iterate the the process and automate pieces of it as you go. You need to be re embedded as the as a company in the sort of the recruiting ecosystem. This is no different than how we've gone to market in the past with any business, right? You form channel partnerships with ⁓ ATS platforms like Greenhouse or Lever. ⁓ you build relationships with HR leaders and hiring managers who trust you. So that's what an AI native recruiting firm actually looks like. And it's fundamentally different than selling a recruiting SaaS tool. Okay. I'll come back to this example a few more times as we go through the the playbook, but let me walk through a few areas where I think ⁓ the emergence playbook sort of gets it right. ⁓ so first is they they talk about domain credibility and domain just meaning the area where they do work, the industry domain, the vertical domain. The article says domain credit credibility is critical to the business model. I totally agree. Buyers evaluate services by the reputation and perceived expertise of the people providing them. And if you don't have that, you start with no trust every time you're trying to sell. ⁓ so I think this is a no-brainer, right? This is sort of brand in in a way. ⁓ in the recruiting example, if you, you know, if you have a a couple of engineers with a clever AI model and no real recruiting background, you're not going to have that trust. ⁓ it to to go to an HR leader and say, hey, we're gonna handle your HR, your recruiting pipeline. Okay. So credibility still has to come first. And by the way, that's not too different from even SaaS companies today. ⁓ the credibility comes from founders who were probably recruiters in a prior life, maybe right, not even engineers. They were actually recruiters, or you hire that kind of expertise early and then you market it. Either way, you've got to compile win stories ⁓ after you start leveraging those people to get into your first accounts. Okay, so you've got to have domain credibility to sell these types of services. The second thing that I I think they get right is something they called Mirage Product Market Fit. ⁓ they define Mirage PMF as the illusion of product market fit created by revenue growth that's powered by human labor rather than AI leverage. ⁓ this is a trap that catches people. It's it's pretty easy to build the services business, grow it. I say it's easy, it's a lot of hard work, but it can be done. It's got a pretty high percentage of success if you really lean into it. ⁓ you can convince yourself if you're driving top line revenue and you're calling yourself an AI native services company that ⁓ that you are that. ⁓ but the the the test of that over time is gross margin. If gross margin isn't expanding as revenue grows, AI isn't doing the work. Automation isn't doing the work. You're just hiring more people, right? So you can also look at your headcount. Is your headcount growing linearly with ⁓ with the revenue growth for the business? You want to test and validate the feasibility of this kind of model before you launch. So in our in the recruiting firm example, that means proving that. AI can actually source and screen candidates at scale before you go sell that capability to a customer. You don't want to get into it and figure out, ⁓ man, there's just parts of this that cannot be scaled by AI, and they're critical to the gross margin profile of the business. Okay. So gross margin, not revenue being the second thing. ⁓ and the product market fit mirage, I should say. The third piece that they that they get right is outcome-based pricing. We're hearing we're hearing a ton about this ⁓ today. They make the point that AI native services companies are uniquely positioned for this because there's no attribution problem, right? It's so hard to attribute outcomes to software products today because there are there are people that have to deliver them. Some most of the time, those are employees of the companies that we're selling it to. So they get to take credit for a lot of the outcomes. ⁓ but in an AI native services business, like a recruiting firm, we get to take credit for the outcomes, which is Candidates placed and folks that or or candidates sourced, whatever the the whatever the customer is buying from us. And we get to deliver the outcome, the service, and price it that way. Okay. So most services businesses are still in this labor pricing model. I got in a I got a bill for my accountant the other day for $300 for a quarterly meeting with him for an hour, right? He is losing money on that arrangement. Because he could charge me so much more. Don't don't tell him I said this, but he could charge me so much more. ⁓ if he really thought about what he's charging me for is his thirty or thirty-five years of experience, not the hour that he gave me. Okay. Almost all services companies are still pricing this way. And that's going to continue to be a challenge, ⁓ especially as AI does more and more of the work behind the scenes for these businesses. ⁓ So the companies who figure out how to price on results while using AI to deliver them are going to have architecturally better unit economics over time. Okay. So those are some things that I really like about the Emergence article. ⁓ a few things that I want to push back on. So the first issue that I have is that this article is really written to make From the perspective of making these businesses fundable by venture capital investors, Emergence is a VC firm. And VCs play a very specific game. Okay. They make dozens and dozens of investments and hope that a handful or even just a couple of them return outsize 100, 200, 500X return on their invested capital. So in my opinion, that view, that funding model view of a services. firm distorts this model. ⁓ and I think as a side note, VC should really be going back to investing in frontier technologies, not new businesses built on existing models and technologies where they're essentially buying and subsidizing market share. But that's a a whole different podcast that we can we can record some other time. ⁓ the article talks a lot about VC type metrics, gross margin expansion, revenue per employee, nonlinear scaling. Those are actually legitimate things to care about. We just talked about them. But I'd argue that these businesses are not even that capital intensive to launch. They shouldn't be. ⁓ most of the people building in this space don't need VC. And so the VC framing might push these companies to do unnatural things to get a business off the ground. ⁓ you know we run we run services business businesses and they are cash flowing businesses. They're not We're not building them necessarily for the enterprise value at exit. We're building them for the cash flows that they're going generate ⁓ every day, every month, every quarter. ⁓ so just interesting perspective there. The best AI native services companies are going to be cash flowing businesses. And if you have genuine domain expertise in a really focused ICP, ⁓ ideal client profile, and leverage from tech technology and AI, you can build a company. like this that generates significant free cash flow without raising round after round of capital. You don't need a large engineering team necessarily. You don't need to hire a large sales team necessarily right out of the gates just to hit a number. You do need expertise and you do need to protect margins. And that's fundamentally a different kind of company than I think what the Emergence playbook is calling for. And if anybody from Emergence hears this, ⁓ would love to chat with you about it and clarify. The second thing that sort of stuck out to me that sort of a critique of the article is is the advice around staffing pilots. So they ⁓ the playbook says that we should staff pilots of of these outcome-driven services with dedicated specialized teams, like Navy SEAL teams, and then transition to different teams for steady state delivery. I understand that logic. Pilots usually involve a lot of uncertainty. And you want your best people on them. But in practice, this creates a lot of problems. And it sounds a lot like a SAS, right? Like we're trying to sell SaaS again. ⁓ although I do think that that that the best people will learn from a pilot and they'll hand off to a secondary team. ⁓ that might work, but you're gonna lose a lot of context when they do that handoff and the engagement, you know, should be starting to compound results for the for the customer. It also sort of sends off the signal that the client is going to get a different team when when the deal signs, and that might just erode trust. So the better model in my mind is building a team that gets better at the full cycle, full life cycle from pre-sales all the way through delivery, pilot all the way through outcomes, ⁓ and and embed people in that who can scale and automate the process so that what ultimately gets built. Is productized into the service itself. And so this is where the forward deployed engineer comes in. Keep the same team on it, but include a forward deployed engineer in that initial pilot, right? So you have them looking at it from a product and a in an AI automation angle. ⁓ the third thing that I I don't I ⁓ point out here is sort of an overemphasis on product development, which sounds a little bit weird. ⁓ But the the article spends a lot of time talking about product roadmaps, North Star product metrics, productization strategies. This makes sense if you're talking about the a the product as the service you're selling, but we have to remember that this is not about a traditional digital product. Okay. This is this is a service as a product, an outcome as a product. And so most of the founders who would be reading an article like this, ⁓ They come to VCs with a with a digital product mindset, not a services product mindset. And so it's going to be a big adjustment for founders to adjust to this mental model of service delivery versus digital product delivery, ⁓ data product delivery. ⁓ which brings me back to my point about the founders of these companies needing to come from services. So with the services background, the mindset is already. That the product is the people. The product I tell that to my team all the time. We are the product. The product is the result that we're creating for the customer. The work you do for a client is the product. There is no separate software layer. ⁓ there's there's not necessarily something that we're going to to tangibly provide as a subscription to the customer that they have to use and operate. But the technology is how the work gets done better, faster, cheaper. And the founders who are conditioned to build software products, they're often stuck in the SaaS layer today, or now even an agent layer. But these are just services companies, right? And a services operation is fundamentally different than a SaaS product operation. ⁓ so it's it's this product development. Can be a trap that costs time and money. And I've seen it happen over and over again in services businesses. You think you're investing in something that's productized and what it really ends up being is a bespoke solution. So best to think of the product as a productized service that's repeatable, scalable, and over time gets better and better. All right. So the article does touch on something that they call the vertical ecosystem, but I think it could go a little further in sort of flushing this out. So the companies that are really best positioned to deliver AI nar native services are ones that are already embedded in vertical ecosystems. So what do I mean? I'll give you a few examples from my own resume that I think ⁓ illustrate this. So years ago I worked with a company called Black Pod based in Charleston, South Carolina, largest ⁓ publicly traded software company you've probably never heard of. ⁓ they serve the nonprofit sector and have have been for decades. ⁓ their customers are universities, churches, hospitals, arts organizations, nonprofit. ⁓ tight knit ecosystem, very specific workflows in the fundraising side of the business, ⁓ of these businesses, very specific language and the relationships in the ecosystem, very, very specific. An AI native services firm built inside of that ecosystem, one that knows how to configure and run BlackBod products, one that knows fundraising operations of a mid-sized, you know, university alumni association, knows what major gifts are. a firm a firm like that is going to have a significant advantage over a generic AI services company trying to serve nonprofits. And certainly. over a SaaS company that's trying to sell yet another product that companies and and and users inside of those companies have to figure out how to learn how to use. ⁓ Same logic or same applies to a company called Higher Logic in the association management space that I worked for, also sort of nonprofit tangential. I worked for a company called Rain Focus in the enterprise tech events industry. Which already actually operates as an AI native services company. They actually deliver a lot of the outcome in meaning they set up the software on behalf of the customer and then they also go help run the software and run the the back office operations of the event when the events happen. So I think actually ⁓ this company, Rainfocus, that I ⁓ spent a lot of time with, I think they're uniquely positioned, they're already doing ⁓ what we're talking about here. They are an AI native services firm. And I haven't spoken with these folks lately, but I would guess that they are really investing in in AI automation behind what it is that they do to drive their own gross margin improvement. ⁓ So those are some examples. And I don't think I don't think this is the same as saying you have to pick a vertical. The article says that, but there's a layer underneath it that matters. You have to be part of whatever ecosystem it is you want to play in. You have to have partnerships. You have to have people certified. And I'm not talking about technology certifications. I'm talking about the association certifications for the ecosystem that you're you're playing in. ⁓ they have to have community relationships, presence at the conferences, all of it. The AI alone is not the moat, right? It's the ecosystem position. It's back to the first thing we talked about, which is brand here. So ⁓ Anyway, I think ⁓ the vertical ecosystem insight is really important here because that's going to differentiate sort of a horizontal catch all solution that's very generic and hard to sell from one that is that is you know very easy to sell because you understand and can clearly deliver the outcomes that that vertical industry is looking for. All right, so let me bring this back to something I've been thinking about for a while because I think it it connects directly to AI native services. but SaaS based customer success has been chasing outcomes for what, two or three decades now? but the fundamental difference between SaaS and what we're talking about here is who's delivering the outcome. So the best organizations I've seen, like Rain Focus, is a good example we just went through. They went a step further than most software companies that I've seen, right? They didn't just tell customers what to do. They actually did it for them. They didn't just teach them how to adopt the software. They actually ran the software for them. They built playbooks, managed the operations. They owned the outcome of the events that they they still do, that they're involved in. And it's it's it's gonna pay off handsomely for them, by the way. ⁓ they became part of the customer's organization in a way that looks more like a managed service than software and support. ⁓ and that's exactly what it was. So AI native services to me is a continuation of that model, managed services. ⁓ and the assumption is that it can scale significantly with AI. And I believe that. ⁓ But the A but the CSM doesn't doesn't sit alongside the customer anymore in this model. The company has to become a services firm and take that CSM and translate it into somebody who provides operational work entirely for that customer or those customers. Okay. And with the with the AI handling repeatable high volume work in the background, ⁓ humans get applied, as we all know, when there are judgment calls to be made and when there are relationships to be managed. ⁓ but the customer ends up getting a result and then an outcome, not necessarily. hours from a person. So if you spent time in customer success ops, you understand what it actually takes to make a customer successful, ⁓ not just, you know, doing enablement during onboarding or not QBRs, but what actually makes them successful, then you've got a head start. ⁓ and you you probably ⁓ you you probably have an advantage over your competitors that that aren't that deeply embedded with their customers. ⁓ You have empathy for what the customer needs. And the question is, are you willing to take accountability that comes with owning the result? So back to my recruiting example real quick. ⁓ So let's say that an AI native recruiting firm in 2026 is running all these operational processes with AI or at least partially. ⁓ Sourcing is largely automated. The models that that company has built know what a strong candidate looks like for a given role, maybe in an industry. ⁓ Searches across platforms automatically, surfaces top candidates before a human reviews them. Initial outreach is personalized. It's not mass market templates, right? That just go and spray and pray to candidates, but the candidate screenings are actually personalized based on the candidate's background and the company that they're at and that the role that they'd be leaving. ⁓ scheduling, follow-up, status updates, all that's automated now. No humans chasing calendar confirmations, right? Interview prep documents for hiring managers automatically generated based on the candidate profile and the role requirements. So There's a nuance here. Yes, we're doing the work for the customer, for the client, but we're also doing it better, right? Because every a lot of the things that I just mentioned are things that maybe we didn't do before. The level of personalization might not have ever been there before because we just simply didn't have time when humans were doing all this. But that's a side note. So where are humans still essential in this process? the final review of the candidates, the relationships with the command candidates on the way through the process, working with the hiring team to drive a a decision on the right candidate to hire, debriefs with the hiring manager, anything that requires high emotional intelligence, we're not gonna delegate that to ⁓ to AI. Not yet. Maybe not never. ⁓ the the that's the current division of labor. And it will continue to change over time. Two years from now, AI might own more of the EQ tasks, but today it owns the volume work. ⁓ that firm, that AI native recruiting services firm ⁓ that embeds in their industry, that's that's the the firm that that's gonna win. Okay. So ⁓ I'll wrap it up here, but let me close with a quick recap from the emergence article. ⁓ and some things worth taking seriously out of it. And then what's take with a grain of salt. So first of all, domain credibility matters more than anything else. Just like any other company, if you don't have it, it's it's going to be hard to fake it. Gross discipline is the non negotiable. I'm sorry. Gross discipline. Gross margin. ⁓ discipline is the non negotiable, right? It's not about revenue growth. It's about gross margin growth. And it'll it'll start lower than you want it to. In an AI native services company. I'm I'm there right now building my own. ⁓ but the labor goes into cogs, and if your margins aren't expanding as you grow, you are not achieving AI leverage. Okay, so gross margin discipline is the key. ⁓ outcomes pricing from day one. Even if you start with a labor-based model, you know, sort of the minimum viable product version of your AI native service, ⁓ but just know that there's a transition. ⁓ over time that you have to go through. And then direct customer relationships and any partnership you build are key. The data that flows out of that relationship is your is your value prop, right? The the we didn't really touch on that, but the since you own the process, you you get to manage and own all of the data that flows through that process, which gives you both ⁓ ability to see insights across the industry you serve as well as The ability to improve the AI components of the service underneath it. Okay. All right. So those are all the good things I want you to take away from from this ⁓ emergence capital playbook. The the things to take with a grain of salt, the VC framing. These can be really good businesses without taking on a lot of capital, especially not on VC terms. So think about free cash flow, not just your top line growth rate. ⁓ Product first mentality, I think you got to be real careful with it. ⁓ of course you got to productize what you're doing, but it's a service, it's a productized service, it's not a digital product. And so don't fall into that trap of thinking that you're building something that you're gonna sell on a subscription to the customer. ⁓ and then the pilot team model. Just just keep in mind when you're delivering services to customers, you know, knowledge transfer is harder than you think. Continuity matters, the the relationship is the real value. Right. The trust is the real value of the relationship that you have with your clients. ⁓ the thing that's not in the article at all that that might matter more than any of this is that the founders who who win in AI native services are gonna be the ones who spent years inside of an industry, built relationships in that industry, understand what customers actually need at the day-to-day operational level. ⁓ I don't think you can learn that from a playbook or build it on the fly. That's the result of either a long career in that industry or a lot of hard work to come up to speed on it. And I've seen both work, right? I've seen founders really dive into an industry, become part of that ecosystem, but the expertise that they end up building, that's the foundation. AI is the way you accelerate it. Okay. all right. So that's AI native services. I hope this was helpful. ⁓ it was helpful to me to sort of unpack what the ⁓ industry is saying, thinking, feeling, doing about this. ⁓ I'll drop a a link to the to the playbook in the show notes. It's it's probably worth a read, especially if you're trying to build something that looks like this or if you have managed services in house today and you're trying to figure out how to scale that. ⁓ definitely a good read. So ⁓ drop us a line, J at customer success.io, Jeff at customer success.io let us know ⁓ what you think of the podcast. ⁓ leave us a a rating or a review that helps get the podcast out to more people. If you are so inclined, share it on LinkedIn ⁓ or wherever you hang out fr online. ⁓ we'd love to to to have you help promote this if you get value out of it. So hope that's been helpful. Hope you have a great week. Talk to you soon.