Shipping revenue: Will Sullivan on modern GTM ops, agents, and operating revenue like software

James Geyer: Hello everybody. We are back for the latest episode of boardroom RevOps, where we’re bringing you valuable tips from RevOps experts so you can make to the C-Suite. I’m James, co-founder of RevOps of the RevOps BI platform. Super excited to welcome a friend of the company and a friend generally Will Sullivan to the podcast.

He’s the founder of Predictive Analytics Partners. Will great to have you on.

Will Sullivan: Hey James. Good. Great to be here.

James Geyer: Will is an absolute expert at the intersection of finance and go to market. Uh, we’ve collaborated on content in the past around things like the importance of ARR waterfalls. But will’s also dive really deep into AI and agents.

And so that’s gonna be the topic for today, talking about modern go-to-market ops, modern go to market processes. It’s gonna be great. Um, before we jump right in, will, do you mind just share your background for folks in some more detail?

Will Sullivan: Yeah. Quick, straightforward. Uh, I lead a, a group called Predictive Analytics Partners.

We, uh, focus on building exit ready analytics for CROs and CFOs. And so for, uh, PE-backed companies, that means helping with M&A preparedness. Diligence integration. And for public companies, we often help move, uh, CROs and RevOps teams to a net new ARR model. I learned about these from working in M&A, worked with some of, uh, James and Josh.

Um, not when they were at Blair, but we used to work a lot of Blair guys and, uh, helping companies, helping buyers look at the go to market diligence and effectiveness and quality of revenue. While I was on the m and a team at PWC. I have a unique background as an operator as well. I, I joined a search fund, acquired business and ran that and sold it.

I helped sell it, uh, as well as I also led and ran a regional oil field service company about $200 million p and l. Prior to that, I led troops in combat. So I have a unique background, but I really love multiple arbitrage, and that leads you to really day to day, day in, day out of sales and RevOps. To drive those multiple arbitrage opportunities.

James Geyer: Yeah, that’s great for anyone listening, multiple arbitrage. Basically buying a company or investing a company at one multiple, improving it from an operational standpoint, selling it for larger multiple, making a lot of money. In the meantime. Well, I think it’s so awesome. Like you do the best job of probably anyone I know of communicating like the value of good RevOps.

I think multiple arbitrage is one great example of it, but in so many cases you’re talking about like improving net retention, improving the customer experience. And I think a lot of people in RevOps, particularly at the mid-level, actually get stuck there ’cause they think so tactically and process wise.

So I think, uh, for everyone listening, tune into this, tune into the content we’ve done in the past because I think you do a great job of that.

Will Sullivan: Yeah. The way I think about that is, you know, the world of RevOps really evolved from like. Salesforce admins, and so that really created a tooling approach so I could focus on tech stack and tooling over strategy and frameworks.

Yeah, and I think when you come from a finance or operator background and then you move into RevOps, you really have a strategy over tools, framework, and focus.

James Geyer: Yeah, that’s great. Well, let’s dive right in. Will, um, so we’re gonna talk like AI agents, all that good stuff today, very topical. But maybe let’s just start with a wide aperture.

Like what does a modern go-to market ops like process kind of look like to you? Like how should folks be thinking about the framework for, for go-to market today?

Will Sullivan: Yeah, it’s great. I think go-to market ops has been a on again, off, again topic now for a number of years and I really think that the agentic revolution is driving a shift to go to market ops, which is really just another way of saying.

We’re gonna break down the silos that goes from customer acquisition into customer expansion. So as some people may call it the bow tie or the full funnel effect, but it’s really, it’s that golden record or, or full visibility in the customer journey. And in order to do that, you need breakaway from these departmental silos.

But these departments have silos because over the last 20 years we’ve been, the market has been selling vertical SaaS solutions into marketing, into finance, and into sales. And they never did a great job of integrating or that you had to have a team that helped you with integration. And so that really was only teams north of two 50 million in ARR.

And so now we talk about go to market ops. It’s really leveraging a agentic orchestration, go to market engineering, or really basic Zapier connections across marketing, attribution and enrichments pipeline notifications, scheduling and booking and follow ups automations. In the past, these were all done through workflow and Salesforce, and maybe you bought Marketo or other out outreach tools.

Now you can really do it from one single vantage point, leveraging clay or cargo to see the full enrichment contacts and touch points across the go to market operation.

James Geyer: That’s great. So to, to summarize, make sure I understand this, it’s basically automating the entire ideal customer journey. And to do that you need to break down silos across all the systems in each of the functions at all stages of the bow tie, obviously underpinned by AI and agents.

Is that a fair summary or did I kind of misrepresent

Will Sullivan: it is I think the, um, another way I’ve heard it said recently is that a good go-to-market operations team can ship revenue. Like they can ship code. And if you’ve ever been a developer and you know how to do, um, you know, push code to GitHub and then push it live to production, that’s the idea.

And so what that means is that you have systems from all across a customer journey that you identify. If I deploy these systems at scale or uh, focus in this area, I will have an uptick on conversion rates or lead engagement. Or meetings booked or whatever metric you’re driving towards, and you can ship that out there.

And in order to do that, you have to really have a go to market ops approach where your RevOps team is working with your marketing ops team. Your CRO is strategies aligned with the CMO strategy. Mm-hmm. And your CFO has full visibility on cac, by lead, by marketing attribution, by segment, by vertical, by firmographic.

And so he understands where to deploy capital, to drive revenue and keep CAC at a reasonable place to drive profitability.

James Geyer: That’s great. I think that’s, uh, awesome Additions. I, I’m hearing understanding like the go to market model when you bring in CAC and like deploying investment to, to, to bring back your, your finance and RevOps kind of intersection.

But it also makes the question, like when you talk about shipping revenue, it’s again, very kind of systems based. Um, it begs the question, where does it make sense to have human sellers versus agents, both one or the other. How do you shake out on that?

Will Sullivan: Yeah, I think, uh, I’ll go back to the finance term that you and I know and love is think about like CAC payback.

And so if you’re going to deploy, let’s just use round numbers, uh, um, you know, 300k OTE sales rep, then you need to have a payback on that. So typically they’re gonna be a line towards a higher price points, so SLG, and it’s gonna be some sort of either enterprise client or mid-market clients. Um, or we may have an enterprise go to market motion.

So that makes a lot of sense to have human touch because there, there just needs a human touch. One thing that agents don’t do well right now is selling to buyer groups. And so it’s a really hard thing to construct. So if AccountAim is gonna go sell to IBM, there’s gonna be a multiple set of people you’re gonna sell to at IBM.

And you can’t just deploy an outreach agent and reach out to procurement or reach out to head of sales. You really need to have a multi-touch view. And typically you need a human to do that. So that’s on the higher end spend of a human. And then on the product led growth, or let’s just call it lower lead score or a non MQL, that’s where you deploy agents because it’s a lower probability of success.

You don’t wanna spend too much money there. You need to also enrich those lower end scores and enrich it so you can identify as scores. And then, um, with, uh, on the PLG side, your course want to have a volume play and agents can be more volumous than a human.

James Geyer: Got it. That’s a good framework. And so I’m hearing in the more complex upper middle market enterprise sale, humans still makes sense because you need to be able to multi-thread, kind of keep a narrative loop.

People in, like manage humans. Whereas at the lower end of the spectrum, like lower ASP, it’s maybe more cut and dry, quick sales cycle, you can kind of afford. To automate it because A, it might not be that profitable to have a human in the seed, and B, it feels like maybe there’s just less risk overall.

Like even if you blow a sale or two, like it’s not necessarily the end of the world, is it?

Will Sullivan: Yeah, it’s a higher volume play, so you wanna deploy lower cost solutions. I wanna like correct you on one thing you said multi-thread. So multi-thread refers to like deploy on multiple channels, like on multiple socials and then also on some ad tracking.

And then. Also on some maybe, you know, event based. And so that would be like a multi-thread. It’s really the buyer group approach where it’s multiple personas and multiple people you’re outreaching to. And that has complexity of like, well, I know when I speak to James, he likes to BS for a little bit. If I speak to Josh, he’s going to be quick and clear.

Cut to the point. So those kind of things, it just, it gets nuanced to build to an agent for outreach. And I know I got that backwards ’cause Josh is,

James Geyer: you know, I was gonna say, yeah,

Will Sullivan: that’s the thought process and you take that from two person example to a eight person example and you need a person involved.

James Geyer: Yeah, makes sense. How do you manage a world of both human sellers and agents? And maybe it begs the question sort of like, how do you manage agents generally? I think a lot of people are still trying to figure that out. And I know you’ve gone pretty deep on this recently.

Will Sullivan: Yeah. It’s, it really gets back down to this, the, the phrase of attribution, right?

And so we talk often about marketing attribution or revenue attribution, but really now we want to, you know, bucket, you know, um, cac, you really want to align. How are you engaging? And interacting with these leads or these customers and tag it to an agent spend. And so early, early days here and what I’ve seen people are doing.

Is if you have a single agent for all of your sales, and there are some out there like a Piper or a Alta and a few others out there, they were, uh, when I speak to these companies, their integration to HubSpot or Salesforce is coming from the integration user. And so it’s really easy identify that this lead was created by integration user and now you know it’s only coming from that one agent you’re using.

But if you’re like me and you’re kind of towing around with more than one outreach or sales agent, then it’s about using. What I do is a, a nomenclature on the campaign, so be a prefix or a suffix. And so I’ll use a prefix for like, let’s just say, um, uh, manufacturing, um, pe portco, CFOs, and I’ll do a prefix of like.

AI and the agent I’m using, or if it’s something like I did myself and I do a suffix or a prefix, depending on how I’m feeling that day, I need to get more organized of my initials. And that way I know that’s been, the campaign has been tied by, it’s been built by a human, or it’s deployed across a, um, from a campaign from an agent.

And if you’ve used one of these outreach agents, you know you can build these campaigns within less than four minutes. Versus a human doing it and you’re going into a flow in Salesforce and building it or Marketo or hey reach and it takes about 20, 30 minutes. So those are kind of a few ways. So I think, uh, right now it’s based upon created by user or if you do a prefix on the campaign, and I’m still toying with that.

And so once you have that, it still kind of begs the requirement of like, how do you have visibility on created by and activities. Paying activities and that kinda gets to like standard RevOps stuff, which most people should be able to handle on this call.

James Geyer: Yeah, that’s a really good tactical tip for, um, how to think about monitoring or managing both of these.

And also how to tactically get to the ability to track it from a metrics perspective. Are we talking just like traditional. BDR funnel sales process funnel? Like which, which metrics are you actually looking at as you post?

Will Sullivan: Yeah, I, I think a lot about activities and so I, I call it demand gen activities.

So how many posts are going out per day tied to a campaign? What kind of campaign is it as a high reach campaign? Is it an engagement campaign or is it. Product education campaign. And then, so are you doing demand gen activities? Are you doing going to podcasts, doing webinars, going to, um, you know, those type of things.

So those are human based, so you can kind of identify that there. And so once again, uh, then there’s the outreach campaigns or also why I talked to another company recently. They’ve had their best return on agent spend on non MQL campaign. Mm-hmm. Anything that hits their website or signs up for something.

And they enriched it and it’s non-marketing qualified. They’re deploying an agent at it and they’re getting a 25% conversion rate up from like a 5% or 3% conversion rate just because the speed to lead and the person’s curious and there’s something with the enrichment either that doesn’t fall into, they don’t fall into the lead scoring or just kind of surprises them, and it just shows that the market is moving.

And that non-traditional people and players are looking to buy some stuff.

James Geyer: Yeah. That’s great. How do you layer in and are there any qualitative things you think about, like, something I hear in the market about agents is like, oh, you’re just gonna like burn through your entire tam or something like that.

And I don’t, you know, you can debate what that even means, but I’m curious, how do you think about like this qualitative, um, the qualitative component of the experience and things like that?

Will Sullivan: Yeah, I think that gets into guardrails. And so this kind of gets into standard RevOps of assigning bags with territories and in in agent development deployment that is in guardrails.

And that’s saying this agent and this campaign will reach out to this very filtered, narrow prospect list. And that gets into how do you manage that prospect list and then what the agent can and cannot do. And if it hits a guardrail, it escalates to a human. And so those type of things. And then when you think about like burning through your tam, so you’ve got a total adressable market of a hundred thousand accounts, and if you’re deploying an agent, you could reach out to them very easily.

There are some natural throttles. So on LinkedIn, I think everyone knows you can’t send more like 30 outreach connections a day. So that’s one to be thoughtful about how you use that. But then you can kind of, um, get around that by creating alternative LinkedIn accounts and, and buying accounts. So there’s other, other things there when we’re talking about deploying systems.

Well, I think I heard it called, uh, systems of a mass disruption or systems of mass confusion, someone called it. ’cause we all get spam and everything is, you have to be even more thoughtful about how you go to market. And that’s where I think what’s exciting is that we’re gonna see go to market operators and RevOps move away from task oriented and be more strategy direction.

And because it’s easier to deploy these things as opposed to spending a week or two weeks building a flow in Salesforce and fixing it and changing it, you can now just do it very quickly in cargo or clay or buying one of these off shelf tools.

James Geyer: Yeah. Um. What happens? So we talked about activity tracking, we talked about some conversion rates.

’cause we think about now or in the future, managing a team of AI SDRs or something. You, you posed a, a rhetorical question a while ago that I think is a really good one. I’m curious how you’re thinking is involved, like what happens when your AI SDR doesn’t hit quota?

Will Sullivan: Yeah. I, it’s, I love asking that to vendors that reach out to me and they’re like, uh, you know, and so really it really comes down to like anything in sales.

What inputs are you putting in to create your outputs? And so I think about it, um, of, you know, if you’re developing a flow or recipe or, uh, you know, a Canvas app to deploy an agent for an outreach, then you have to figure out what’s the prospect list? Did you have the prospect list correctly? Did you have the right sequencing?

Um, correct. Did you have the right touch points? Correct. Were the right context library you were using, was it correct? What were things you’re saying? And so really get some of the, the inputs. And then the activity once again would be the sequencing and the touch points and how that goes. Does it feel very agent like, is it on purpose or is it not on purpose?

Are you trying to, um, make it sound like a human and, and confuse them? Um, and then often the thing right now is about speed to lead is like how fast these queries run and how fast they, they go out. But it really comes down to like who owns it. And that kind is a question that everyone should be asking a vendor right now is if your agent does not meet quota.

How do I improve the process? And they’re gonna say, well, we have a customer success team. Great. Does that customer success team have an expert in my domain, my industry, and with that per person will help me review these flows. And what you’re seeing is that there’s actually an increase in price on service of these agent companies, the outreach agent companies, or tagging on service requirements so that the end user gets a full ROI.

Hmm. ’cause in these early days of fragmented market of agents, brand reputation is everything. And so they don’t want to just sell you an agent and say, go have fun with it. They wanna sell you an agent and make sure it’s successful. And so that’s why you may see some higher price points, but those will also be the ones we have.

Customer success will, um, be better. So long answer to say whoever owns it, the vendor, the go-to-market engineer, or the uh, four deployed AI team, they’re the ones responsible and they’re the ones held accountable.

James Geyer: Yeah. Sounds like, uh, a pip a performance improve improvement plan, essentially. Right? You’re still gonna go through, back to basics that you would do with a human, but, um, just,

Will Sullivan: but it’s not necessarily on the, on the agent though.

It’s, it could be on the flow, and so you may have like seven flows crushing it and one flow doing awful. Mm. And so that’s, it’s, it’s not, it’s a, it could be a pip, but it could be a PIP on a workflow.

James Geyer: That’s a good distinction. Yeah. Interesting. Another interesting, interesting thing in there that you mentioned, I, I feel like with vendors tacking on services, it really just starts to feel like an outsourced SDR firm, which I get hit up by all the time, and I think it’s been around for a while.

It’s funny, I feel like AI broadly is looking more like services and software rather than software as services. Right? Yeah,

Will Sullivan: I, I know that term has been thrown around lot this past year. I really prefer the agent enabled service. Because that’s what we’re talking about here now. Yeah. Is like one, once again, if you’re a software without an agent, you’re already living the past.

And I know there’s a number of companies out there that do not have a conversational agent, and it’s like you, you need to have one, right? Just for customer success or for onboarding, just needs to be upfront on your, on your page. And so this idea of agent enabled service gets back to service and you’re solving a problem.

You’re not solving a job to be done. You’re not. Running a model, building something, you’re actually helping a company grow and evolve and, and service them. And so I think it’s this idea of agent enabling service, and this is where you see difference between sales and marketing agencies who just, they’ll sell you something and then they’re going to deploy some cl, um, some clay spreadsheet, prospect list and workflow for you and charge you way too much on it.

Or it’s gonna be someone who actually will help build out agent for you internally. Or build out an agent for you, uh, alongside you, so you have a custom agent within their software system with a customer service person helping ’em with that. Mm-hmm. When, once again, going back to this idea of like strategy over tools, being able to identify someone selling your tool or someone helping you execute a strategy and advising on it as it evolves is, is the critical component.

James Geyer: Are you finding the AI SDR R is a common term right now? I think it’s been normalized. I don’t hear much about like the AI ae, like where are you finding the AI process from a GoTo market perspective stops today? Yeah. AI is not closing deals yet, is it?

Will Sullivan: Let’s define these terms. Ai, SDR is basically the BDR and SDR, right?

So it would do a prospect list. It would do the outreach. It would do a emailed outreach and a. LinkedIn outreach. The key on the emailed outreach is a good software or agency vendor will provide, uh, cold email warmup, which is critical. Um, and so I, they’ll just go through the engagement or, uh, prospect awareness engagement and get them to the meeting.

And so that’s where like the AE typically comes in. I have seen buyer agents where you go log into a site and a buyer agent will walk you through a step flow and allow you to ask questions that are non-scripted. So like there’s buyer agents I’ve seen, once again, this is for lower cost PLG motions and I think the, we’re still gaining momentum in the market is these kind of buyer demo spaces.

Mm-hmm. Where you spin up a demo space just for that buyer and they can upload some data and have a proof of concept in the moment. So that’s not an ae. It’s not like an AI AE or agent ae, it’s really just like a, a demo space that’s spun up. And so the ability to spin up that space is done through obviously code and some, probably some sort of agent to code.

But that’s where I kinda see where like the AI SDR ends either at a buyer agent for PLG reason for low cost, or into A SLG or enterprise environment where there is a demo space that is built for that company’s. Um, domain and industry knowledge.

James Geyer: Yeah, I think that makes sense too. And I, I, I do subscribe at least for a while from now, AI making sellers much more productive in like the mid-market enterprise versus being replaced completely.

I’m curious if you agree or if you think sellers.

Will Sullivan: I agree. I, I think, uh, the, once again, this company, I’m, um. Uh, I’m using their agent and I talk to them often is they’ve seen, uh, their daily meetings going from four a day to seven a day, and then they get to eight a day. And so once they start crossing that seven a day threshold for weeks on end, then they go out and hire someone.

So when you look at these companies that have, um, they’re selling agents and they’re hiring, they’re probably hitting 7, 8, 9, 10 meetings a day. And so those are kind of key signs of a high growth, um, company. And that’s where I think, uh, these reps are so critical because it is a hard skill to listen to someone and keep a framework in mind.

I’m like, am I doing a challenger sale? Am I doing command to the message? How am I being empathetic to this, uh, buyer? Yeah. How am I looping them back on topic? You know, how am I getting ’em to say yes, those are skills. That a human, um, is, I think in my, in my mind, for a number of years, would be better at, and just, it’s easier and more personal to interact with the human instead of a glossy faced agent that you see, uh, like a human interface, um, uh, approach for an agent.

James Geyer: Spot on. Earlier we talked about CAC and like the economics of, you know, 300k OTE and like when to think about hiring a human versus an SDR. And you also mentioned the skills needed to close out a sale, basically anything. And, and the number of meetings before hiring a new rep I think is a good triangulation.

Like anything else to add on, like how to think about the investment case for humans versus agents, like an annual planning. I feel like we covered it pretty well, but curious if there’s anything we, we missed in that discussion.

Will Sullivan: Yeah, I’m think like step back from that before you get to the investment case, you gotta have a framework.

And so I, I, I wrote about in my article a few weeks ago of, for 2026, are sales team’s gonna have a, um, automation approach, like a autonomous agent where you hire an AI SDR and some can go from 10 K to 30 K or more. And so that’s a pretty big, um, investment. Um, not as much as a hiring a human, but it’s still like you’re spending 30 K on a tool.

And you need to know it’s deployed. And then you still need to have this understanding of like, if I’m gonna have this tool, I’m probably gonna have 10 to 15 campaigns running. How do I manage that? Mm-hmm. And so then you get this idea of like, maybe I need, uh, someone smart to manage this. You have it in-house, you hire for it, or there’s a framework of I’m gonna hire a go-to market engineer and orchestrate across existing tools.

So I have an hay reach or an outreach. And then I have my CRM and then I have ZoomInfo, and then I have whatever else it is. And you connect all these systems. And that’s kind of kinda a, a Lego approach of connecting different things you need to go to market engineer for, either as, as a hire or an agency.

And so those kind of framework is where I’d start is, are you gonna connect the dots through orchestration or you’re gonna buy something out of the off the shelf and deploy it and you feel comfortable, you have process in place and a person in place to manage that. So on those regard, then the question is the investment case and are you trying to drive down cac?

Are you trying to drive lead engagement? So right now maybe you’re only engaging with 50 or 75% leads, you wanna engage a hundred percent of leads, or are you trying to improve net retention rate by making sure you’re getting the right customers? And it might be an enrichment play and then a nurture engagement play.

James Geyer: Yeah. Uh, makes a lot of sense, I think. Is there anything that you think AI is best suited for across those three kind of options?

Will Sullivan: I, I think the easiest one right now is enrichment. I mean, if people are not engaging all their leads and not enriching all their leads with an ICP score, firmographics and Technographics that align to their ICP and if people aren’t having every two months a ICP review.

Like we have so much ability right now through the low cost option of Apollo or Clay and using credits of just enriching your leads that hits. And so that should be like number one. And then number two is, the other hard part is how do you identify dark funnel leads? And so leads that hit your website just through an IP that you don’t know and is trying to get better visibility on those dark funnel leads.

And that really speaks to the CMOs office. So enrich what you got and then go out and reach, um, uh, more of the leads that get there on a dark funnel. Enrich those and then you can align the right campaigns. And so those are, I say, top of the funnel stuff you should be doing

James Geyer: sadly. We’re almost out of time.

Last question for you. How can RevOps folks become more skilled in implementing agents? I think you dropped a lot of great knowledge here. Anything else you’d recommend if, you know you’ve gone through this journey yourself? How folks could replicate it or, or dig in?

Will Sullivan: I think there’s a lot of like low cost options out there of, um, you know, following someone on LinkedIn to joining a school for 20, 30 bucks a month.

Um, going to like a, a more higher in school like GTM Engineering School. But really what I do is whenever a sales rep calls you from an agent company selling you something, talk to them and talk to ’em again and again. Ask them these questions. How do you do this? How do you deploy it in your business? Can you show me your CRM.

I’ve learned the most by asking these sales leaders of these agent companies to show me their CRM and they’re the ones figuring out firsthand. That’s where I’ve learned the most.

James Geyer: That’s great. Great place to cap it Will. Thanks for stopping by. This was, uh, super insightful, so I appreciate it.

Will Sullivan: Thanks, James.

Glad to be here.