Craig Rosenberg on this episode
Craig Rosenberg, Chief Platform Officer at Scale Venture Partners, brings deep expertise in go-to-market strategy and revenue operations from his work across portfolio companies and enterprise clients. He led Scale's extensive State of GTM AI Report, combining quantitative research with in-depth interviews and hands-on hackathons to identify what actually works in AI-powered revenue teams.
The gap between AI productivity gains and revenue impact is wider than most teams realize: while 85% of GTM teams using AI see some benefit, only those implementing phase-two use cases—like AI-powered campaign analysis and account scoring—are seeing 3x to 5x lifts in pipeline velocity and conversion rates. Most companies remain stuck in phase one, using AI only for individual productivity tasks like drafting emails or researching accounts, without connecting those efforts to strategic business outcomes or revenue metrics.
The breakthrough insight is that successful AI implementation requires defining specific business outcomes first, then selecting tools to enable those outcomes. Rosenberg emphasizes that teams winning with AI know exactly what they want before selecting technology—not the other way around. This means starting with strategy and process (who to sell to, how to engage them, what messaging resonates), then layering in AI to accelerate proven motions. He highlights that marketing teams are three to five times more likely to unlock downstream revenue impact when they combine AI with data-driven segmentation, account scoring, and campaign analysis.
The episode digs into concrete use cases that are driving real revenue lift: micro-segmentation at scale using tools like Clay, AI-assisted pipeline reviews that surface deal risk in seconds, data enrichment via call recordings and email signals, and capacity planning that focuses reps on high-probability accounts. Rosenberg also covers why RevOps leadership and dedicated go-to-market engineers are critical to moving AI from ad-hoc experiments to systematic, repeatable processes—and why human-in-the-loop review remains non-negotiable for maintaining brand voice, deal quality, and customer trust.
What we cover in this episode
-
0:32
Phase One vs Phase Two AI Phase one focuses on individual productivity (doing more with less); phase two drives revenue impact through campaign analysis and account scoring.
-
3:55
Marketing Leading AI Adoption Marketing teams are 3x to 5x more likely to unlock pipeline and conversion lift when using AI for segmentation and campaign analysis.
-
6:00
Strategy Before Technology Successful teams define outcomes first (what they want to accomplish), then choose tools; AI amplifies strategy but cannot replace strategy.
-
8:42
Defining Specific Outcomes Winners know exactly what output they need—like a four-page QBR with data from six specific sources—before deploying automation.
-
11:28
Human in the Loop Requirement Every AI-generated output needs a human 'tastemaker' to approve messaging, deal quality, and brand fit before customer contact.
-
14:16
Campaign Analysis Drives 3x Lift Teams using AI for micro-segmentation, account scoring, and campaign data analysis see 3x pipeline volume gains and 5x conversion improvements.
-
18:19
Data Quality and Enrichment Reliable data sources (emails sent, meetings booked, call recordings) enable better AI decisions; enrichment via third-party tools fills gaps.
-
23:22
Hardest GTM Environment Ever Current market is more competitive and crowded than ever; AI helps efficiency but cannot overcome fundamentally harder buyer environment.
-
27:27
RevOps and GTM Engineers RevOps-led AI adoption is 20-22% more likely to succeed; dedicated go-to-market engineers scale ad-hoc experiments into systematic workflows.
The lines worth sharing
We know someone's going to be really successful with AI when they know what they want and decide how to use automation.
Craig Rosenberg · 8:42
The people most successful with AI figure out strategy, people, process, and then tech—in that order.
Craig Rosenberg · 6:00
Marketing teams using AI for campaign analysis and account scoring were 5x likelier to see increased conversion.
Craig Rosenberg · 14:16
Sales reps were spending 70% of their time on busy work and 30% engaging customers—AI is solving for that by freeing up time for real conversations.
Craig Rosenberg · 24:30
Common questions from this episode
What is the difference between phase one and phase two AI in go-to-market?
Phase one AI focuses on individual productivity—doing more with less through tasks like drafting emails or account research. Phase two drives downstream revenue impact through campaign analysis, account scoring, and data-driven segmentation. Only phase-two implementations show 3x to 5x lifts in pipeline and conversion.
How much lift do teams see from AI in go-to-market?
While 85% of GTM teams using AI see some benefit, only those in phase two see major revenue impact. Marketing teams using AI for campaign analysis and account scoring are 5x likelier to see increased conversion and 3x likelier to see pipeline volume gains.
What should companies do before implementing AI in go-to-market?
Define strategy, process, and people first—know who to sell to, how to engage them, and what you want to accomplish. Then select AI tools to enable those outcomes. Most successful teams know their exact desired output (e.g., four-page QBR from six data sources) before choosing technology.
Do companies need clean CRM data before using AI?
Data quality helps, but it's not a blocker. Teams can start with reliable data sources like emails sent, meetings booked, and call recordings. Third-party enrichment fills gaps. The key is accessing trustworthy signals relevant to your decision, not waiting for perfect data.
Which teams are best positioned to lead AI adoption?
RevOps-led AI adoption is 20-22% more likely to succeed. Teams also need a dedicated go-to-market engineer or center of excellence to move experiments into systematic, scaled workflows across the organization.
What role does human review play in AI-driven go-to-market?
Every AI output needs a human 'tastemaker' to approve it—whether messaging, deal quality, or brand fit. AI should accelerate proven motions and free reps from 70% busy work, but humans remain the final arbiter of what touches a customer.
SEO meta description
Craig Rosenberg on why 85% of GTM teams see AI benefits but struggle with revenue impact. Discover phase-two use cases driving 3x-5x pipeline lift.
Target keywords
Craig Rosenberg
Scale Venture Partners
State of GTM AI Report
go-to-market AI
phase one phase two AI
campaign analysis pipeline lift
account scoring
micro segmentation
RevOps
go-to-market engineer
AI ROI measurement
human in the loop
Read the complete conversation
Auto-generated from the episode audio and lightly cleaned. Expand to read or search inside.
Read the full transcript · 47 KB · Craig Rosenberg
CRAIG ROSENBERGWe know someone's going to be really successful with AI when they, instead of saying, hey, we're going to go use AI, they come in and say, we'd like AI to create a QBR that's four pages long, that has this on the first page, this on the second, this on the third, this on the fourth, and grabs from these six sources. Like they are the ones that are going to be successful because they know what they want and then they come and they decide how they want to go use automation.
SPEAKER_03Welcome to go-to-market science. There's an art and there's a science to go-to-market. And in this podcast, we talk about the science by interviewing CROs, private equity investors, and other sales and marketing experts, as well as talking about what we learn every day in the trenches helping to build go-to-market engines.
EDDIE REYNOLDSAll right, welcome to another episode of go-to-market science.
EDDIE REYNOLDSToday's show is on how B2B SaaS companies are leveraging AI in go-to-market. And I have my special guest, Craig Rosenberg, the chief platform officer at Scale Venture Partners. Thank you for joining me today, Craig.
SPEAKER_04Thanks for having me, Eddie. I'm excited to dig in with you and it's a pleasure to be here.
EDDIE REYNOLDSI'm excited to dig into this as well. So today, as I mentioned, we're going to talk about how AI is actually being used in go-to-market. And there's a lot of talk right now about AI and a lot of it is fluff. A lot of it is folks going in and writing prompts and chat GPT, and it's not super tangible things
EDDIE REYNOLDSthat people can take away and go action in their own business. But today's different. We're going to talk about some very specific examples. Craig, I know you and your firm have done a lot of work in this, a lot of research, and I'm excited to dive into that. But before we do, you guys have written the state of go-to-market AI report. And I read through it to prep for this, and I wanted to share my take on it, give you an
EDDIE REYNOLDSopportunity to respond. But there's a lot of really good stuff in there.
CRAIG ROSENBERGI love it. Let's do it. By the way, just for everyone to show that I'm an old man, I have to print the very large version of it for my copy of it that I carry around because I need it for my eyes on the digital version. So yes, I'm ready to go. I've got it in front of me, as you can see.
EDDIE REYNOLDSI think that's better because I'm over here trying to read off my screen, and then it looks like I'm not maintaining eye contact with the camera. So I like it. All right. So here's what I took from the report. Two thirds of teams are using AI and go-to-market today. And of those teams, 85% of them are seeing a boost. However, they're struggling to measure ROI. And the reasons for this, there's two phases to AI as you guys describe it. Phase one is you find ways to do more with less. And phase two is you drive growth, or as you guys termed it, the downstream impact. So phase one, we're looking at reducing costs of output. So focusing on more quantity. And these things tend to focus on ease of adoption and frequency. So for example, we can all pop into ChatGPT and ask it to help us draft a message or research
EDDIE REYNOLDSan account.
EDDIE REYNOLDSReally easy to do and something we're going to do very frequently. And so that's a very easy example of a phase one go-to-market AI use case. But phase two is unlocking the impact of revenue and actually improving the quality of the work. Examples being data analytics, analyzing campaigns to see what are more effective, or helping sales reps improve their skills. And you mentioned that marketing teams are three to five times more likely to improve pipeline and conversion if they're leveraging AI in this way. We also talked about how to get there. And you shared four steps to getting here. Defining the current state, which I'm anxious to dive into if there's time in this podcast. Letting RevOps lead, hiring go-to-market engineers, and not getting hung up on buying versus building because either can work right now. So now I'll stop with my soapbox. Anything you want to add to that, Craig?
CRAIG ROSENBERGNo, that was an incredible update. That summed up the report better than I could have. Excellent. So no quibbles here. That's...
SPEAKER_21Did you have anything that you had a quibble with and what you described or...?
EDDIE REYNOLDSNo, it resonates with me. I think my team and I have been a little hesitant on AI for a few years now. So much so that when we design some of our frameworks, we put AI like kind of at the end of the list saying, here's all the things that you need to do before you even think about AI. Now, I see a lot of exceptions to that. And we're going to go into some of those today. But I think that I still feel a little bit disenfranchised because it feels like there's this concept that a bunch of folks are like, let's buy everybody chat GPT licenses and tell them how to go like ask it to do account research. And at the same time, like my own team, we have... And you guys lead in this report with marketing being sort of like the highest impact zone. Marketing is the first place that we implemented AI in our own company. Every single thing that we do in generating content from like ideation to actually drafting things is using AI, but none of it is using AI without human intervention. And so I think your report really resonated with what I'm seeing with the companies that we work with and what we're doing internally here.
EDDIE REYNOLDSAnd I was just anxious to learn more about the things that I haven't learned yet and don't already know.
CRAIG ROSENBERGCan I just throw out a couple of things based on what you just said? Of course. So one is buying people chat GPT and having... The one thing I definitely saw... So the way we did the report, we did the quantitative, right? It's like 270 something people. But then we tried to interview between three and five people a week. So we would get the qualitative, right? So there was folks from the portfolio. There was folks outside. I mean, if we saw that someone was doing something, we would want to learn. And then we also have these essentially what we call go-to-market AI labs, which are hackathons. So we'd have people come and show what they built or bought or whatever that might be. I think the one thing that is important now in phase one and as you enter phase two is that everyone needs to learn AI, right? That we have a quote from Jordan Crawford, which is like, step one is everyone needs to learn how to talk to the machine. So I don't think it's a bad thing. I don't think it's the solution, which is what you're saying, right? Which is like, you're not AI if you just go, hey, everyone do use chat GPT and go do this, which I'm going to get to your second point, which is dead on in a sec. But there is a different point, which is like, who cares? Get everyone playing and understanding, learn how to talk to the machines, as Jordan said. That is how you unlock, not just whether you're going to build more stuff or you're going to buy more stuff, but you got to, this is a unique point in time where you do have to understand what AI can potentially do. And you do want to become versant in it as you sort of take that next step. So that was, because the second thing you brought up, which is really important is AI should be last in the list of things that you guys do on your framework, because the people that are most successful at it, figure out the strategy, people process, and then the tech, right? It's that that rule still applies. But the difference here is you're going to do this from a organizational perspective and a people perspective. You're going to look at like, what are the people doing it? Because like the phase one is about individual productivity, but you don't want them just doing a whole bunch of stuff. We have to figure out like, how do we work today and how do we want to work? And can AI help us go do that? It might not. And then you want to solve for that. And then we want to look at the entire revenue process. Like I'm sure you guys do with every client, right? You have to look at everything and what are the steps along the way and where's the latency? Where are the things that can be improved? And what is the problem? What would we like to do differently? And is this an automated or not automated way? And to be successful with AI, you have to start with understanding the current state and making good business decisions or you'll never get ROI, even on the individual productivity level, because it's not a toy, even though it seems like it is a search engine for a lot of people. To be really effective here, you have to start with this idea of what are we trying to do? How are we going to try and go do it? You brought up the sales part. If you want AI to help you with messaging, you have to understand what the messaging is. This thing is to right now is to expand and make you faster, bigger, better, stronger. But like if you haven't figured out who you want to sell to, right? And what you want to say to them and how you want to engage them, then you won't be successful with AI because that's not the magic right now. The magic is people that know what they want to do and they know what they want to get out of it. Those are the people that are going to be more successful. Well, I talked to Dan Gottlieb at Gartner, right? On the enterprise level, that's what he's like, look, we know someone's going to be really successful with AI when they, instead of saying, hey, we're going to go use AI, they come in and say, we'd like AI to create a QBR that's four pages long, that has this on the first page, this on the second, this on the third, this on the fourth, and grabs from these six sources, Salesforce, Snowflake, you know, whatever. They are the ones that are going to be successful because they know what they want and then they come and they decide how they want to go use automation. So to your point, I mean, you're just being smart, maybe you're hesitant, but like you're just being smart. That's the smart way to go buy any technology and AI is no different than that. The only difference I brought up before is right now, not many people fully understand the possibilities. So like along the way, you kind of want them to play with it and understand and to do things with it that they've never been able to do before.
EDDIE REYNOLDSYeah, I love that. It resonates a lot. And I want to dive into some specific use cases, but I also want to touch on what you just said, which is also a specific use case.
EDDIE REYNOLDSAnd I think the dream is like the outbound AI STR, right?
EDDIE REYNOLDSYeah. And so to your point, if we're not targeting the right accounts, then we spin up an AI messaging tool and we send out this messaging to the wrong people in the wrong companies. And we all know what results are going to come from that. And something that you just reminded me of that you said at the conference that I saw you speak at was like, we've been here before. This isn't the first time new technology has come to like radically change business or B2B SaaS go to market. And you think about like the sales loss and the outreaches of the world. It's the same thing. Like I remember people telling me when these first came out, like, this is amazing. You can just press send all and send out like 500 emails or drop voicemails. And how well did that work for everybody? This is no different. But if you know what good looks like, because you have human beings that have done it successfully, and then you try to reverse engineer the process and say, okay, so like the human that is doing this successfully goes to the website and looks for these things and then goes to LinkedIn and does this. And then they do this. And LinkedIn might be a bad example, because I'm not sure if you can access it via the API, but you have this step-by-step process. And then you can ask the intelligent questions of like, where and how can we use AI to help with this?
CRAIG ROSENBERGThat's right. I mean, I have a buddy, Dave Brock. I don't know if you know Dave, been around for a long time. And his favorite story is when CRM came out, he's like, you know, the presidents and the CROs, well, used to call them senior VPs of sales. They'd go, yeah, well, we just bought CRM and it's weird. It didn't increase revenue. And it's like, yeah, because it's there to support your strategy people process. And yeah, the same thing with outreach and sales loft. It's like, oh, well, we have an SDR problem. So we're going to go buy outreach or sales loft. It's like still going to be a problem, if not worse. And so, yeah, you know, that we're in that, it's a very similar situation. And so, yeah, we can talk about use cases because, you know, when you look at some of the state phase one use cases, the productivity ones, they all make a ton of sense, right? Because those are things that are tedious amounts of work that if we could accelerate, because your last thing, oh, that's what I want to say. Everyone needs human in the loop right now. And there's been a really cool set of interviews that we've had on the, we have a founder, David Boscovich, who's just crushed it with AI. And he talks about the humans are there. The new human you hire is a tastemaker. They are the final arbiters of, they have to have taste and go to market because the AI could do so much stuff. But if that, the human is the decider from a taste perspective on, is this going to work and does this look right and does this fit? That is the truth. That is the human in the loop on SDR messages and those things. It's like the tastemakers have to make these decisions on whether this is going to, this messaging, this look and feel like that. It's right now, like for a lot of this stuff, it's not a big, like go ahead and AI SDR as a human in the loop enabler to make SDRs faster. It's great. But if we have this pipe dream that they're going to be better than the greatest SDR, outbound SDR in the world, they're not, not yet. Right now, it can make your greatest SDR do more work faster and spend more time engaging with prospects. And that's because you have to have the human in the loop. And this idea that we're looking for now in employees in this AI first world is their taste and whether they can be that final arbiter of whether it should touch a customer or not. It's pretty interesting. It's part of the new way we're thinking about things. I don't think it's in the report, but it was definitely brought up multiple times. It sounds like some other, by the way, podcasters have used the word taste. So I apologize to them. I can't refer to you, but people have told me that.
SPEAKER_01So it makes a lot of sense.
EDDIE REYNOLDSIt resonates with our personal experience. We started in marketing and it really was, how do we come up with ideas, draft content, go through our entire production process for marketing content. And every step in that process, it's like, we need, we and I specifically need to double check it and say, is this really what we want to say? Is this our voice? Do we believe in this? And so that resonates. But let's stay away from the super high level stuff, because I think everybody would love to like, just press a button and have an outbound or inbound SDR or just generate all of our marketing content. And you just can't do that today without human in the loop. Yeah. But let's get into like more specific stuff.
EDDIE REYNOLDSI've got a whole list of things from your report that I can go through, but are there any other specific use cases you would like to start with?
SPEAKER_19No, you lead me down the path.
EDDIE REYNOLDSAll right. So let's start with campaign analysis. Tell me a little bit more about what you've seen there working and not working.
SPEAKER_35So in a phase two use case. Yeah. So when we looked at just really quick from a data perspective, the folks that were looking at their campaign data, and there's not a ton of them, as I'm sure you can attest to.
CRAIG ROSENBERGThis is like, the reason we call the phase two is it's just happening. Okay. But when the folks that did that saw incredible results, so like if you looked at marketing teams using AI for campaign analysis and account scoring, which is a lot of things, right? It could be signals. It could be segmentation. They were 3x more likely to see increased pipeline volume and faster pipeline velocity and five times likelier to see increased conversion. That's real data. That's really happening. That's not a lot of folks, but it is the start, right? That's why we believe there's a phase one, which was, let's get everyone productive. And now let's look at phase two, where we start to have these downstream changes to metrics. And that, you know, when we looked at the data, the folks that said they were doing both of those things, they saw the most lift for sure. They were the ones who had those downstream metrics. We saw a variety of different things on the campaign analysis front, right? Where, so it was hard to sort of pinpoint the, I can't just give everyone a tool because everyone was doing different stuff. Using AI before the campaigns went out to look at the various mix of touch points that worked for them in previous campaigns. Look, definitely a ton of more work done on micro segmentation, which I think was like kind of bleeding into the account scoring thing. We definitely saw that when folks were using data to gather smaller amounts of accounts to run their campaigns against and programs against. We talked about AI was helping them do that. They were making the funnel, but they were able to take things like external signals, internal third-party data, internal data to choose like, here are the hundred accounts. It's just, I'm making up that number that we're going to go after. And then from a business perspective, that makes sense, right? Now we've got a smaller amount of folks that we can concentrate on and we could tell the AI to go help us create content for those things because we're not trying to boil the whole ocean and ask AI to find the most general middle of the road thing. We want folks that are sort of in a small cohort and people are using data to go do that. That seemed to show up a lot in the outbound SDR teams, marketing teams that were being more effective. So it's like, I kind of lumped campaign analysis and the account scoring data that you see a lot of folks referring to the data use in the report. But those two things were like the predominant. So like I said, a lot of people using data to figure out what's working, what's not working before they go and after. Two, using data to find micro segments that they could be more targeted with. Because the third part is, you know, how can I use AI to go like Sydney Sloan's like, man, they do tons of micro segmentation at G2. And she's like, without AI, we can never do this micro segmentation.
SPEAKER_22So the big thing I want to ask about this is the data quality.
EDDIE REYNOLDSSo I'm with you on like the dream here. But the challenge I've always had with some of the companies that we work with is if they come to us when the data is just an absolute mess, is this possible? Or do you need to like really get your process and data like all dialed in before you can think
EDDIE REYNOLDSabout this?
CRAIG ROSENBERGYeah, I mean, I've lived in your world too, of course, right? Especially if you want to use first part, like your stuff, the data is really messy to come up with good conclusions. In a perfect world, that's straightened up. And but there's a reason you guys have had a really good business for a long time is that the data quality thing's always been a challenge and it's hard to do anything without the good data. But I would say in general, it's a requirement. However, I work with startups like you. They don't have time. So, yeah, they can bring on folks like you, which I would recommend that they do so that they've got that established foundation for them to make smarter moves going forward or their data is not quite as messy yet, which I'm sure you'd cool with. I'm sure everyone shows up in there. And if they could just start with third party data, to me, just to start, that's a luxury that we've never had before.
EDDIE REYNOLDSYeah, I think it's a spectrum, right? Like you can enrich a lot of data. And, you know, if I need to know, like what industry or X, Y, Z is happening with my leads and my closed one deals, if I go enrich that data, I can do that.
CRAIG ROSENBERGYeah, that's right. You can. But it's still your next a thousand things that the sales team will. I mean, look, I actually I'll back up a little bit.
CRAIG ROSENBERGI am very encouraged by some of the things that really messed data up in the past.
CRAIG ROSENBERGLike one was the sales reps. So this idea that we can, you know, there's applications that are automatically updating the CRM for you. That makes a ton of sense and will help with data quality down the road. I have a ton of just even startups, folks just like using gone calls or call recordings for their decision making. Pumping those things even through a horizontal platform to find things to read out on. So, like, I want to agree with you, but I want to say that there's other things that people are using to make decently better decisions across the board. And that's pretty exciting.
SPEAKER_22Yeah, I don't disagree with you.
EDDIE REYNOLDSAnd I think one way I've tried to wrap my head around this, because this has been an ongoing debate across my entire team for multiple years now, is there are certain aspects of data that you can get that are reliable. For example, if you can get access to emails, like those emails were either sent or they weren't sent. Those meetings happened or they didn't happen. What the sales rep recorded in their call notes is a harder thing, but that gong recording is there. So, and that industry code or that revenue range that comes out of whatever tool is providing that data, those things are there. And so, there are reliable sets of data that we can use. Yeah. And so, it comes down to trying to figure out, like, well, what data can we access with the
EDDIE REYNOLDStools that we have? And how is that relevant to the question we're trying to answer?
CRAIG ROSENBERGYeah, I mean, yeah, I would say your job is going to be more and more complex as you guys help people wrap their hands around the data, but they have access to sources that are way more powerful than they've ever had access to before along the way. There's the traditional data pen, the shit that used to break us all the time. Yeah, that's out there. But like, actually, I'm actually just forward looking going, there's all these killer data sources now that you can help them go sort out and make sense. There's tons of examples. A lot of the people we talked to were just grabbing, they were using something like, there's a ton of apps out there that are capturing the call data into the CRM. There was that. But mostly, it was like, you know, reading emails, looking at recordings, looking at third party data and being able to cobble together something that was relevant. And then when they get really good and they have like a RevOps team or consultancy that can help them go continue to add different sources to make better decisions, that's where we're really going to see the ROI flex. Because if you think about it now, everyone's doing something, but it's not what it could be. As we start to get better and better using all the data that we have to be able to make those decisions, then you're going to start to see increased conversion and even more pipeline or closed business along the way. Because the data's there. We just got to be able to get it into one place and be able to look at it. And that's starting to happen, right? And we'll just add more to it. It'll be more complex, but it will be uncomplex via AI and folks like you guys to help them go do it. And we're just going to see the numbers continue to flex. Remember one other thing too on the flex. It's also one of the hardest times to do go-to-market regardless of whether you use AI or not. So AI has come out at a moment in time when the buyer environment changed drastically over the last couple of years. And so a lot of the stuff that were in previous playbooks has to get re-sorted out in order for people to be successful. I didn't really talk much about this in the report. My team didn't either. We've really focused on the data and what to read out of it. But if you looked at what we wrote, it's exciting to see people getting phase two complex use case lift because right now there's not a lot of people getting a ton of lift, right? Regardless, because the buying environment, it's so hard. You can't get emails read. SEO is dying. Everything's changing. And so that is combined with this idea that you want to go make a more efficient revenue machine using AI, but like it's actually coming out in the face of one of the hardest times to do go-to-market actually I've seen in a while, right? It's just way more competitive. It's way more crowded. It's just tough out there for a lot of teams that, frankly, veteran folks I've known for years that have a way of doing things and those ways aren't working as well right now. And so that is another factor where we'll see more lift is as new best practices start to sort themselves out of the new playbook starts to manifest itself and AI goes and helps accelerate
CRAIG ROSENBERGthat. That'll we'll see even more lift there.
EDDIE REYNOLDSYeah. Well, and it's interesting you touch on veterans. I don't know what you mean by veteran, but I feel like from, I don't know, 2012 ish for the next 10 years thereafter. Like we had a really rich environment and you have folks that have been in go-to-market for a long time that are now facing like the first time it's gotten really challenged that
EDDIE REYNOLDSweren't around during the global financial crisis.
CRAIG ROSENBERGYeah. I mean, we can go there, but like even those folks I'm from before then, I mean, the guys in my generation, they're the same thing, right? You, you can't just roll out a bunch of reps from SAP and have them call the CEO and get a $40 million deal. Like the game all across the board, the game's changed. So I know it's a big deal. Like the SAS run up to Zerp was like a lot of people had no idea how hard this game was, but I'm actually talking about something even more different, which is like even the folks that had been around before the hard times still lived off a playbook for the last 10 years of outbound SDR sending a ton of emails, marketing, doing a ton of digital SEO, all these things to bring in, you know, 40 to 60% of the pipeline, just beat the crap out of reps, bring home deals, big deals, hire big guys. Those things are still in play, but they're not the secret sauce. The people that are doing really well right now have really great products. It's not like Cursor has a great go-to-market. They have a great product and they're sucking everyone in and that's creating this massive tidal wave. But like everyone else who's doing traditional go-to-market, it's hard. It just is. It's not the same. My point is, it's not just the folks that never learned what hard times are like. That is a factor. It's actually the people that have been through hard times too. They still had a way of doing things. And right now we just have to put everything back on the table. AI allows us to put everything back on the table and have different options and different ways of solving that problem that we've never had before. But my main point is, it's harder today to do go-to-market no matter what. So like to push new, like people coming back and saying, we're going to get better win rates. I think you can get better win rates, but it's not going to be exponential because the reality of the game, like we did a survey on just go-to-market metrics and folks selling AI have an 8% lower win rate than that of traditional SaaS. And that's because that's crowded, lots of options. It's like a, it's a brave new world for tech sellers. That is, that's hard. I mean, you could do AI, you could probably be more efficient. You probably get some more points on that, but like not back to the levels of 38 to 40% conversion rates because it's like it from a sales perspective, it's just harder out there. But anyway, we can go back into more specific use cases, but it is an important point, right? Which is AI is coming out in a rough time from the B2B go-to-market perspectives.
EDDIE REYNOLDSNo, I think it's a very valid point. I feel like you were touching on a number of actually more than just one use case, more than just this campaign analysis. You touched on account scoring. I think you kind of touched on even like defining ICP and buyer personas and segmenting. So like, if we just take a big step back and we think about how do we figure out who we are marketing and selling to, and how do we narrow our focus into just the absolute best folks that we want to market and sell to? Are there any specific tools that you're really excited about or use cases or aspects of use cases that we didn't dive into that you're really excited about within that realm of just where do we focus our sales and marketing efforts?
CRAIG ROSENBERGYeah, I got to be a little careful on vendor naming, but here's what I would say is that I think that Clay and the lookalikes to Clay have driven better decision-making across. It's a little bit of a different weapon of mass destruction than the sales engagement platforms were just being able to buy contacts in that you have sort of wonky. We always say there's like three kids in Brooklyn using Clay right now that have figured out that if we sell to these 500 accounts, we're going to be, these are like the ones that are doing X, Y, Z, and we could be more successful with. So I have tons of people come to me and say, oh, ABM's dead. And then they describe what they're using their tools to go figure out. I'm like, that is ABM. Like you are targeting the right types of folks. So I've seen folks that using Clay or some, you know, like a Freckle or some of these other things that are out there to go do or the ICP side, maybe, you know, a Revic, right? Where they're doing, they're looking at data and they're actually making good decisions on the overall target market and they're constantly reviewing it. But also I think Clay has taken a bunch of young folks that have no go-to-market experience as actually it's kind of refreshing to hear them present and say, oh no, we segment, you know, using like the tool has led them into segmentation and that's a good thing for the world. So when we talk about ICP, there are tools that allow you to go figure out what your ideal customer profile is. But you can, you know this, you can also just bet initially if you're a startup and then sort of watch the data and see how that goes. But then Clay has allowed them to segment the segmentation and allowed them to be more successful. So when I think of like account scoring, we kind of left it bigger. So lead enrichment, scoring, these things, actually those are key components to how people are tackling their more successful campaign because in many, not many of them have said, oh my God, well, we're ABM, we're going to hit targets and we're going to segment. They've actually started using the tool and then realize they could segment. I don't care how you get there, but it's like the, one of the most important things you can do. So that's one part of the use case that's really important. Micro segmentation. And there are tools that allow you to go do that. That's, I've seen that across the board on outbound. I'm telling you, man, like, unless you've got the brand of the century, you want to be in micro segmentation mode and taking on smaller chunks of folks. And yeah, I believe the new tools have actually led people down that path. Whereas the old tools where you had access to the world, that led us to try to email the whole
SPEAKER_12world.
EDDIE REYNOLDSI mean, this is exactly what my team is doing internally. Exactly as you describe it, micro segmentation, super tiny lists. We do 20 to 40 activities a day and book one or two like meetings a day. So our conversion rates are super high, close rates are in the mid thirties for a service.
SPEAKER_19That's amazing.
EDDIE REYNOLDSJust because it's so hyper focused in the exact way you just described it.
CRAIG ROSENBERGSo that's what we're talking about. Strategy, people, process, like the business as, but it's interesting, right?
CRAIG ROSENBERGThat tech has led many folks into the right business decision. But yeah, micro segmenting. When you look across all the use cases that are working, even the SDR side, it was phrased a little differently there, but updating systems and data quality, these things are part of that process here. When we looked across marketing and SDR, that micro segmentation, that ability to use tools, you know, like the clay and the lookalikes there to go do that, that has been one of the most effective things that people can go do. And that's where we definitely saw the, like on the SDRs, we saw, you know, we have a meetings crisis in SDR land and we saw folks that were reporting two to three X the number of qualified meetings they were sending. Those folks are all doing some version of what we just talked about.
EDDIE REYNOLDSYeah. I just, I can't think of anything more powerful than just getting hyper-focused on the right accounts, the right people in the accounts, the right number. We haven't touched on capacity planning, but something we're really pro on is, and this is an issue I've had in every sales job I've ever had. And I've spent my whole career in sales, aside from starting this company is every job I get, here's 2000 accounts to call on. And I'm like, I can maybe cover 200 this year if I'm lucky. So which 200 do I need to call?
SPEAKER_22And now all of a sudden I'm doing data analytics, which I do not think is what you want sales
EDDIE REYNOLDSreps spending their time doing.
CRAIG ROSENBERGYes, that is correct. By the way, one other thing I did see some of was on the SDR and on the sales rep side, on the pipeline numbers was the phone channel and using data like the Titan X's of the world to go find out of the 200 that we just chose, who are the 50 that are most likely to answer in the phone and going after that. Oh, that's cool. That has, you wanted sort of more specific, that I see, I've seen that, right? So if it's exactly your, you triggered me on the use case, because we're always assigning sales reps 200 to 400 accounts in their territory when they have to work X amount. And there's apps, Titan X will tell you of those 400 accounts and these personas, here's who's most likely to answer their phone. So you can, these people can manage their day against those folks that have that, that have that ability to go do that. And then, you know, whether they use a dialer, you know, or not, you know, they at least have a better, more efficient way of going to reach people. So that's another example of something sort of new and that people are using to be effective there. But anyway, that was a little different than what we talked about, but same idea, which is like, can we take this big and make it small with better odds of either engaging and or turning into pipeline or closing.
SPEAKER_22So let's say we've done all this. We've mapped out our ICP and our buyer personas, capacity plans, territory plans, dialed everything
EDDIE REYNOLDSin. We've got all of our reps with the right accounts. Maybe marketing has a list. Maybe we're doing ABM, but we're hyper-focused on a narrow set of accounts. What's next? Let's talk about account research. What, what did you see working best in that regard?
CRAIG ROSENBERGWell, I think account research is one of those things where good sales reps have always needed to do it, but they never, you know, only like 10 to 20% would actually do it and understand it. It was hard. It was laborious. And AI has allowed you to, to go do that in a way that makes you faster, even provides ideas, whether you choose to use them or not. But in like right now that is like the predominant sales use case. In my opinion, that's a phase one thing because like you can do it. We, you and I can go do it right now. We could ask Gemini, Hey, I'm about to go meet with this company. What are the three things happening in their business and what should I go lead with? I mean, it'll give me something, whether you choose to use it or not. That took a long time before it was the right thing to do, but it took a long time. And now you can do it really quickly. We all do it. I mean, I'm sure you do. I do it all the time. I mean, I'm like, I call in and find out what the heck am I about it? Who am I going to go talk to? So, you know, that one, I think will likely be best served in the horizontal platforms. The account research. We'll see if someone wants to make mods to some version that they have on top via their own internal vibe code are great. But like, that's a phase one thing that works. That is something that has always worked. That's makes for a better buying experience and seller tactics. And they can do it. You can do it much faster than you've ever been able to do before. So that's really, to me, there's, that's a very straightforward part of the story.
EDDIE REYNOLDSLet's talk about pipeline reviews. Have you seen anything super effective in terms of whether it's from a management perspective or a rep perspective, reviewing the pipeline, seeing risks and deals, seeing deals that shouldn't be in the pipeline, anything to just improve our overall quality of pipeline and where our
EDDIE REYNOLDSreps are focusing their time?
SPEAKER_35That is my favorite thing for all my buddies in the game. The operationally minded sales leader that's playing with how they get their pipeline reviews
CRAIG ROSENBERGusing AI, that's been a game changer. I don't know, you know, we had Ash from Anthropic come on and at our recent event, he was just going through his pipeline review. And it's like, I mean, by the way, he works at a place that sells AI, obviously, but like he wasn't grabbing, it wasn't mad. He was grabbing it from Salesforce, from Gong, just like from email. And then like, it had accounts that were in trouble, just it lined everything out, put it into the format. He'd optimize the format that he'd wanted and he was able to do it really quickly and then have it at his fingertips. And the reps did too. And it's been like a huge breakthrough. Those are my favorite stories I'm hearing from management where they've basically using the various data sources and bringing them together with AI into the way that they want it. Because one of the issues for the vendors right now is that AI provides insane, if you build, it's a customization economy, right? You're not sort of stuck in a rigid workflow. You can create what you want. And if you see kind of what's happening with those pipeline reviews and how these guys are using it, it's amazing. And it's only going to get better. And so, yes, I've seen, those were some of our favorite stories when we were doing interviews with folks that, and they were sending us examples of their pipeline reviews and they were sending them in advance to the rep and, you know, the rep loves it.
CRAIG ROSENBERGThey love it. I mean, it was, it was amazing.
EDDIE REYNOLDSAgain, this is all just tying together for me. If you think about it, if you're focused on the right buyers and you could have AI leveraging all this information to account research and helping you draft like really effective messaging and you've got human in the loop to review it. And then it's a double and triple checking your work as you manage deals through the pipeline. You now have your time highly focused on the right prospects and the right deals with all
SPEAKER_22of this help to save you time and doing the things that the best sales reps were always doing manually.
CRAIG ROSENBERGYeah. I mean, I'm writing a recap right now and I, I've had these, I've put myself out there on LinkedIn and I've had all these guys make these stupid comments calling things like busy work and all these things. And I'm just going, guys, do you remember three years ago when the data was coming out that sales reps were spending 70% of their time on busy work and only 30% of their time engaging with customers? That's what we're solving for here because the tastemakers, the end of the line, these are the folks we need to be engaging and thinking critically about customers and by freeing them up to be able to go do those things and to giving them tools so that they can be even more deadly in that conversations, engagement or whatever. That's what we're trying to do here. We've been trying to do that for a while and we've made with AI, we've made incredible strides, but you are right. You do have to make decisions on good decisions on who you sell to and how we're going to approach them and those things. If you've nailed those things and you know, all the other stuff we're talking about here today can make you even better.
EDDIE REYNOLDSNow with a lot of this stuff, are you seeing teams like manually loading all this data in or are they building all of these workflows across the organization?
CRAIG ROSENBERGWell, I think what we've seen in this sort of round one has been a lot of self-made work in the horizontal, you know, using a horizontal AI app and that's changing really fast as rep. We saw if RevOps was involved, then all the AI work was like, I think they were like 20, 20, 22% more likely to have high impact. And it's because of what you're just talking about, which is like, I think the cool thing about AI, it's been very democratic. So we've had forward thinking, ambitious folks build stuff. And then, you know, now the organization, the RevOps folks can come in and sort of take, because the workflows with the connections to other stuff, it's getting easier, but you still want something sort of standardized and run, you know, run via an operational arm of the organization. But I would say the vast majority of examples that we saw were someone just building it over a weekend. That was the vast majority.
CRAIG ROSENBERGYeah, that makes a lot of sense.
CRAIG ROSENBERGI mean, that will change. You can't, it doesn't scale, but like, that's why we say, just start talking to the machines. Once you start to play with this and see what this could be, and then you can do things with the other thing. The data tells us, you know, like I said, one, the data told us RevOps, when they lead, it's going to be likely more successful. Number two is that having a vibe coder, go-to-market engineer, someone that's dedicated to the, you know, to AI in either the operational groups or within the functions or in some kind of center of excellence that accelerated things as well. And so to be able to effectively execute what we're talking about, which is taking random acts of really interesting AI stuff and putting it into sort of systematic and always on type of work or putting it into the overall workflow, that's going to require having some type of person that's dedicated fully to AI in the organization. And frankly, an ops team that understands how to make these things work and, you know, roll them out. But like I said, the inspiration has definitely come from, at least what I've seen from individuals.
SPEAKER_63That doesn't surprise me.
CRAIG ROSENBERGYeah, for sure. And hackathons, a lot of people are doing hack, like just put 10 sales reps in a room and go build something together. Put 20 marketers. Maria Pergolino just left here. She was in EIR. The first thing she did, we're flying to Miami. Oh, good. Marketing kickoff. She's like, no, hackathon. We're all going down there and we're all going to build stuff. And it's not because they're going to use all that stuff.
CRAIG ROSENBERGIt's just, this is how you sort of, you know, you find new ideas and these things.
EDDIE REYNOLDSI think it's really cool. And I would say this stuff doesn't surprise me. And at the end of the day, I mean, salespeople know exactly what is taking up the most of their time and is hardest. But I agree with you that at scale, you want to architect this across the organization. So you get these ideas from salespeople, like what is dragging your capacity? What is the thing that is hardest for you to do manually? What's the thing that can have like the highest lift if we improve this?
EDDIE REYNOLDSAnd then you would want to architect that across the entire organization, especially if you're not in a startup with two salespeople, but you got 200 salespeople. Should each of those 200 salespeople be writing their own unique custom prompts in the LLM or should we build something for the team?
SPEAKER_01That's right. Yeah, I'll give you one since you help folks with their, on the operational side. Remember, there was a moment in sales tech when everyone, you know, like sales reps would
CRAIG ROSENBERGget used to an app at their previous company and then come over and they were just miserable. Either they didn't have it or they had something else. This will be interesting because you're going to have a lot of reps that have built something they prefer in a horizontal platform that's theirs, that they've built for themselves. Like they just love it. They can't wait to use it. And then we go in, it's a 200 person org and ops wants to rein it in. We've got governance issues, all these kinds of things. That rep's still going to use their own thing, right? And is that okay?
EDDIE REYNOLDSI've been that rep. I know exactly what you're talking about. Dude, you have to deal with it all the time. I'm sure for the last X number of years.
SPEAKER_65No, when I worked at Salesforce, I was like, I think we didn't have like a note taking app
EDDIE REYNOLDSand I had to like hack the computer to get Evernote on my computer. Cause I'm like, I'm not taking notes instead of Salesforce. Yeah, that's a good one. Just not going to do that. Yeah. Well, anyway, we are at time. So I've plugged your go-to-market AI report, which we will put the link in the show notes. Where else should people look if they want to learn more from you and from scale?
CRAIG ROSENBERGWell, I would say from scale, you can follow the scale venture partners LinkedIn page because we put two things and go-to-market things out a week. And, you know, I typically will funnel through them. The other thing to follow is my podcast is the transaction that you could follow the LinkedIn page there. Just go to your podcast app of choice and follow that.
CRAIG ROSENBERGThose are two really easy ways to do it. You can follow me too, but I feel weird talking about that. But yeah, the, I'd say the scale page, the transaction page, and or just listening to the podcast. I think those are two really good channels for you to pick up the stuff that we're, that we're doing.
EDDIE REYNOLDSWe'll try to put those links in the show notes. Thanks so much for joining me today, Craig.
SPEAKER_41Thanks for listening to the show. If this resonated and, or you'd like help with anything we talked about in the show, please reach out to us. You can find us at unionsquareconsulting.com and the info will be in our show notes.
