Unlocking Hidden Revenue Signals in Your Contracts and Calls with Keith Rabkin

In this episode, PandaDoc President Keith Rabkin shares how his team is rethinking sales productivity with AI—not by adding more tools, but by unlocking insights from what they already have. He and Emir Atli dive into how contracts and call transcripts reveal critical revenue signals that CRMs often miss, why reps shouldn't waste time filling in picklists, and how a “friction-first” approach to automation actually drives adoption. Whether you're in RevOps, sales leadership, or just AI-curious, this conversation offers a practical look at building smarter GTM systems without slowing your team down.

00:00 Introduction and Guest Welcome

00:13 Keith Rabkin's Background and Overview of PandaDoc

00:42 PandaDoc’s Framework for Implementing AI

01:50 Addressing Sales Workflow Friction with AI

03:46 Building vs. Buying AI Tools

05:14 Examples of AI Implementation

06:52 Future of AEs with AI

11:40 Encouraging AI Adoption in Sales Teams

14:02 Ownership and Impact of AI in Sales

19:18 Hiring for AI-Driven Sales Teams

22:55 Conclusion and Final Thoughts

[00:00:00] Emir Atli: Hello everyone. Welcome back to the second episode of AI for Go-to Market. I have Keith Rabkin from PandaDoc here. Keith, welcome to the show. How are you?

[00:00:10] Keith Rabkin: Doing well, thanks for having me, Emir.

[00:00:13] Emir Atli: Of course. Do you wanna do a quick intro?.

[00:00:15] Keith Rabkin: Yeah, sure. I'm Keith Rabkin, the president of PandaDoc. I've been here about two and a half years.

[00:00:20] Keith Rabkin: PandaDoc is the leader in SMB and mid-market agreements, so we help you close deals faster, get your contracts signed, and extract useful information out of those contracts. Prior to this, I spent three years at Adobe, and before that, 10 years at Google.

[00:00:37] Emir Atli: Awesome. Cool. It's our first time meeting as well, so it's very nice to meet you.

[00:00:42] Emir Atli: The, as I mentioned before this is AI for go-to market. So we talk about how other leaders are using AI and their go-to market motions, especially on the sales and marketing side and account management, customer success side. Right now to pick things up, how are you using AI in day to day with your team?

[00:00:58] Keith Rabkin: Yeah, we use it quite a [00:01:00] bit. We tend to have a, like a framework for how we should apply AI. There's so many different ways that you could use it and, instead of just picking a spot, what we've done is tried to take a top level view of all the places we have friction. In our go to market processes, like my ultimate belief is we hire really smart people.

[00:01:21] Keith Rabkin: We want them to focus on the things they do best to drive the revenue and customer facing outcomes of our business. What's taking them away from that? And we start to look for those places of high friction in manual tasks and then apply AI in those places. And by starting there, we can then figure out what the right tool is instead of using the tool and jumping into.

[00:01:45] Keith Rabkin: How that tool solves a particular case. So that's the overarching approach. Some of the things we've tended to focus on are what's taking our sales team away from spending time with customers, 'cause if they're not selling, they're not bringing in revenue. [00:02:00] And so a lot of that is inputting data, finding information on prospects, and then carrying that data between systems. So we've been looking at that AI note taking, which I think everyone is doing, but more importantly, how do you push the information between systems? And so using automation there's all sorts of tools out there to take what we hear from customers, push it into systems that then make it accessible for other teams and for our data teams to find information and patterns on.

[00:02:30] Emir Atli: There's always a like difference in opinion between like where sales leaders are thinking these are things that our reps are wasting their time on versus what the reps think. And usually in sales, most sales reps like AEs and SDRs are resistant to change one. How do you actually find out what they're wasting time on?

[00:02:50] Emir Atli: And two, how do you like break through that resistance to change?

[00:02:55] Keith Rabkin: I tend to believe what my reps are saying. So if my reps are telling me something is [00:03:00] taking their time, I definitely wanna understand that and take it seriously. So we do start with what the reps are saying from the field and then validating it, from a leadership and ops perspective to see how we can improve it.

[00:03:12] Keith Rabkin: I agree that reps are resistant to change, but ultimately. You have to win them over. And we've actually had cases where the reps are banging on our door for change with certain tools that make their jobs easier. And to me, that's a real win. So I try to find that like win-win situation where the rep wants something, we believe that something will make them better at their job and then actually roll it through knowing that they'll adopt it.

[00:03:42] Emir Atli: Yeah. What do you think the right of right order of action is? Do you first, I don't know, build like a go-to-market workflow? Then find out what is taking more time than it should be then kinda go out and buy tools to automate those or what is like your kinda like workflow there? Do [00:04:00] most people, for example, I am on the camp that you first need to build stuff internally, but it's like a $20 chat GPT instance where you can build custom gpt to kinda like test your validations versus some people just go out and then buy like a tool, they implement it in a couple months and then they put into their workflows.

[00:04:17] Emir Atli: What do you think about that?

[00:04:19] Keith Rabkin: Unless you're a really new company, you're gonna have some existing workflow. So I tend to start with those workflows, figure out what the friction points are, and then figure out the right tool to solve it. My thinking has changed a little bit. We started off thinking maybe there are like tools on the market that we could go buy that just solve the end-to-end workflow.

[00:04:40] Keith Rabkin: And by experimenting, I think what we're learning is that it's better to build your own or DIY, the approach by finding point solutions, whether it's chat GPT or something else, and then connecting it with scripts that will push the information. I think that is, if you have a good rev ops [00:05:00] team, a better way forward than buying like a, not a heavy to implement tool, but a tool that's gonna require some implementation be fairly expensive and maybe not that customized to your existing like way of doing business.

[00:05:14] Emir Atli: Are there any examples that you can tell us without giving the secret sauce about the use cases that you automated or tools that you built internally?

[00:05:22] Keith Rabkin: There's, like I mentioned the note taking. There's so many note-taking options out there. But I think finding ways that you can use existing tools.

[00:05:30] Keith Rabkin: Like we use Gong. How can I just use Gong? Gong's got great AI transcripts. How can I push that information straight into our CRM with a script and at the level of detail we need, like we use a sales methodology. Gong does support that we can have the transcript like broken out into the sales methodology.

[00:05:49] Keith Rabkin: And then I need that pushed and structured in my CRM in a way that it can break that information down or closed lost reporting. Today, we like scan thousands of gong [00:06:00] transcripts but we've gotta manually pull all this information out or like custom script it with cps and we're starting to do that.

[00:06:07] Keith Rabkin: But it becomes really powerful when you can act on this large corpus of information. I don't need to go buy like a closed lost tool, I can just pull that information in aggregate, run an LLM model on top of it, and then spit it out into a format that digests all the information that was in all these things, and like, why does that work?

[00:06:25] Keith Rabkin: This is a friction point that maybe my reps don't know about, they have to fill out closed lost or they have to dq things and it's a pick list. Your pick list is always only gonna be as good as how you set it up to be. And so you lack some of the fidelity that, your reps would put in the notes, but frankly, like they just don't have the time to put in the notes.

[00:06:42] Keith Rabkin: So this solves that problem for us, and we're getting much richer information from all these transcripts of every single call that we're recording.

[00:06:52] Emir Atli: Do you think a question that I'm thinking about a lot as a sales leader as well is I don't know, if you think about what makes a great AE.

[00:06:59] Emir Atli: It [00:07:00] is like following your sales methodology, finding the gaps, then working on those gaps, creating an account plan, multithreading, stakeholder maps, all that stuff, and then running a really good sales cycle. If AI is able to, like your example, for example, AI is able to understand the gaps in the med picc, push them into the CRM.

[00:07:18] Emir Atli: Then basically probably at some point it will also be able to guide AEs into the right direction to fill in the gaps and also help the AEs understand like who to reach out to at the company, who, how to, all that stuff. What is gonna make a AE a great AE in let's say four to five years, if those gaps are all filled with AI.

[00:07:38] Keith Rabkin: Some people believe that you won't need an AE in four to five years. I don't know if I'm there yet. I think people, at least for now, although five years is a long time in tech world, still like buying from humans and there's something about reading the room. The best AEs aren't just the ones who follow the sales methodology.

[00:07:53] Keith Rabkin: They're the ones who understand the customer, the mindset. They know how to negotiate. You can get a tool that [00:08:00] can do all the multi-threading and all that stuff. No problem. It's that like human interaction that I think really separates the best AEs. I think the piece that is missing maybe is like a real co-pilot.

[00:08:13] Keith Rabkin: I. And I'm starting to see pieces of this, like outreach has kaya that will listen to the calls and give information in real time. We have a tool called Tribble that's really powerful that our AEs can go into Slack and ask a question and it'll search our corpus and actually pull up information about anything in real time about whether our product supports that.

[00:08:34] Keith Rabkin: I think if you can combine those, you're gonna end up with something that lets AEs have a real time conversation that's much richer. The other thing I think could be really interesting is something that customizes the demo and some of this exists, but not at the level we need. Like the best AEs are selling by understanding customer pain.

[00:08:54] Keith Rabkin: And when you show a generic demo, like it's helpful, you get a sense for what the software will do, [00:09:00] what the product will do. But when it's really customized and tailored to the unique use case of the individual, somehow it enlightens them as to what might be possible. And I think it's that. What is possible that really helps sell.

[00:09:14] Keith Rabkin: And so finding a tool that will do that I think will be a game changer. So those two things, how do I answer questions in real time with confidence in detail? And then two, how do I show what this actually looks like for the customer?

[00:09:27] Emir Atli: So you think like the, I dunno if AI can, I fundamentally believe like the human interaction or like understanding customer pay and all that stuff makes an AE a great AE, but there are a lot of other aspects that they also spend time on that basically contributes to understanding the other prospect or understanding their pains and all that. So you think the AE that is doing really well, hitting quota over achieving their quota is the one that's gonna focus more on human interaction and four to five years, even if if we have AEs still.[00:10:00]

[00:10:00] Keith Rabkin: I, yeah, I mean I do, but I think there's there's different paths to success. I think it's, I shouldn't say there's one size fits all. 'cause I have great AEs who are masters at that human interaction and they just know what a customer is looking for. They know how to work a deal, they know how to structure their pipeline.

[00:10:15] Keith Rabkin: They tend to be the ones that are like actually the worst at updating their CRM. And it drives me and my sales leader a little bit crazy, but they like, they keep it all up here and they like plan their deals ahead of time. They know when it's all gonna close, but then there's others who are truly mastercraft outbound. And they know how to pull the enrichment data. They know how to multi-thread like they're masters of breaking in and creating new pipeline. And so I think there's room for both of them. And they'll use AI in different ways. So I think that's, I don't think there's gonna be a one size fits all, but I think people will get creative in terms of these templates.

[00:10:52] Keith Rabkin: I don't know if anyone in any sales leader has ever done like an overview of the different kinds of sellers, but I think that'd be really [00:11:00] interesting. Yeah. Because I think there's definitely these archetypes of, Hey, this is the person who just like is the natural seller and this is the person who's like the mastercraft of the process.

[00:11:09] Keith Rabkin: And the really good negotiators and they all have different ways of getting to P Club. I think that'd be a cool little podcast to do

[00:11:17] Emir Atli: if you're, yeah, if you're allowed to have only one of those types in your sales team, let's say you have a sales team of a hundred, all of them are one type.

[00:11:24] Emir Atli: Which one would you choose?

[00:11:26] Keith Rabkin: Oh, that's so hard. We're mostly inbound today, so I probably wouldn't put the mastercraft outbound or even though like inbound does require mastercraft, I think it's probably the relationship seller today. But I reserve the right to change that.

[00:11:39] Emir Atli: Okay. And how do you, let's say you don't, you might have tools as well, but you still would probably want your team to solve their problems on their own without needing to have a new tool or without having their sales managers or yourself in the room.

[00:11:54] Emir Atli: How do you incentivize that with the team?

[00:11:57] Keith Rabkin: Yeah, so one of the things we did [00:12:00] when AI just really started picking up was we codified it. We put that AI was gonna be a cornerstone of our company strategy, and it wasn't just in terms of making our product better with AI, which we're in the process of doing and I think is really powerful, but we wanted it to matriculate into every single part of how we run the company from an operation standpoint. And so I think one of the first things you have to do is make sure everybody knows that this is part of your charter, this is part of the operating model. And it's an expectation. It's an expectation for leaders, and it's an expectation for ICs, and it's going to be something we experiment.

[00:12:40] Keith Rabkin: Like I don't expect that we're gonna get it right all the time. But this is something I learned back at Google, whatever, 15 years ago, you launch and iterate. Get it out there, play around with it, figure out what works, what doesn't work, shut it down, move on and try it again. And so we expect that we freed up some budget.

[00:12:59] Keith Rabkin: [00:13:00] We looked at things where we could, shut down things that weren't performing and reallocate the budget to experimenting with ai, and it's really paid off. It's led to a proliferation of tools in all sorts of different areas, enrichment, outbounding, messaging, note taking, and I think that proliferation is really good in the short term because it encourages experimentation and you're only as good as what you try and what you know.

[00:13:27] Keith Rabkin: The challenge is later you do want to, solidify around areas of expertise. 'cause otherwise you've got like a little bit of chaos. You've got your data floating everywhere and you maybe have some legal challenges on that front. And then you can codify these things and create efficiencies around them, build them into these processes.

[00:13:44] Keith Rabkin: Like without what we were talking about, it's hard to have a DIY system of pushing information from seven different note-taking tools to your CRM, but one tool. That you align on really easy to do. So that's it. It's, I think it's expectation setting and then [00:14:00] creating the budget for it.

[00:14:02] Emir Atli: Who owns or who's gonna own AI internally in your opinion?

[00:14:06] Keith Rabkin: I think, I don't think there's one owner is my answer. We definitely have a C-level sponsor for it. Our CTO is the sponsor for AI use in the company and spending a, 80% of his time thinking about how we use AI. But then within the revenue org, we have the rev ops scheme thinking about it.

[00:14:27] Keith Rabkin: I think about it all the time like it's fundamentally a game changer for the world. And I don't mean that in a hyperbolic sense. It just like really is a game changer. And if everyone's not thinking about how to use it most effectively, we're gonna fall behind.

[00:14:42] Emir Atli: And traditionally, sales teams only spent their budgets on CRM, like gong, like a call recording software, like the usual, outreach, those types of tools. And they weren't really like marketing, for example, marketing has 10 x more tools than a sales team. Do you think that is [00:15:00] gonna change with the AI and the new use cases for the sales seems to be more efficient.

[00:15:05] Keith Rabkin: I do, but I don't think it has to be a question of expense. A lot of these tools are very reasonably priced and I think AI in general is gonna be deflationary and bring down costs. So I think there's an opportunity for sales teams to adopt more tools but connect them in a way that makes them highly efficient.

[00:15:27] Keith Rabkin: And that does require some time to set up whether you're using clay or you're, whether you're using Zapier or make but finding those ways that you stitch things together built on the big tools of your stack, whether that's Gong or your CRM or in our case PandaDoc .

[00:15:42] Emir Atli: And PandaDoc is in Go to Market Tech, and it's one of the.

[00:15:47] Emir Atli: Probably oldest and more, more like established categories in go-to-market tech. How do you see go-to-market technology changing with AI? Because what I'm seeing in the market right now is everyone is, at some point, [00:16:00] like any type of marketing tools competing with any type of marketing slash sales tech and sales tech is competing with MarTech.

[00:16:07] Emir Atli: It's every single tool is adding more and more features all the time because building technology is easier. How do you see the future of go-to-market tech?

[00:16:16] Keith Rabkin: So I think everything comes back to how does it actually improve the life of our sellers and customers. And so I think because AI is allowing so much to be possible with fewer tools, you'll see a decrease in the number of tools is my hypothesis.

[00:16:34] Keith Rabkin: As a lot of these tools like incorporate some of this. So like we mentioned, Gong already does so much for us but it's like collapsing things that we used to do in other systems. I. So now we use Gong for our forecasting. We use it for our conversational intelligence. We use it for pipeline management.

[00:16:52] Keith Rabkin: These are things we used to just do in Salesforce, and I think it's the same with PandaDoc. PandaDoc historically was a proposal tool that helped you [00:17:00] get the best looking proposals out and get the signature back as quickly as possible. But with AI, we're starting to give you insights into which proposals work best.

[00:17:09] Keith Rabkin: We are doing smart workflows that help you route for approvals, meaning you probably don't need a deal desk in the future. These are things that you know fundamentally change how go-to market teams work, make them more efficient, which then helps you serve your customer better.

[00:17:26] Emir Atli: When you say we will see fewer tools, do you mean if maybe Penoc for example, taking more and more ownership of other functions in go-to Market and then Gong taking more ownership of other go-to market functions, like centralizing in a few tools?

[00:17:41] Keith Rabkin: I think it's, I think that's true. I think I haven't thought as much about centralizing across functions, but within a function for sure. Like I think it's pretty, it's not necessarily from an AI perspective, but in the past you needed a forecast tool and a conversational intelligence tool and a CRM.

[00:17:58] Keith Rabkin: Do those get [00:18:00] unbundled? Or I guess do they? I guess they get bundled. But do they get bundled together and do things that used to rely on structured data that was human input, like a CRM become less and less important? We believe that some of the most important information is actually in the contract itself.

[00:18:16] Keith Rabkin: And with LLMs you can pull this information in real time. Same thing with. With like conversational intelligence, when I've got a recording of the call, it's much less important what a rep puts into the CRM. If I can pull in real time and access the information that's in all the calls I've ever had with a customer, that's much richer and meaningful information than whatever my rep chose to put in the moment in the CRM.

[00:18:41] Keith Rabkin: So I think that's what I mean, like how do you go to the source of record of the information, which is either the contract, the conversation, the pricing, those things become inherently more valuable when you can access them at real time and at a level of fidelity that doesn't exist today. Because today most of the information is [00:19:00] input manually.

[00:19:01] Emir Atli: Yeah. Yeah. Contract, contracts is a really interesting space 'cause as you said, there's a lot of missing information in our CM that is actually in our CRMs, which. It's kinda impossible without El s to read. That is a big pain point for sure, for a lot of companies, including us. And my final question to you, which is also something that I'm thinking about every single day, is how does your hiring change with this shift, like how do you find people who are gonna adopt AI or curious is going to excel in this new environment?

[00:19:34] Keith Rabkin: So if you ask my team, they will say there are very few things more important to me than hiring the right talent. I'm a big believer that when you get the right people in the organization, they make good things happen. And so for making sure somebody can thrive with AI, I. One of the things I looked for before AI and I'll continue to look for is people who are curious.

[00:19:54] Keith Rabkin: Because if you're curious, you start asking questions and if you're curious, probably like to experiment or you like to [00:20:00] try things, especially things that are like relatively easy to get started with and that allows you to just, start playing around with these things and finding what works. I think the other piece of it is comfort with ambiguity and making sure that you're okay with the unknown.

[00:20:15] Keith Rabkin: So I like to test for how people can handle if they don't know what the solution is or if things change. That adaptability is really important because the AI solution of today, it's probably gonna look really different than the AI solution in the year. And you've gotta be okay with things changing.

[00:20:30] Keith Rabkin: Whether it's a change in how many people you hire or the tool set you use, or how the process works. The people who are really resistant to change I think are gonna be the people who struggle. And then I put all of that together with, and, talent that likes to challenge convention.

[00:20:46] Keith Rabkin: They don't want things to work the exact way it's worked in the past. And to me, those are the people who will thrive and honestly have been the best employees that I've hired.

[00:20:56] Emir Atli: Do you have any best practices or anything that you'd like to [00:21:00] ask to test for these things? Any favorite interview questions or test case studies?

[00:21:06] Keith Rabkin: Yeah, arc, yeah. I would say from a question perspective, I'll try to ask try to put them in a situation like, Hey, I've asked you for this information, or I've asked for this project halfway through. Information comes out that suggests, we maybe only have a 30% chance of being right. How would you adjust?

[00:21:28] Keith Rabkin: 30% still decent, but does this person want to change? Are they open to it? Sometimes that's hard to understand in an interview question. You could do it in a case, like in the case, maybe not make the case about changing, but change the information in the middle of the case if the person gets annoyed with you.

[00:21:45] Keith Rabkin: I haven't actually done that, but it's a decent idea. I like asking in the case interviews, if the person wants to meet with me to ask me questions. Because you'll see whether they continue to ask questions are they fine with the first [00:22:00] level of data that you gave them and they just operate on that?

[00:22:02] Keith Rabkin: Or do they ask more questions? Or sometimes I'll even tell 'em like, don't worry about answering that question, but what else would you want to know? Because then you just start understanding whether they're a naturally curious person.

[00:22:14] Emir Atli: Yeah. Yeah. The. Interviews are getting interesting and interesting, I think.

[00:22:18] Emir Atli: I find, I also find a lot of different case studies and interview questions and also different tests that you can run with AI as well. Like you can put them in, I know, build me a GPT that would automate X, Y, Z and then show me results. And I think what's getting really interesting is. Like time goes by and then there's sales leaders that need to adapt, but also I think it's very exciting because there's I think a new generation of talent that are basically really comfortable with ai, really comfortable with the tools they can do out of the things that they need by themselves, which is extremely exciting for me as a sales sitter.

[00:22:54] Emir Atli: Yeah. Awesome. Keith, thank you very much for your time. This is awesome. Is there anything else you [00:23:00] wanna

[00:23:00] Keith Rabkin: No, that was, that went by super fast, so I appreciate it. I think about this every day, so it's fun to talk about.