Invictus Guild Summit: Building your AI Pipe Gen Engine: How Agents & Avatars lead to Maximum ROI

Hello, everyone. Wanna introduce Scott Kane, one of our newest Guild members. I will let him give his bio. I’ve had the fortunate or unfortunate of working with Scott for years over time at Salesforce. com in the early days. I also brought him in as CEO of Bitly when I was on their board. He’s done a fantastic job, and he’s gonna speak on what you see on the screen.

And one of the things you’ll see a little bit more of is we continue to invest and evolve the Guild, and as folks’ tenure and time on the Guild come to an end, we are starting to add new members. We actually have eight new Guild members in the last 90 days. We’ll be announcing a few more in the coming few weeks, that natural evolution as we continue to align the Guild to the most pressing needs of our portfolio. So, with that, I’ll turn it over to you, Scott.

Right on. Thank you so much, Eric. And thanks, everybody, for joining. Really excited to share this content today. I have been on an AI- first journey in sales development for some time, as I’ll share, and really excited to share some of those learnings with all of you. I’m gonna ask that we hold all the questions until the end. We do have a lot of content here. I believe we will have time for questions at the end.

If we don’t have enough time to get to your questions, I’m more than happy to take questions offline, schedule time with you, and go through and talk about how all of this actually aligns with your particular needs and your situation as well in your portfolio company. So, without further ado, we’ll dive on in. So, as Eric mentioned, he and I worked together at Salesforce way back in the day. This is now my entire career. I’m just going back 20 years. But it’s definitely the most relevant.

So, I started at Salesforce when we had only about 1, 000 employees and there were about 300 million in revenue. So, it was still a really small company at that point. And I built out the first marketing operations team there. So, I was very systems and process forward from the beginning. And then eventually I moved over into sales development there. We had an inbound team of SDRs that I grew from about 60 to 180 in a mirror. I then went to Google Cloud and I built out the global sales development team there.

When I first got there, I only have one SDR following up on all inbound Google Cloud demand because that’s how much there was. And now it is, I think, a $ 60 billion business, which is incredible. So, I spent some time there, did go to Bitly where I ran all of go- to- market and as the COO and was the interim CEO for a little bit as well. Sales marketing, sales dev, customer success, RevOps, really the entire span of it. And then I decided to go back to sales development again with Databricks and was there for about four and a half years overall.

Started there when they had about 800 million in revenue, left at the end of last year at 5 billion in revenue and really grew and transformed the organization there. Started off as a very human centric sales development model, moving from 60 people up to 300. And then we began to work with AI and I saw the writing on the wall last summer and thought, hey, I really like, the future of sales development is no longer gonna be human centric, it’s gonna be AI first.

And so gave them my notice and decided, how do I go and start to really help other organizations start to lean forward and build out sales development, go to market with agents and AI first. And so that’s what really has brought us here together today. So let’s dive in, let’s talk about, what are we gonna talk about today? So first I’ll lead off and provide a little bit of a background on how AI can maximize the ROI of your pipe gen engine. Talk about some proof points, what did we see at Databricks?

And then I’ll go in and talk about, well, once you believe that you should be using AI, then what does the sales dev tech landscape look like and how do you go and choose from the different options that are out there? And then I’ll give you a specific blueprint for getting started with AI. What are the things that you need to go and do right now, especially on the sales development and outbound side in order to make sure that you can make a successful choice and then begin down this road?

Well, a few key takeaways and then I’ve got some tangible next steps for you here as well. So let’s talk first about the problem that everybody has, which is like everybody on this call is trying to build an efficient and effective pipe gen engine that is gonna maximize ROI. You need pipeline in order to grow your revenue, pipeline is oxygen. And really there are a few, you can think about this in a couple of different ways. So stepping way back, you have net new business and you’ve got existing customers.

On the new business side, you’ve got inbound leads, outbound prospecting, partner referrals, and then existing customers if you’re thinking about renewals, add- on and expansion. Today, I’m really gonna dive in and talk about outbound prospecting the most because that’s really where AI has been having the biggest impact on the pipe gen engine so far. It’s one of the areas that AI has been best at in terms of going and augmenting and eventually replacing a lot of the things that human SDRs and BDRs have been doing up until now.

So, but what I do want to point out is that, look, there’s been an outbound evolution that’s been going on for over 40 years, and AI really is just the latest advance in that overall evolution. So we go all the way back, we’ll put ourselves in the Glenn Gary, Glenn Ross days, coffees for closers, always be closing. Look, at that point, the AEs handled all aspects of the sales cycle. You know, they were using the Rolodex, they’re making phone calls, pounding the street. They were handling all the inbound leads, doing all the outbound prospecting.

There were no SDRs, there were no BDRs, it was just the AE. And then we began to see a shift in the 90s when Oracle and CA began to build out the inside sales model. They started to think about, hey, how do we do inside sales teams? How do we do more outbound calling? How do we expand those capabilities? And then really a huge transformation happened in the early 2000s when Eric and I were at Salesforce.

And that’s really where Salesforce popularized the SDR, BDR, AE split to do dedicated roles to say, hey, let’s specialize and let’s have the lowest cost, lowest cost resources doing inbound follow- up, do triaging, and then save the higher cost resources for the AEs and let them go and close deals and really work with our prospects and with our buyers. And then the 2010s brought on a lot of evolution and a lot of advances on sales engagement.

Now you had Outreach and Sales Loft and Gong, and they really helped to make those SDRs and BDRs more productive, but you still had to use SDRs and BDRs to go and do that work. And now we finally have come to the 2020s and now we’ve got companies like Relevance AI, OneMind that really are leveraging the fact that now we have LLMs, you’ve got autonomous agents and even avatars that can really go and execute the majority of a lot of these very repetitive sales development tasks that we have.

And interestingly enough, we’re now getting to the point where the SDRs and the BDRs are, you still have the role, it’s just that it happens to be AI that’s doing the work and you’re starting to get back to the point where the AEs can work with their AI, SDRs and BDRs and basically kind of run the entire process themselves. And again, I’ll talk a little bit more about where is that going and what were we seeing at Databricks. But fundamentally around outbound, there are still just three things that you need to do.

As long as humans are still buyers, you have to find a way to connect with your buyer, you have to find a way to inspire them to engage with you and your company, and then you have to qualify the opportunity. You have to work with them to identify, does it mutually make sense for us both to spend our time evaluating and then moving forward with this evaluation in order to figure out whether the solution can actually help your buyer overall.

But again, the game changer here overall is that now AI agents, LLMs, avatars, they can actually go and execute a lot of this work. And again, we’ll talk a little bit more about that. So I’m gonna take you through my journey at Databricks in becoming an AI first sales dev and go- to- market leader. And, you know, ChatGPT came out in November of 2022. It was about nine months after that that the CIO at Databricks went to the CEO and said, look, I can’t get rid of half the BDRs with an LLM and tools based on that. He never talked to me.

He didn’t understand anything about my workflow. He just said, I can do this. And so I said, well, hold on, let’s start talking about this. I don’t believe you. And he was wrong at that point, but it really kickstarted the relationship with IT to start going down this path. And because we were Databricks, because we’re a data and AI company, we realized that we had to be on the bleeding edge of AI and agents in order to figure out how do we do this so that then we can go and share that journey with other customers of ours.

So about six months later, we finally, we had an LLM tool that was able to personalize follow- up on inbound leads. And it finally reached parity with what the BDRs were doing by themselves. And so we’re like, okay, well, we have a proof point. Now let’s figure out, can we go and take it even further? And then later on that summer, we launched an initiative to figure out, hey, can we do sales development work with, can we do touchless and have no humans in the loop? And how can we think about doing this with humans in the loop and help to augment them?

And so we went out, we took a look at 20 different offerings that were out on the market. Ultimately, we ended up narrowing it down to just one that we were able to do a POC with. And that was with a company called Relevance AI. And that POC was started just about this time last year that we went and kicked that off. And by the summer, we had, we built a team of three AI agents in Relevance that could go through and do, they could conduct research, write custom outreach and LinkedIn messaging.

And I’ll show you the results in a few minutes, but we realized that we were onto something and the initial results that we had were so good that we didn’t even do a huge POC. We just said, let’s just start rolling this out.

And now where Databricks is, is they’re now beginning to launch inbound in relevance to globally, because they’ve realized how much time they can save on the hundreds of thousands of leads that are coming in by leveraging an agent to go and do that. And by the end of this year, they’ll have outbound automation globally rolled out as well. And I will bet you that in 12 months, they will have half the BDRs that they have now. So the CIO was right.

He was just about two and a half years early, or I guess it’ll be three and a half years early by the time that we eventually get there. But let’s talk about, so what did that agentic AI workflow look like at Databricks? So I wanna call your attention over on the left- hand side. It does still require input. You need to have a lot of things that have been identified. So you have to have your ICP accounts. They have to be prioritized by score.

You need to be able to identify what are the accounts that you’re going after in all of this, because you still have to go after those most important accounts first. You have to have your personas identified. You have to know who is it, who are the buyers, who is it that, what do they care about? What problems do they have? And you have to be able to go and identify that persona for each of the different segments, each of the different verticals that you might be going after.

And you also have to develop your own, your value proposition, and you have to differentiate that by vertical, because if you’re going and reaching out to one particular persona, you have to make sure that your value proposition is very relevant to them. And you also have to have customer references and proof points overall, in order to be able to prove that you can do what you say you can do.

And ultimately, these were the six things that we automated and were able to save a tremendous amount of time for our BDR. So the first thing was contact discovery. And you’d look in Salesforce to see, okay, who do we have as contacts? And then we had an agent that would go out to LinkedIn, go out to Zoom Info, go out to Sumble, find and add more contacts. We had not, and the agent would then go and enrich all of those contacts that we currently had with brand new data, any new data that was, that we could go and update the existing contacts with.

We would do account level research. We’d go out and search Google, 10 Ks, any reports for any unique insights and key triggers, especially for Databricks. If anyone had data or an AI initiative, that was something that we could go and lean into. We’d then go and do contact level research in order to get fodder for YUYU Now Messaging that we would go and wanna reach out with. We’d look at LinkedIn for that. We’d look at blog posts, anything that was out there as a digital footprint on the individual person that we were reaching out to.

And then we had a different agent that would actually go and write all the messaging and take all the research on the company and on the person, specific messaging that would end up aligning with that. And then we had another orchestration agent that would go and lay all of that into a sequence. And we didn’t just have personalization on the first step. We have personalization throughout the entire thing. And so all of these things blew my mind. And this is what really inspired me to say, look, the day of human BDRs is over.

You still will want some BDRs in order to be the A’s of the future, but the volume of BDRs that you’re gonna need is gonna go down dramatically. And here’s the way that I knew that. Number one, when we rolled this out, we immediately began to save 10 hours per week per BDR on all of that repetitive time that they had been spending doing the research, scraping the contacts, writing the messaging, 10 hours per week were gone. And of course, immediately the CIO said, great, now we can get rid of 25% of the BDRs.

I was like, hold on, we’ll get there eventually, but we’re not quite ready yet. We had a 33% response rate lift in the AI power personalization versus just some of the BDR written outreach. And then on the inbound side, we were able to dramatically reduce the overall speed to lead. This is with humans in the loop. We were able to cut it in half.

Eventually they’re gonna get to the point where they just are able to immediately have follow- up on all of that, on all of the inbound, because you no longer have to have an SDR checking and saying, all right, is an AE already engaged or not? What’s the last time that we reached out to them? All of those things can be automated and can dramatically save the time and manpower that it has historically taken for that.

And so, as you are thinking about moving forward with an initiative here, really the metrics that matter, the MTM that you need to be taking a look at here, you know, start immediately with, well, how many human hours has this saved if you’re replacing people that you’ve got?

And also, if you’re comparing to what you currently have in place, you need to take a look at the response rates, take a look at the engagement quality, how many people are interested, really go and baseline that and make sure that as you’re rolling something out, that you’re at least able to meet your existing baseline before you end up like really pushing something, pushing something out and going big. And in the longer run, really this is all about AE productivity.

And this is all about ensuring that you’re able to get more pipeline through to the AEs and that the existing AEs that you’ve got can close more deals because they have more pipeline that’s coming in and it’s more highly qualified pipeline as well. Okay, so hopefully I’ve convinced you that you can maximize your ROI or your pipe gen engine by starting to introduce agents and avatars. Now let’s talk a little bit about what does that sales dev tech landscape look like today?

And then how do you start thinking about, well, what might I choose and how do I want to move forward with it? So I’ve broken it down into four different categories here. So first I’ll start with the AISDR platforms. So these are kind of all- in- one platforms that are designed to fully replace with like AISDR and Artisan, your BDR team or augment your sales dev function like a Reggie AI by doing auto researching and initiating that outreach.

So these are platforms that you sit down, you configure, you put all of your playbooks in, you put all of your responses in, it takes a few weeks to go and set it all up. And then you begin to go and roll that out and you don’t have to have BDRs, actually in the loop for that overall. So that’s one, and by the way, these platforms have really come a long way over the course of the last couple of years. When we were looking at them, they were effectively unusable, even just about 18 months ago.

And now they’re really starting to get a lot of traction in the market. The next one is thinking about workflow and intelligence orchestrators. So these are tools that allow you to go and provide the logic and also build out agents in order to be able to go and execute on a lot of this outbound prospecting. So that’s like relevance AI, that’s what I’m most familiar with because I have been a customer of theirs.

And as I have talked to some of the portfolio companies, I understand that some of you are using Zapier, like NoviLabs, I know Peerspot is currently using Make. These are all ways to go through and orchestrate, connect through to different applications and start to build agents. The agents for Zapier and Make are a little bit further behind what I’d say relevance is right now, but it’s all speeding up. And as the frontier models end up changing and adding additional capabilities, I’m sure that they’ll speed up pretty quickly.

Clay is a little bit different, but I have it here as an intelligence orchestrator. They’re kind of a Swiss army knife. You’re able to do prospect sourcing, enrichment, AI research. You can also do personalized message drafting there, as well as pulling in intense signals. They’re starting to move in the direction of a relevance in Zapier and Make as well, in order to be able to go and do like full- on orchestration. But they have come from a place of really being more of a data intelligence operator. So, but keep watching Clay.

Clay is really emerging in this space. Then down on the lower left- hand side, you’ve got Agentic AI within major platforms. So this is if you’re deep in Salesforce, if you’re deep in HubSpot, or if you’re in Apollo, then you may want to take a look at some of the Agentic AI capabilities here. I was talking to Dave Kanellis over at Axia the other day. He mentioned that they were trying to really use some of the capabilities in AgentForce, and it wasn’t really doing what they needed to do.

I can tell you from experience and working with Invictus that Apollo, they have some Agentic capabilities that are a little kludgy, but again, they’re moving along pretty quickly. So this is really going to be, if you’re deep in one of these ecosystems, and this is really what you’re focused on, there may be certain use cases that you need to go and build out in these different platforms. And then finally, I want to highlight the AI avatars and some of the human- like sales reps.

This is a little more bleeding edge, but as these capabilities and as these technologies and solutions continue to emerge, I continue to keep your eye here, especially if you have a lot of inbound demand where you might want to have one of these avatars out on your website. These are solutions that are trained on all of your customer- facing data, where you can have someone come in, have all of their questions answered, and you can also proactively qualify those prospects while they’re on your website.

These are also solutions that you can eventually go and put into emails and offers out to your buyers and prospects and allow them to engage with your company at any time. They can do it in their pajamas if they want to, 24- 7. And these are very capable. They’ll continue to go and move forward. And it’s a really exciting space that I’d recommend that you all keep your eye on.

So now let’s think a little bit about, okay, well, when would you choose one of these and when might you want to steer away from them? So Artisan, AISDR, Reggie AI, the AISDR platform.

So this is when you have a sales motion that is really high volume, very transactional, or maybe if you’re mid- market and below, and where, again, you’re gonna need to have a well- defined ICP and proven messaging for all of these things, but especially here, because if you go and take those things out at scale and you haven’t nailed them, then you’re really gonna have a problem.

This is an area where you want to avoid here if your sales cycle is highly complex, enterprise level, a little bit more consultative, and if your ICP and messaging are still being figured out. And one thing here to watch out for is kind of to set it and forget it. Just because you don’t have human BDRs in a loop doesn’t mean you can just put this out and never go back and train it again. Again, I haven’t worked specifically with Artisan AISDR. I did take a deep look at Reggie, but this is an option for you if you want to go and take a look at that.

And I think that they are actually capable platforms at this point. They’ve started to get a lot of market traction. So now we’ll go and we’ll talk about the Workforce and Intelligence Orchestrator. So these are the ones that you want to consider when you want to build a custom differentiated outbound model rather than going and buying off the shelf, where you have multiple signals that you want to pull together around intent data, product usage, CRM data, social into one workflow.

And if you’re a really fast growing company, which I know all of you are, that need flexibility as your go- to- market motion is evolving and being able to go in, figure out what’s working, what’s not working, and go and make those changes. This really is a very interesting area to be in right now. Now, you do need some operational bandwidth to be able to go and build and maintain these gigantic workflows that it’s coming down.

Last year, when we were going out and building in relevance, it would take weeks, if not months, to go and get a single use case up and running. They now have a lot of tools that are beginning to rapidly advance that. And I think that all of these, as the agentic AI workspace space is moving forward, they’re all becoming a lot faster to move forward with. And again, you can now begin to do, get some of these up and running. It’s not gonna be as fast as a platform that, as one of the AI platforms that I mentioned just before.

And again, in terms of things to watch out for, these are only as good as the data that are going into them. So you need to make sure that you really are, you’re really tight about your ICP, you’re tight about all of the information that you’re using to put into them overall. So I’ll briefly touch on agentic AI within the major platforms. So again, consider these if you’re deeply ingrained in one of these platforms already, avoid them. If you really need best- in- class AI capabilities, each of these platforms is definitely lagging here.

And then finally, there’s also a lot of AI washing where they’re like, hey, these are AI agents and really you need to go in and make sure that you understand what’s automated versus just a system. And then finally, the AI avatars. This is one, if your product is super complex, would benefit from demo or visual explanation. And especially if you would benefit from a two- way engagement around that, then you might wanna take a look at this.

Also, if you have a lot of website visitors and you wanna be able to handle those 24 seven, if you don’t wanna have to have an SDR that’s always manning that, then this can be a really good way to go, especially taking a look at OneMind and Qualified here. And finally here, really these tools work as best of a broader motion. This is not the first area that I would go and lean into here. It’s probably more of an add- on and something to take a look at a little bit further down the road.

Okay, so let’s get down to brass tacks and talk about how do you get started. So as I mentioned, really your blueprint here, you’ve gotta start with a strong foundation. You cannot skip these things. AI cannot figure out your buyer, cannot figure out your value proposition. So understanding your buyer, fortunately you all have Diane, so you know which companies are in your ICP, but you really have to understand what pain or opportunity do they have? What are their top priorities? What messaging is gonna resonate with them enough to engage?

On the value proposition, how does your solution help them to achieve those goals? How are you uniquely positioned to go and do that? And what proof points do you have that you’ve solved this for other companies that is going to help this company to understand and believe that you can solve it for them as well?

And then finally on the systems and data infrastructure, you’ve got to have a clean CRM. You need to have some sort of engagement platform, whether that ends up being Outreach, Gong, Sales Loft, Apollo, Actively, something there overall in order to be able to go and reach out. Data enrichment platform, you have to have a place to be able to go and pull data in. You’ve got your ICP model with Diane.

And then finally, if you can also get signals data to help you to identify when is the best time to go and reach out to somebody from Sixth Sense or from Clay, then it’s a great place to go and add that in as well. Humans still are going to have to own some of these things right now. Number one, sending individual LinkedIn messages, something they’re going to have to do, LinkedIn and their terms of service still does not allow for that automation. There’s some tools that are out there skirting the rules right now. I would not recommend them.

But what you can do is you can have AI go and tee all of these messages up so that they’re ready to go. They’re personalized. You just have to go and click through them and go and send them out. So it can dramatically decrease the amount of time that you need for that. Making phone calls, assuming that your team is still making calls and calls are still really effective for some companies. The voice AI is not quite there yet at the quality level.

The last time I checked, although I understand that Eleven Labs and Bland are moving along very quickly here. So I would recommend you continue watching the space. It’s something that is probably going to be ready, I would say, within the next 12 to 18 months, if not sooner. And then finally, qualifying non- AI ready prospects. As you begin to go and move down this path, and I would say that today there are probably a lot of people would say, look, I’d much rather talk to a human than talk to an AI agent.

I think that once people begin talking to AI agents, they’ll realize that they never really wanted to talk to a human SDR or BDR in the first place. And they’d much rather be able to go and actually have all their questions answered whenever they want and be able to stop a conversation with an avatar if it is going in a direction that they don’t want it to. But for right now, let’s assume that you did have some of these avatars up and running.

You just need to have an off- ramp for your prospects to be able to go and talk to humans if they don’t want to talk to avatars. Okay, so tangible next steps. One, audit your foundation. Do you have a clearly defined ICP? Leverage Diane here. Is your value proposition proven and quantified? Is your CRM data clean, reliable enough to use? Go through and make sure that you’ve really got your house in order here first. Then just go and pick one use case. Just don’t try and boil the ocean. Don’t try and do everything at once.

Just get one MVP, minimum viable product, up and running, see how it works, and then just continue to go and iterate from there. You’re going to learn along the way. There’s no way that you can know exactly how all of this is going to work when you first begin, but just get started. And then choose your deployment path. Do you want to buy a platform to get up and running quickly? Do you want to go custom with relevance, AI, Zapier, Make?

Or do you want to start hybrid and say, look, we really need to go quickly, but I’m going to want to be able to do an experiment. And then finally, as you’re going through this, really review the metrics that matter. Take a look at the human hour save, response engagement times, A productivity. You need to be taking a look at these constantly in order to continue to tweak your model and as you’re going and building this out overall. And then finally, key takeaway is one, the process hasn’t changed. You’re still doing the same main three things.

Just the execution is changing now. Start with a foundation, then bring the AI. Choose your path strategically. Start where the humans are most expensive. And then finally, just measure ruthlessly, iterate quickly as you go through this. It is something that is, it’s a space that’s constantly changing and is rapidly evolving. And it’s one that I think is a really exciting time for us to all be on and continue to work with. So with that, those are the end of my slides. I think I’m right at 930.

So happy to open it up for any questions for the last 10 minutes or so. Eric, do you want to?

Yeah, that was great. So feel free to just jump in, either pop in the chat or join the call. I think sort of get in front of two things. One, a bit of a new format today, and hopefully people are receptive. The Ask of past Guild Summits has been really focused a little bit more on tactical, third- party usage of apps, etc. So I think Scott did a great job of that today. So we will share this deck. I think Scott has the right to comb some of it. I don’t know if there’s anything in here, Scott, you don’t want in.

But we’re also going to start chunking out a lot of these decks into what we call sort of Invictus best practices. So that if you want to know about the orchestration piece or just a chunk of this, it will all be on our Navigator platform. We will not be emailing this around. So please log into our Invictus Navigator. There’s discussion forums around this. We’ll post the materials in there. And that will be our central sort of collective community that we will continue to use inside of Navigator. And any questions, go forward for Scott.

I’ll jump in. Hey, Scott, good to see you again.

Thanks so much.

Yeah, thanks for spending the time. I think, you know, based on our previous conversation, we’re onboarding a tool called Warmly, and you mentioned, you know, AI agents first talking to human VDR. We’re kind of going to do a crawl, walk, run, and have almost a human- in- the- loop approach when we have a lead. But you mentioned that people are getting more comfortable with that, and I know that’s changing also with the chatbots are a zillion times better than they were four years ago.

Are there any statistics or just based on your kind of, like, feel of the market that people are becoming very successful with saying, like, those chatbot sessions they’re seeing on their web page are driven to a white paper? And are those, you know, starting to become more trusted than having that human- in- the- loop, where people are actually, like, self- driving through the process?

Yeah, so I’ll give you one stat that I remember from my research around OneMind. So OneMind has actually been deployed at HubSpot in their trial flow, and they’re seeing an 88% engagement rate with the avatar in that flow.

first present an avatar

and somebody’s like, whoa, I don’t want to see a face like that, then you can say, oh, let me just get kind of the glowing, like, ball. And then if you don’t want that, you can say, hey, I just want to do chat only. But yeah, I think that there are a lot of, as I’ve talked to a lot of decision makers, a lot of decision makers are like, ooh, I don’t like talking to avatars, so I don’t know if my customers will. But I think that you’ll begin to see that it’s definitely coming based on some of those engagement rates.

Yeah, I think so, too. I think it’s a generational shift, too, where mine was like, there’s no way I’m plugging information into some sort of autopilot. Also, because the prior experience with chatbots was so irrelevant, but they are getting just much smarter exponentially. No, thanks very much. And yeah, I think kind of a follow on to that was just to ask, like, same thing. There’s a lot less pensiveness, I think, now for people to engage with bots. But I also like that option where there could be a human in the loop.

And one thing we got concerned with when we did initial persona tracking and just saying like those, you know, they were always called big brother, just making sure if someone’s looking up a competitor, hey, we can follow up. Any suggestions on time relevance? Like I always say, guys, don’t hit them up in five minutes so they know that we’re tracking all their activity. Maybe I’m old fashioned that way, or people just used to the fact that they know people are always sniffing on their activity.

So I don’t have any specific data on that around how long should you wait. And again, I think it probably ends up being different. If you fill out a form, then, you know, you have permission to get back to them immediately. Yes. And especially if you have kind of a synchronous form where somebody has said, hey, contact me, or, you know, it’s a little bit different if somebody just does white paper. You don’t want to follow up with them immediately because they’re like, I haven’t, I’m just looking at a white paper. I haven’t had a chance to read it yet.

So part of it comes down to on the forms. It comes down to what is it that they’re actually downloading. I don’t have any data yet on if you’re if you’re sniffing and you suddenly see that somebody’s come on your platform. And like how quickly are you following up? I need to think about it. I would think that you’d probably want to do more of a softer touch. It might even be something where if it depends on how many times they’ve been coming on your website.

Get them into a marketing sequence.

Exactly right. Yeah, get them into a marketing sequence. Get them into a LinkedIn ad campaign, something along those lines. If it’s very early and they’re just kind of in the awareness stage, then treat them as though they’re in awareness and don’t jump right to, hey, let’s get married.

Perfect. Okay. Thanks, Scott. Hey, Scott, it’s Andrew.

Good to see you again.

Hey, Andrew, how are you?

Good, good. And thank you for this. Look forward to our next call where we can do a little bit more deep diving. We made I’ll throw this out to the group. We made a mistake when we launched. We were doing outbounding to cold prospects who didn’t know Peerspot thinking we had a reason for doing that at first. And then had the aha moment when I saw somewhere on the interwebs where it’s like, hey, you actually should start out the easier route with people that know who you are. So maybe MQLs that didn’t convert, webinar leads that didn’t convert.

So we’ve actually now flipped and started reaching out to warm people. So I just want to throw it out to the group. Don’t start out your cold outbounding to cold people. Start out cold outbounding to warm prospects. Maybe everybody else had thought about that, but we missed that and all of the other intricacies that we were putting this together.

Yeah, definitely sage advice. One thing that we started off when we were trying to do things totally touchless at Databricks is we weren’t going to super cold people that we didn’t care about at all, just to use that as kind of our test base, but then 100%, if you wanna actually have conversions when you’re launching it, then definitely warmer is definitely better.

Yeah, thanks, Andrew. My question was gonna be around the same thing. So I’m glad you already have data behind this, right? Because with AI and automation, it suddenly makes scale no longer a problem, right? You’ve got the resources, you’ve got the agents that could do it, but now the problem becomes, do I wanna go that broad or do I wanna stick really targeted in your case, Andrew, with the folks that are already showing some interests that I wanna go after versus sending the entire world’s messages?

I wonder, Scott, like obviously that’s gonna be a lever that you can use as you start to try to grow and create more demand and give more of those opportunities to the agents to open up those doors for you. I like Andrew’s idea of starting out with warm. Is there an approach where you start to move into the cold territory and start to say, we’ve used up our warm, we’ve got to start digging deeper, going broader, going further. And then what sort of guardrails would you suggest putting in place?

Yeah, I think part of it has to do with how, I mean, so first of all, I’d even go even, I’d say go even warmer, number one, and begin to, I mentioned, look for signals that are helping you to identify, wow, this is really gonna be somebody who’s potentially hot now. You can now look beyond six cents. You can begin to go and identify very specific things that are happening in the market that may be relevant only to your particular customer base. And those are some of the things you might be able to go and get out of clay.

So the problem with all of this agentic AI is that because it’s gonna be so easy for everyone to go and send all of these personalized emails, then over time, the entire channel is, and both on the LinkedIn side and on the email side, are gonna become more and more saturated and then it’s gonna become even harder to be able to go and get through. And so I would say that you could, yes, eventually, if you run out of warm, then you have to go cold.

But before going cold, I’d say, try and warm up the ones that are cold, maybe through more marketing outreach, because otherwise, you’re just gonna be going and wasting your time and kind of burning through your account base with messaging that is not terribly relevant to them. And then they’re gonna tune you out even more than they would have before.

Thank you.

Sure. Hey, Scott, any different level of seniority, right? So like, hey, we, you know, director level AI worked pretty well with, it was terrible with VP level, never got a C- suite person. Any stats you had around that? That’s where we’re, I’ve ran into in the past.

You know, I don’t have any specific stats. I will tell you that, I will say that, I would say that anecdotally, the higher that you are trying to get within an organization, the more likely you’re gonna connect through a personal connection rather than through an outbound email or LinkedIn. I would say LinkedIn probably is gonna be better than just going and doing cold email. But, you know, those people get hit up so often.

Now, one other thing that I think that is, that we were starting to have some real success with was, and now with agents, you can actually begin to go and scale this a little bit more, is delivering like mail, FedEx, et cetera, kind of going back to, I mean, I’m not saying go back to sending the driver cover and you get a driver if you take a meeting with me, but going and delivering and cutting through the noise and having very specific personalized messaging that is landing on someone’s desk that can resonate with them.

You might start taking a look at that for some of the VPs and C- level and directors.

Super helpful. Yeah. And we’ve definitely found the, AI is better at coming up with gifting ideas than the AE is, right? So that’s super helpful.

Yeah, absolutely.

Hey Scott, quick question. You talked about like Reggie AI for like an inbound and then I guess relevance for outbound. So do we have to like, is it like very secluded in the sense like you have to use both to automate both or like one can do a little bit of both and you just pick like the one that makes most sense for you?

So Reggie, and Reggie first off, Reggie is actually, it’s more outbound focused, but to your question overall, you can, and so Reggie actually will be less capable on the inbound side as of right now, but there definitely are tools, relevance certainly can help you do both on the inbound and on the outbound side. I believe that AI SDR I know has some very specific capabilities around being able to keep up two- way, keep up conversations, kind of going both ways without actually having to bring a human in the loop.

But no, these, the platforms, they definitely have both inbound and outbound capabilities. It kind of just depends, their maturity just depends on where did they come from? Did they come from the inbound space and they’re moving toward outbound or they started outbound and moving toward inbound? But, and by the way, I’d be happy to talk about your specific needs and at some point and help you figure out which of these solutions might be best.

Yeah.

Appreciate it. Thank you.

Sure.

Any other questions?

Yes. Mr. Kane, it’s great to see you. I have a question for you around leadership and ownership of this type of initiative.

What’s the role of the executive team?

Like where should this type of initiative be owned? Is it the business? Is it IT? Can you share some thoughts on that?

Yeah, I’d be happy to.

I was actually at a, I was at a relevance customer event yesterday, and this is one of the things that we were talking quite a bit. First of all, I’ll tell you what worked well for us. And what worked really well for us is that at Databricks is that this was an initiative at the CIO and president of go- to- market level, and it had a lot of visibility on it. And as a result, we were able to make a lot of really good progress. So as we began to build this out, we had somebody from the business side on my team who really understood all the workflows.

We had somebody who was on kind of our growth and innovation team, and then we had IT engaged as well. And they were all working as a team in order to go and build this whole thing out. And to me, and then now where it is, they eventually took the leader that was reporting into sales development, put her into an AI ops role. And so she is now basically create, she’s now owning sales development as a service that is now for all the sales development reps and all the AEs.

But the ownership of the agents that are being built really ends up sitting within the business and within business operations, not within IT. And what a lot of companies were struggling with was they were seeing that IT and the business were trying to do this as any other kind of IT project where it was like, okay, you give us the business requirements docs, and we’ll do a PRD, and then we’ll go and build it and try and iterate. And it just doesn’t work very well.

Because what you really need as you’re going and building these agents is you need somebody who is deeply involved and has deep knowledge of what is the business process? What is the workflow? What are the problems you’re actually trying to solve? And so I think it’s really important, number one, to get that executive buy- in. And then number two, to have a team that works very iteratively together on this overall. The old model doesn’t really work. Does that answer your question, Mike? Good to see you too, by the way.

Yeah, thank you very much.

Absolutely. Thank you.

Bye. Thanks, Scott.