Russell Witham [00:00:00]:
You need the structuring of the information. You need that top down approach. You need a graph of information that's backing it to really power and create an agent that can actually do the work. Do this core work of sales engineering.
Narrator [00:00:14]:
You're listening to The Unexpected Lever, your partner in growing revenue by doing what you already do best, combining your technical skills with your strategic insights. This episode was taken from a LinkedIn live series about sales engineering with our CEO Matt Darrow. We hope you enjoy.
Matt Darrow [00:00:33]:
Thanks for joining us today. I'm Matt Darrow, Co-founder and CEO of Vivun. I started Vivun after a career running global sales engineering teams at private and public companies. I'm here today with Vivun Product Leader Russell Witham, who now you know a little bit more about as well. He's also an expert in building decision making systems with AI. He lives and breathes data ML, AI, especially making the tech into fun and actionable products. So today we're going to talk about the future of AI in sales engineering, specifically what it will take for AI to truly move the needle for SES and the sales teams that they support. If you are in sales or presales and you're looking to keep up with AI's increasing impact on your role, you are in the right place right now.
Matt Darrow [00:01:21]:
All right, so in case you missed, last time I spoke with Vivun's Chief Data Scientist Joe Miller about Genai's impact on sales engineering. And as we wrap that discussion, our audience wanted to know. All right, given all of this, what tools should I be using as an SE so I don't get left behind? Well, Mark Benioff gave us a big hit last week at Dreamforce, which is now branded agent force. And in case you missed it, the most important SaaS CEO at the most influential SaaS company is reiterating that Agentic AI is fundamentally changing the way that we buy and sell, starting right now. Now at Agentforce, Salesforce told us how virtual and human team members would begin working together in brand new ways. And they focused a lot on the bookends of the customer journey. SDR agents help finding prospects and bringing them to your door and support agents helping them resolve their trouble tickets.
Matt Darrow [00:02:17]:
But business and more importantly revenue generation actually happens between the bookends, between when somebody is interested and has shown up on your doorstep, but has not yet taken the step to becoming a customer or expanding further. This is the realm of technical sales, where AE's and SES work together to drive revenue. This is the work to be done that wasn't covered at Agentforce and it all has to do with the work that sales engineers focus on every single day. So Russell, you've been working with SE leaders at some of the world's best and biggest tech companies for years now. At Vivun you've seen a lot and I could actually say you have a perspective broader than probably nearly everyone. How are companies trying to use AI for sales engineering work right now?
Russell Witham [00:03:07]:
Well, there's a sea change as we know. And you mentioned agent force just a second ago. Matt, I. And comparing ourselves to where we are today from a few years ago. Now, large language models, that's a commonplace term. When you say LM, people know what that is and that's reflected in what we see sales teams doing. And they're starting to experiment with the public foundation models, our quads, our OpenAIs. They're actually taking a look at those applications and trying to figure out what could this mean for my organization.
Russell Witham [00:03:36]:
And then likewise there's a bunch of tools that are starting to emerge that are around the periphery of what means to do sales engineering work. Right? Answering rfps product question and answer. Right, these, these things that are part of the sales engineering role. But the big challenge, right, Matt, is that's not actually the SE role, right. It's not just answering a product question here and there. It's not just filling out an RFP, right, that's just a portion, but it's not the core strategic work. And right, it's totally missing the essence of what it means to be an SE, really doing the impactful work that's going to drive tons and tons of value and allow organizations to actually facilitate. Here's the software I need to go and buy.
Matt Darrow [00:04:19]:
Yeah, well, it's nice too that even after the last like six months, LLMs and experimentation with models is becoming more commonplace. People are okay with some of the sort of security settings that are out there now and be able to access this stuff. And you're right. Like as a prior SE leader and se myself in a past life, I was always frustrated how reductive it seemed that my role was just to tell you how an integration works. In your opinion, what is the essence of the SE role?
Russell Witham [00:04:46]:
Well, beyond those edges we were just mentioning a second ago, it's really about crafting a solution, coming and understanding the goals, the problems that an organization is encountering, and how we can uniquely deliver back something that can actually resolve those problems. That's such a unique understanding of the customer's real needs, all the inputs that they're coming to that conversation with. Right. Every requirement that they have, there's all these details that go into it that's so complex and so highly technical. Right. That's why we've typically used this word engineering about. It's really crafting a solution to a problem. And that expertise is hard to develop.
Russell Witham [00:05:27]:
We know organizations invest a ton in how they ramp and train their team members to be able to have that expertise. And that compounds across years. But that's what makes it valuable. Right. It's this thing that's hard, it's complex, and it's been difficult how they can do this more efficiently, how they can help their team members be able to really do this at the scale that's now required. We have more need and demand than at tons of points in the past, and that requires this role so crucially right now.
Matt Darrow [00:05:56]:
Yeah, well, it's funny that solution is actually in the name. And when starting Vivun, we were kind of humming and hawing around. Should we talk about pre sales? Should we talk about sales? Engineering solutions, consulting. That's a common term. It's the most amazing profession that has like 100 different titles that we call each other. But I think you nailed it there, Russell. And that was from my own experience, too, as an SE leader, is that solutioning is the job and is the most important characteristic of the role. And when I think about what is a solution, it's that amazing intersection of customer goals and problems intersected with your product's unique capabilities, differentiation and proof point.
Matt Darrow [00:06:37]:
While you and I can sort of sit here and pontificate about our experience and our conversations, you actually have some data on this, too. We're lucky enough to run one of the largest biannual studies of pre sales benchmarking. And we had over 500 SE leaders, actually, in our last August breakdown, and it actually broke down. Where do SES actually spend time doing work? So could you share a little bit more about that, Russell? Because after all, the whole promise of AI is to do the work that we do as humans. So where are they focused?
Russell Witham [00:07:09]:
Yeah, and Matt, you referenced that amazing dataset that we're fortunate to be able to work on with our great customers here at Vivun. And that represents tens and tens of millions of different actions that team members have taken through time. And so when we start to break that down and dig through all that fun data, what do we find in there? And we mentioned things like RFP answering earlier. That's what some solutions are capable of doing today. But that's a really small subset of the overall sales engineering work. It's about 8% or so of the time. When we look at that in aggregate across lots of different organizations, that begs the question about where are sales engineers spending their time? Where does it go? And we see that's really predominantly in two major buckets. Bucket one, what we've been talking about of the role of solutioning.
Russell Witham [00:07:55]:
So, right, digging in on those customer problems, crafting back the solution that's actually going to solve those. And then, yes, compiling that solution to here's a compelling narrative that I have. Here's a presentation, actual demo, I want to deliver against that. And those two core aspects of crafting the argument, building up that narrative and then delivering it expertly, that's where the real heart of the work is. And that's where this unique skillset has been so valuable within the technical selling process.
Matt Darrow [00:08:26]:
Yeah, that's awesome when you can bring some of the data focus behind that, too. Russell and just drawing back to the beginning of this conversation, you've run into people experimenting with foundational models. You mentioned Claude and chat GPT. Can you just ask these foundational models to build you a demo to build a customer solution? Is there merit to that approach?
Russell Witham [00:08:49]:
Well, you can. You can ask them and they'll give you back an answer. These models, they'll give you back just about anything you ask for. But is that answer actually any good? And that's sort of where the problem is. Like, we've seen this amazing transformation, and the piece of innovation with these models is phenomenal, and that's going to continue to be the case. But there's two core problems that the models have, and these will continue to be true even as the models get significantly better. One, they don't have a top down understanding of what technical selling actually is. So we talk about this intricacy, all of the things that you actually do from sales engineering.
Russell Witham [00:09:24]:
What makes that valuable? How do you think about that process? And that's not really a characteristic of these models. They don't have reasoning in that sense of. And so you're going to get back an answer based on just the pre trained bulk of information, right? The statistical modeling. You'll come back with some idea of a solution, parts of that will even sound reasonable. But if you then tried to play that back to an actual buyer on a real enterprise deal, it's not going to give you, it's not going to give you the result you're looking for. And so you have to go deeper than that this nature of the top down knowledge, that's one of the, that's one of the key problems. And then the other problem is the actual information that we're relying on. We have a core belief.
Russell Witham [00:10:03]:
One of the things you touched on in your last session, Matt, with our Chief Data Scientist Joe Miller, around the importance of knowledge graphs, and that's really key to our philosophy. Even when you start to think about the quality of the outputs, they are improving from the models. We could talk about OpenAI's recent release of their zero one line of models and the chain of thought reasoning that's happening with those. But even those are not going to give you what you want because you need the structured, detailed network information that you're really compounding in a very unique way. And that's what we're doing at Vivun. We have to structure that information so it's actually usable and consumable as part of solutions. As we mentioned, we're going to continue to see these foundation models get better and there's lots of solutions that are going to come out into the market that are going to be really enabling. Right.
Russell Witham [00:10:51]:
You mentioned agent force. We're going to consistently see this across the market as there's going to be general purpose solutions that get brought out for teams and that's going to become more commonplace. We have seen releases just in the last month like Anthropoc's enterprise offering, where you can have your own enterprise version of Claude. Likewise, OpenAI has similar effort not too long ago, but that's, as we mentioned, that's not really going to allow you to do the work because it's only going to assist you on a portion of that work. If you want to improve an email, sure. Great. You can plug that in, you're going to get an answer back. But as we just mentioned, that's not the core of the work, it's the solutioning is the delivery of that solution.
Russell Witham [00:11:33]:
That's the core of the work. It's something that we need really detailed modeling and structuring of that data to be able to do. As we talking about this, I think about our customers and this topic and how does that gel together, right? How do we go from the importance of this data and actually delivering value back, seeing that actually manifest in the market?
Matt Darrow [00:11:53]:
I think a lot of agreement on the importance of the sales engineering work being sort of centered around the solution, because that drives everything. It just drives the executive summary, it drives the discovery questions that you ask, it drives the demonstration that you do for the prospect or the customer. And I think it's almost funny, given how transformative AI has become, that people have a little bit of a general malaise and are feeling a little underwhelmed. And what they're running into is actually what you're describing, Russell, is that they're these great big, powerful foundational models that are sort of just generically average about everything without the ability to actually reason and accomplish work that we would do as a human. A lot of approaches that I run into are, hey, let's put a bunch of documentation in a vector database, put chat bar on top of it, and you get just a slightly less annoying version of clippy. And youre just recalling information you already know. And at best youre recalling it at like 70% rate because its hallucinating a little bit because its just trying to find the next best work that is totally different than knowledge representation and what youre describing. And that was one of the things that I was excited about when I saw agent force last week, was the big push from Salesforce to say agents are the next wave humans and digital team members working together.
Matt Darrow [00:13:12]:
They were very much focused on those bookends where sales engineering work is focused on the meat of revenue generation. And it is only going to be possible if you can have some true knowledge representation that will let you do the work and the work that ought to be done when it should. When I talk to SE leaders about this too, I normally see them actually gravitate to four really potentially quite substantial benefits if you can actually do this sales engineering work with AI. And the first is, might be a little surprising, but it's actually all about a independence. There's probably a lot of SE leaders and sales leaders out there if you're watching this or watching this on a replay later, that you have like volume bases of your market, commercial SMB, mid market, lots of deal, high volume, high transaction. And you know, what kills a deal or stalls it or pushes it two to three weeks is if the AE can't actually drive it forward with the velocity that they need. So being able to take the work that SES can do and give it to the hands of the account executives becomes really, really powerful for those market segments. The other thing is also enablement.
Matt Darrow [00:14:20]:
This one actually surprised me. I wasn't expecting this to come up in conversation around where AI could play a really big role for sales engineers. But I was talking to folks that have pretty technical products. They were like, that takes like nine months for an SE to become effective. And if I can get them out in the field working in one to two months, and I could take the quality of everybody and shift it right. That's just a massive benefit for the whole team in general. Had the fortune of speaking to folks that TAM is so large and growing so fast that they actually can't keep up with chasing it because it's almost like the expanding universe. And they're saying, well, if I could take 30% to 40% off my team's plate, it's not that I'm going to take my team and shrink it.
Matt Darrow [00:15:02]:
What's going to happen is I'm going to go chase that Tam that I wasn't able to run at fast enough. And that actually is a contrast to folks that maybe be part of PE backed companies or folks that were acquired who actually do want the better ratios. How do I go from a two to one to a four to one? And these are all things that happen when. To your point, Russell? When you're not focusing on the 8% of the work that SES do, but you're focusing on the 60% plus of the work, which is how do I craft a solution for the customer and deliver that in a compelling narrative with the right demonstration visuals along the way?
Russell Witham [00:15:34]:
All right.
Matt Darrow [00:15:36]:
Wow. I just went for it there, but it felt there's a lot to say. So, Russell, we've been talking for like 20 minutes here, so if people remember only one thing from the conversation today, what do you want it to be.
Russell Witham [00:15:46]:
That we're in this transformational moment? That's incredibly exciting, Matt, but the current approaches alone are not going to get you to the place where you can create that leverage on those use cases you mentioned. The current approach is also they have a reductive understanding of what it means to be a sales engineer. It's, oh, that's the folks who are going to go and answer my product questions. That's the folks who are going to go fill out my rfps. And as we just discussed, that's not the real work that a sales engineer is doing. That's not the value they're bringing to the table. As you're talking about the value of scaling that up across my tam, it's not, I need to scale up answering more rfps. If that were the case, fantastic.
Russell Witham [00:16:25]:
But it's, I need to scale up more people who can efficiently understand the problems of my customers, be able to deliver back solutions on that. And you're just not going to get that out of a generic LLM, as powerful as they are, as fast as they're going to continue to evolve. You need the structuring of the information. You need that top down approach. You need a graph of information that's backing it to really power and create an agent that can actually do the work, do this core work of sales engineering.
Matt Darrow [00:16:56]:
If you're curious about what does it mean to do the actual real heavy lifting work of a sales engineer? And if you just like this discussion around how AI is transforming sales and pre sales along the way, join us at unexpected. The theme this year is the rise of the aise. We'll be airing it virtual on October 2 and will also be bringing unexpected to a city near you. Five cities, starting with San Francisco, will also be in Seattle, Austin, Boston, New York. Register with the URL you see on the screen. We'd love to see you live. We'd love to see you virtually. We're going to continue this conversation.
Matt Darrow [00:17:32]:
Mark Benioff and the Dreamforce squad that's now agent force did an awesome job of basically kicking the whole market into gear that agents are the future. And while they're focused on the bookends, let's talk about the meat of revenue generation. Along the way, our whole focus is how can AI change B2B selling for good. That's why Russell and the rest of our team are working to ensure that the future of AI works to drive the pre sales profession forward. Thanks for joining us, and we'll be able to see you guys next time.
Jarod Greene [00:17:59]:
For additional resources, check out vivun.com. and be sure to check out V5, our five-minute soapbox series on YouTube. If there's a V5 you'd like us to talk about longer, let us know by messaging me, Jarod Greene on LinkedIn.