Meet Ava. The World’s First AI Sales Engineer

Julia Nimchinski [ 03:00:52 ] We are transitioning to an actionable part of our summit, AI tech demos, and we’re going to see Vivun in action. And. And. And the world’s first AI sales engineer. Welcome, Matt Darrow, again. How are you doing? Thanks for having me back from this morning. I was able to get some water, get some tea. I’ve got a really easy goal, which is to make this 15 minutes the most memorable 15 minutes of the entire event. It’s a high bar, but I think we can do it. Let’s do it. Okay, let me first make sure the screen share is coming through correctly. Looking good? All right. So, hi, everybody out there. I’m Matt Darrow. I’m co-founder and CEO of Vivun. We’ve raised over $130 million to build critical tools for technical sales teams.

Matthew Darrow [ 03:01:48 ] And we’ve created the industry’s first AI Sales Engineer. Why we do that? Well, this is our new reality. In the next year, AI is going to do over half of the work that SEs’ team is going to do. And that has massive ramifications for your sales team. The first is just way more capacity to go and chase after your target market or dramatically reorg your teams. The second is full coverage. Every seller, every account executive getting on-demand access, more power and independence. So instead of running 4 to 1, 5 to 1, I’ve heard of 40 to 1 AE to SE ratios. That’s a thing of the past. And 70% of the sales cycle is spent on technical validation. That point between I’m interested and I’ve signed a contract.

Matthew Darrow [ 03:02:38 ] So if you’re going to do major damage to the velocity and your experience that your buyer has, this is where you want to focus. For us, this is the work. These are the 12 things that the world’s best sales engineers do. And last month at our industry event, we actually showcased AI development. And we’re doing all of these. And honestly, it freaked a lot of people out because they thought, hey, this might happen five years from now. And we showed them that it happened yesterday. And this is where now culture and technology are on this collision course to catch up. And adopters are really going to spring forward. So the rest of my time, another 10 minutes or so, I actually want to give you the broad basis of what you’re going to see.

Matthew Darrow [ 03:03:24 ] And if you’re interested, you want to talk to me, you can follow me on LinkedIn. I’m Matt Darrow. You can visit us at Vivint. com/ slash A-I-S-E. Get in touch with us. I’d love to introduce you, though, to Ava, the world’s first AI sales engineer. And working with Ava is unlike anything that you’ve ever experienced before because it’s like working with your best teammate, except that best teammate is always available, never on vacation, and knows a lot about everything that they need to do in their job. And that’s the fundamental promise and premise of GenAI. It’s doing the work that we do as humans, not necessarily how we’ve worked with SaaS in the past. And that’s why it’s so different. Now, you train Ava and you onboard Ava just like you would a new hire.

Matthew Darrow [ 03:04:12 ] So how would we do it as a human? And because she has a point of view of the work that she’s hired to do, in this case being the world’s best technical seller, she also knows what we haven’t yet covered. In this case, discovery. Super important part of any sales process is to really uncover customer needs, and we’re going to help Ava get up to speed just like we would a new hire human direct report. I’m going to use our 10-page Word doc that we have here at Vivint that describes how we do discovery, and I’m going to have her review it. That way I can course correct, I can validate, I can make sure that the understanding is there, and she’s going to give me the TLDR on what is discovery all about at Vivint.

Matthew Darrow [ 03:04:52 ] And again, she’s going to net down what are the pains that we’re searching for, the questions that we like to ask, why is it relevant for different parts of our product, and the most important thing that she is going to do is append this new knowledge to her full fundamental graph of understanding the work that ought to be done. I talked about this a little bit early on the HSC Summit, but this is where agentic experiences come from. It is not from the LLM. Yes, Vivint uses LLMs as part of our stack. We rely on OpenAI. We use Anthropic for different use cases. But the way that you build an agent is you need to impart knowledge, memory, the work and understanding of what ought to be done in different scenarios, and that’s how you can actually go the full mile and do the work that we do as knowledge workers.

Matthew Darrow [ 03:05:42 ] So let’s get to work. We could spend hours more talking about that approach, how it prevents hallucinations, how it allows you to build agents, but let’s have Kevin, our account executive, actually pair up and get that 100% coverage. And that is how we can really leverage and completely change how they are able to attack their sales process. Now, the second thing that your best team member does, instead of being always available and extremely knowledgeable, they are normally super proactive. They want to come to you with how they can help and how they are going to move things forward. And that is what we have done as well. So AVA integrates with your core traditional systems around conversational intelligence from providers like Gong and Chorus to CRM systems like Salesforce to productivity suites like Google and Microsoft Outlook.

Matthew Darrow [ 03:06:26 ] And we are pulling all this information in to make sure we know what are we going to do next. And in this case, in the next three days, we have got this upcoming meeting with Hooli. So great. AVA is going to get me ready to go. Who is on the call? Who are we meeting with? What is important to them? What is latest in the news about this company? This is all basic stuff. But the thing that is the most important because of the knowledge of the work that ought to be done is this deep understanding of what should we be doing in this scenario. Because of the meeting, the context, the people that are on it, AVA knows what is the work we should be doing is going to help me out with that.

Matthew Darrow [ 03:07:03 ] In this case, because it is a deep-dive demo with pretty high-level executive stakeholders, we want to first start with an outline so we get our message across. And because AVA has already been on every prior conversation from the very first call that we ever had with the customer, we are already starting to identify really important new critical pieces of information. Like things that our product doesn’t do that we want to potentially proactively address. In this case, this first gap around use cases becomes really important. So she is going to take all this situational awareness, take my guidance as somebody that is helping sort of approve the work that we are doing, and then actually do the work.

Matthew Darrow [ 03:07:43 ] And this is what’s so important in the premise and the promise of Gen AI is doing the work that we do as highly paid knowledge workers to a fidelity and an accuracy that you never expect. This is something that we have experienced before. Now, you could go to Claude or OpenAI and say, ‘Yeah, I’m an SE that works at Vivint. Build a demo script for me.’ And you will get something. And it will be generically terrible. Because that is just how the LLMs work. All it is doing is co-occurrence on language with no really understanding of the knowledge of the work that needs to be done. But not in this case. Because we have this great top-down understanding of the work to be done, powered by the only information about your company.

Matthew Darrow [ 03:08:24 ] Now Ava has a point on how we are going to do this demo. These are the key points that we want to hit. And here is how we are going to get them across. What we are going to show, say, and do. The objections that we can expect. The traps that we want to lay. And we didn’t want to stop here. We said, well, great, this is good internally that we could riff on back and forth. But why don’t you just take it the full mile and then build the assets that I’m going to then go and deliver to the customer. And this is, again, earlier in the year when some of the foundational models got really phenomenal at sort of vision recognition. You can compile all this.

Matthew Darrow [ 03:08:58 ] The talking points are all completely AI-generated. But all of the content, all of the material, how are we going to hit the main points with pretty boring, isoteric concepts around solution architecture but more important concepts around our product, this is what we want to focus on. Now, I don’t have time for it today. But one of the things that we did on our Vivint Roadshow when we really showed people how fast this is all moving is Ava didn’t stop here. She actually joined the Zoom call with the customer, built a demo for this outline that never existed before using our real application, provided a talk track on top of that, and had a fully interactive experience.

Matthew Darrow [ 03:09:39 ] And I mention that because one of the things that I talked about earlier today on the HSC Summit, the number one thing that I think a lot of GTM folks are getting wrong is how quickly this is all moving and how quickly they can start to adapt. This isn’t five years from now. It is not science fiction. These are things that are happening right now. To bring this home, I want to shift gears a little bit and sort of jump three days into the future. So now imagine I’m that rep. I was able to do all that autonomously. I have the best assets at my fingertips. Now I have that meeting with the customer. And all that information now comes into Ava.

Matthew Darrow [ 03:10:14 ] And what she’s doing behind the scenes, is doing the long work without anybody telling her to do it. And it’s the work that is the most important thing that an SE does, which is generate the customer solution. That is the intersection of the client’s goals and challenges and problems with what the heck your product does, how it’s going to meet those needs, and how it’s differentiated. This is why companies pay SEs 200,000 plus north a year because this domain number is the most important thing in the world. We have to establish an expertise of synthesizing all of these stakeholder requests, understanding what we have to offer and why it’s relevant. What are the proof points that actually showcase we can do it better than any alternative or competition?

Matthew Darrow [ 03:10:59 ] And by the way, if there’s any gaps, how do we position and sell around it? This is the long work that is super critical. And without this, you can’t get anything else. You don’t know how to drive discovery. You don’t know how to build a deck. You can’t put together a demo because this all flows from this deep understanding of what the customer needs. And the brilliant thing about AI is how interactive this all is. This is all great for a long form, but if I just needed specific sets of information around, you know, how does our JIRA integration work again? I’m just going to go and ask Ava those questions. She’s going to tell me.

Matthew Darrow [ 03:11:37 ] So all of this knowledge and repository that she’s continuously learning, not just from when we onboard, but every interaction and every deal, it’s something that a human being just truthfully does not have the capacity to do. Not only recall and remember and know how to do the work, but then be in every situation simultaneously to learn from all those interactions to drive innovation. To drive everybody forward. Again, super short demonstration today. I only had 15 minutes, but what I was hoping to do was showcase to everybody here a couple of things. The first is what AI is capable of doing and what’s going to happen in the next 12 months is over half of that technical selling work is going to be able to be done by AI.

Matthew Darrow [ 03:12:22 ] And that’s going to give you new power and dependence for your sales team, and lift the quality of work of everybody else, and dramatically change how you think about your GTM team. And the second thing that I’ll leave you with is it’s not five years from now. So much of what I talked about, those 12 core jobs, the way that you work with a new team member, this isn’t something that might happen three years from now. It’s happening yesterday. And if you guys want to learn more, want to talk to me, hit me up on LinkedIn, come to Vivint. com backslash AISE, and we’d love to be able to chat more. So, Julia, I pride myself as somebody that before founding Vivint, I built and ran global SE teams, most previously at Zoraa, on their iPhone, and now I run an IPO-run.

Julia Nimchinski [ 03:13:01 ] And the SE skill set and doing demos is close to my heart. So, ending on time, really important. And I appreciate you having me for the 15-minute time slot. This is so cool, Matt. Really love the tech. People want to know if you are using it internally. Oh, of course. It is actually one of the most fun things. If you engage with Vivint, part of how we demonstrate AI is we showcase how we work alongside of it in our deal pursuits for all of our other products as well. And it’s a blast to do because people go wild for that. And another question we’ve got, how did you make your LLM so precise? Well, remember, we use; so we’re not an LLM provider, right?

Matthew Darrow [ 03:13:47 ] So that’s where all the billions of dollars are going into funding, you know, Gemini and Lama and Claude from Anthropic and OpenAI, ChatGPT. So those are just like base tools that you use. But how you provide it, how you prevent hallucinations, how you actually codify the work that needs to be done in knowledge, there’s a whole other set of tools and approaches and techniques that you need to do that ride alongside of the LLM. Think of the LLM as the mouthpiece. It’s just giving me this human interaction, but it ain’t the brain. The brain you need to build, and that’s where the IP comes from. Really cool. Thank you again, Matt. Everyone follow Matt Darrow and check out Vivint. We shared all the screenshots in Slack and we are transitioning to our next demo.