7 Ways I Use Jev + Grokbot to Run My Business

If you combine Jev with Grokbot, you've got an AI superpower. You've got AI advantages that other people just don't have right now. And what I'm going to do in this video is I'm going to show you how I'm using Grokbot and Jev and how you can also think about applying it to everything else that you're doing. All right, so first and foremost, the way you want to think about Jev is it is a classifier, right? And so what this means is that it can classify a lot of the things that you're trying to optimize for. So let's say you're trying to optimize for people that you're trying to recruit, or you're trying to optimize for a certain set of leads that you're going for. So you want to have only ideal client profile fits in your lead list. So let's say you have a 60,000 list of leads, you maybe only have 2,000 that are ICP fit, but you don't have to sift through all of them, right? From an AEO SEO standpoint, maybe you want to make sure that all the content you're pumping out is not related to content that you've already created, right? Or sometimes maybe you do, but you can add all these classifiers and Jev can classify it very quickly for pennies on the dollar. So the way this workflow works is that then you have type-safe Jev. This is your shared layer over here. And then you can basically implement, hey, is this a choice? Is it like a multiple choice, like A, B, C, D, D, F? Field close, like is it a yes or a no situation? Or it can be a confidence score type of thing. So you can have different levels of classification. You can have a lot of layers of classification. And then from there, you can decide if something is a hold, a suppress, a route, a clear, recommend, and then other agents can take it over from there. Because once you have it classified, then you can take it from there. And then you have a human in the loop here jump in and they add in their judgment layer and then this can continue to be better over time. And then you hand it back over to the bot and the bot executes. And this is just the future of work. The way it works for us is I'll just give you an example here for recruiting. We use this pretty heavily when it comes to recruiting. It's been awesome. Okay? So again, I'm mostly talking about this from a work standpoint, but for example, if we're reaching out to someone, if we look at their profile and their profile doesn't have at least 2 promotions at 2 different companies with an average tenure of 3 years, we're not going to try to shortlist those people. Okay? So we have a talent bot that manages that. And the talent bot here, Jeff can decide if it's a pass, hold, a miss on the role cards. Okay. We send them 5 questions. To kind of gauge where they are from an AI fluency standpoint, right? So we'll ask them, hey, um, you know, what percent of your work is done by AI right now, you know? And also, what are your daily AI drivers? How much of your own money are you spending on AI per month? We ask those— them those questions, right? After those questions are asked, if they respond, then it'll go ahead and evaluate them, okay? Then only then, if they pass, then it will move over to us. And then, um, you know, we have a software that we use for, for agent recruiting called JuiceBox. It'll actually manage JuiceBox as well. It actually manages LinkedIn Recruiter as well. So it can manage outbound there. Now that is at your own risk if you want to risk your own account to do that. But we also use HeyReach for outreach as well. So Jev is really strong at refusing rather than filling, right? But when it comes to recruiting, what are you doing? You're rejecting most of the time because the idea here is that you're trying to recruit the top people that are in the top 1%, probably in the top, even the top 0.1% of people that join your company because you're trying to build an organization of high agency, curious people, adaptable people. That's what you're looking for, right? So recruiting to me as the founder and the owner of the company, that is number one for me, right? I want to make sure that we have the best team. And so that's been a godsend for us. If you're enjoying this right now, you got to check out the Single Brain Gateway. This is free access that you can connect all of your tools in with. This is basically API gateway and you can have all your Connectors, all your tools, all your keys in one spot, in a secure spot, so agents can talk to it, so your team can access it as well. And there's so much more that you can do with it, and you'll find other growth workflows within the SingleBrain Gateway. So just go to singlebrain.com/gateway to learn more, and we'll see you on the other side. So the other thing that we have over here is, you know, we have a Revenue Outbound OS. Okay, so you can see here, what does this do? Okay, so it looks for, you know, warm tier 1, tier 3 leads and Basically, it looks for people that if we're reaching out to them, it's going to suppress people that we've spoken to recently. And basically, it's going to block customer outbound before we send, right? Because in the past, people would just blast whatever and it's like, hey, what if we talked to these people already? What if they're actually within our— what if they're actually about to close with us and then we're reaching out to them for net new sales? That makes us look bad. It makes us look unprofessional, right? And then by the way, like if there's news and we haven't talked to them, okay? If there's warm news in the last 30 days or so, we'll reach out to these people and say, hey, you know what? Congrats on becoming CMO. Because when you become CMO at a new company, people want to buy new stuff, right? And so, we'll look for news and then if they don't have the news, we'll still reach out to them if they're ICP fit, okay? And so, just adding that little personal touch goes a long way. Website outreach. So if someone reaches our website, they seem to be ICP fit, we'll decide, Do we admit, hold, suppress, and do we push them to our CRM? If they are ICP fit, then Jeb is going to help classify that again. And speed to lead, that's okay, cool. When someone looks like they're a fit to work with us, we will decide if do we dial them, do we hold, do we skip them? Do they look like spam? Are they already a customer? Are they already ICP fit before we dial them within the first 60 seconds? Guess what? The speed to lead bot within GrokBot actually calls them using the xAI voice and it sounds really good. And by the way, guys, like all these bots that I'm telling you about right now, you're not going to figure it out the first time. Sometimes it's days of work or weeks of work to get it right. Okay. LinkedIn intent. Okay. So basically we'll look at people on HeyReach. So HeyReach is good for LinkedIn outreach and we'll decide, are they a pass or are they a go to reach out to these people, right? Before we queue them up here. So next one that I was talking to you about is this Content AEO operating system. Okay. So you can see here, we will mine for content ideas. So it can be from content that I'm putting out on my podcast, my YouTube channel, It can be from our internal calls that we're ingesting, client objections or prospect objections, internal calls that we're having. Basically, a content spike is maybe something that's worth looking at in terms of creating content around, right? And so every single day now, this workflow has gotten way stronger where it's giving me, it was giving me 40 ideas a day, but I couldn't judge if we had produced this content already, if it looked like AI slop, If it had cited statistics, it had done all these things. I couldn't classify these things, right? And so the problem with AI before is like you're just going to publish a bunch of slop. But now that you have this, we're making sure we have statistics in there. We're making sure that we have some images in there. We're making sure that is there AI tell. We're making sure that are we backing up our claims in there as well, right? Are we making sure that we're pulling from maybe recent statistics and news as well? And then image slopgate. Okay, so if we have— we can tell through the AI tell, Classifier here if we're publishing a bunch of trash where you can't even read the text, right? We don't want that. If you're enjoying this, you got to check out Single Grain with a G, singlegrain.com. If you need help with marketing, AI-enabled marketing on the paid media side, on scaling creatives, on AEO SEO, on building cold outbound infrastructure, on having these agents live inside your Slack, building these managed marketing agents with a team, with people that actually know how this stuff works. Go to singlegrain.com and we'll see you on the other side. And also, by the way, it'll review the ideas initially, it'll run a classifier, and then once it finishes drafting, it'll run like a publish gate classifier as well to make sure that this fits our guidelines. So for this SEO AE one, I'm going to show you this real quick, right? And so you can see it working in actions. I've continued to improve this where it's like, hey, for the content that we're writing, I want you to make sure that we have an H1 for each of these, right? It basically will put like a criteria gate saying, hey, this is actually ship ready. And so anything above like a 0.35 is considered ship ready. And then these are a hold over here. So very quickly, I can trust the judgment of this. And maybe I don't have to click all these. Maybe I'll click like a few of these up here and a few of these down here just to make sure that the judgment is good. Is this already live? MOFU content. Here's the headline. Here's the idea over here. Here's examples of where you actually covered this. Is this meaningfully better than what you have out there? We have this haggle bot. Okay, what the haggle bot does is it negotiates deals for us where, hey, are we going to renew this SaaS product? Are we going to downgrade it? Are we going to cancel it? So Asana, it took a look and it's like it actually combined Jev and it's like it classified everything and said, hey, based on this classifier, it doesn't seem like many of your users are actually active. You should probably downgrade it. So I couldn't classify that before, but now it's like, okay, well, not only can I see what we're— can this haggle bot see what we're paying for? It can see who's actually active within these tools and then We can decide what to do there, right? One recommendation I have for you, one lesson that I learned from all this as well, is that you don't want to burn the expensive tokens first and then put the classifier. As often as possible, you want to avoid that. What you want to do is you want to— if you're going to evaluate a large dataset, then you want to chunk the data. Okay, so let's say you're looking at a bunch of Gong calls, like that's sales intelligence. So instead of going through all the transcripts, going through all of them, maybe you just chunk it, right? So maybe like a couple minutes of each of these, and then you run like a cheap tag on each of these. Like a numeric result, an operator pain, upsell signal, AEO spike, nothing, right? Now, what I had done was I had queried Jev inside of our Slack agent that anybody on our team can access. And I just said, hey, look at all of our gone calls and then tell me where you see opportunities for where we can kind of map out case studies, right? And then that's where it called it out. It's like, hey, you actually shouldn't be doing that. And Claude actually called it out when I was testing Opus, the new version of Opus. And it said, hey, it's actually way cheaper if you do it this way. And then gates beat softpass LLMs. Helpful model softpass, Jev refuses when something is very thin. Thin being that it's low quality, so it saves time there. And then soft does not equal ready, does not equal send, does not equal publish. 4 verbs, collapsing them is how fleets leak. You don't want these to overlap with each other. Obviously, these are all different actions over here. So anyway, I would just say that it has sped us up ultimately by having this classifier. So highly recommend that you use GrokBot, number one. Highly recommend that you use Jev, number two, but also highly recommend that you're using Jev within any of the harnesses that you're using. I do believe that this is going to become table stakes in the next, probably within the next 3 to 6 months or so as of this recording. So anyway, that's it for this video and we'll catch you in the next one. Thanks, Eric.