GPT-6 Astra Sales Agent OS is WILD!

Julian Goldie (0:00)

Today, I'm going to show you a powerful Sales Agent OS. This is a sales agent operating system I basically built to number one, track the calls from my team. Number two, to easily coach my team up and automatically analyze with AI what we should improve. And also, it uses our sales standard operating procedure system to analyze how our reps are performing versus how they're actually, you know, how they should be performing based on our processes. So this is really powerful. Number one, for analyzing the calls. Number two, for seeing what's been done recently. Number three, for actually automatically rating our teams. It's basically like an AI-powered sales coach that helps analyze our team's performance and also give me real-time data in terms of what we can improve. And I think this is massive for any company running a sales team or if you're doing sales calls day to day, this is really powerful because basically automatically pulls in the latest data in terms of sales calls and then analyzes what went wrong, what to improve, and how we can get better. Now, this data is confidential. If I click that button, I can actually open up the data for my team, and we can refresh this report. Now, it's actually running on a schedule. So this is running every 24 hours, which is pretty useful. And then you can see, for example, here we can look at our teams, we can look at our reps, we can see how they're performing, and also the coaching opportunities. So we can see rep one, rep two, rep three, and then also over here on the course section, we can analyze our teams performance. So what it's actually doing is it's looking at what worked, what to improve the next course, the next steps, and then also analysis of, for example, the discovery process and how the team is doing everything else. Now, you might be wondering, okay, how is it pulling that data around? Let me show you an example. So basically, this is something that she built this afternoon. I went in to GPT-6 Astra, as you can see right here inside chat, GPT. I said, can you create a clean sales agent OS tab inside agent OS that pulls in the latest calls from my team and all the latest conversations my team have had on Google Drive in my team's Google workspace? And it should analyze how many calls we've had, who we've had them with, how the conversations went, and what we can improve. I also want to make sure it's updated every 24 hours and tracks each dates, so the data is stored and we can look back progressively over time. And actually attached an example from a recent call to break down the data, and included information from my team. So it pulled in the latest calls and everything else, and gave me a preview, plus created a routine, as you can see right here, to update the agent OS daily, so that we can get that information in. You could do the same, by the way, with any AI agent or harness. If you prefer Claude to do this, you could do that with Claude. If you wanted, for example, Hermes agent, Hermes agent would be great, because you could use Jupyter 6 Astra as a brain, Hermes agent as a harness, and then create this as a cron job, so a scheduled task that analyzes your AI agents.

And then you can see here that it's pulling in the latest data. Now it did have a bit of an issue connecting the data to this. So what I actually did from there, is I gave it access to the latest calls in the Google Drive folder, and the calls from my team are automatically saved, which is super useful.

Then from there, we also gave it access to the Sales SOP. And basically, I have a three stage process for building and automating anything, including this Sales Team Agent OS system. So basically what you want to do is, you want to identify what you need to build, then test and iterate it, and then track it for seven days, and just see how it performs and whether you actually use it. Because if you create an automation that you don't use, you might as well delete it, but it's worth testing. So sometimes we have like an automation in mind, and then we create it, and it's not actually useful in reality. So what I've said here is, can you add a rating to each call in terms of what was good, and what was bad, and how it can improve? So it's breaking down what we can improve in each call, the good points, the bad points, so I'm prepared, and I can check out individual calls if I'm like, you know what, that one wasn't so good. And then also you can see here, I said, when you update daily, we don't delete anything, we want to keep the history in the system. So we want to be able to see past days, past weeks, past months, see the progression of our sales team, see how they're performing. And then also I've said, for the coaching points, are you able to look at the attached SAP, so I attached my sales training SAP over here, and then use that to give feedback.