I'm Andrew, founder of SagaSolve, and this is the second go-to-market update. In the last stretch we shipped the product UX page, ran our first product interviews, and put cold email outreach into motion with Stephanie driving it. This update shows how those pieces connect: an intake form that filters lead quality, an agent that turns interview notes into Jira tickets, a coding agent that executes the changes, and a social media agent that broadcasts the whole loop.
An intake form built for lead qualification
The landing page exists for ads and cold email outreach, and it asks a short set of qualifying questions on the side: whether the prospect is paid, whether they take Bitcoin, what vertical they're in. The answers route people to our calendar link so they can book an interview.
Running the form has already taught me things I couldn't have known in advance. Signal quality on booked calls was a real problem — people would book and the audio would be garbage, or they'd show up on a phone. So the form now communicates expectations up front: you need a desktop, and there's a short human introduction video so they know who they'll be talking to. Filtering happens before the meeting, not after.
What product interviews actually reveal
Watching a prospect use the product, live, is eye-opening in a way no survey can be. You see their behavior, then you ask them why they clicked something — that why is the critical information. We collect it on the landing page and on the onboarding flow.
One pattern came up again and again: most people are on narrow laptop screens, and the product wasn't designed for them. That's a frustrating experience, and it means real UI adjustments are needed. You only learn that by watching, which is why I'd call these interviews a must-do, not a nice-to-have.
From interview notes to Jira with an agent
There's an honest challenge here: agents are early in the price cycle and nobody actually knows what an agent is yet. There's a cognitive leap I'm asking people to make, so part of these videos is simply showing what an agent can do.
The demo: an agent hooked up to my calendar, Google Docs, Slack and Jira. It identifies the product interviews, pulls the documents, reads them, summarizes them and sends the summary into engineering for the human review team. Once reviewed, it delegates to another agent — a chat underneath the first one — that creates Jira issues for each finding in the product review. The coding agents that pick these tickets up can even reference the interview recordings for context.
A coding agent executing the feedback
One piece of interview feedback came up multiple times: people want the prices first. Time to make a decision on that. So I took it into a coding agent running Claude Code with Haiku — though honestly, Haiku is far behind; GLM Flash performs better — and it pulled up the Jira, understood the layout change, and moved the savings and pricing section up the landing page. I refreshed, and there it was, prices at the top.
That's the full loop: product review produces information, information goes to the team, Jira tickets get created, and a code agent executes the trivial ones. A human makes the decisions at each hand-off.
The social media agent closes the loop
Go-to-market also means communicating what agents make possible, because that's the headwind I'm fighting. Our custom social media platform is itself an agentic product: it records, outputs horizontal and vertical formats, and cuts chapter-based shorts from markers I set. It can generate speaking notes for another pass, upload to YouTube, automatically create the blog post, and broadcast to social media through Buffer.
Then analytics feed back in, so the agents change how they write based on how posts perform. That's the whole building-in-public loop.
The whole cycle in 20 minutes, with humans at the center
Interview an ideal customer profile, filter them, get feedback to the team, into Jira, tracked, changed, and documented on social media — and the video you're watching gets edited and animated by AI on the way out. I've done this whole cycle in about 20 minutes.
But notice how much human work remains: I run the interviews, I decide which Jiras get made, I review the information, I choose what to execute. And I am not an AI avatar. Putting AI in between human relationships would be our downfall — we absolutely cannot do it. The agents operate across our tools: Google Calendar, Docs, Jira, the coding agent, the social media agent, Slack. The humans stay in the loop, and the loop moves fast.
Transcript
Hey, I'm andrew, founder of saga solve. And this is go to market in the AI era.
All right, so I'm going to showcase progress, uh, of where we've, uh, made. Uh, so we published. Let me zoom in here. Uh, we published the product UX page and we've done product interviews today. So I've updated. Uh, Stephanie is doing cold outreach on email. On email. So she, uh, passes people to the, uh, intake, uh, as well. And we're updating the process of updating our ads. So, uh, we go through here. Uh, and this is better lead qualification for us. Um, and this is the, uh, page. Uh, there's like some buttons here. Ah. And then on the side, uh, just a very, uh, you know, ask some questions. We, uh, want to know if they're paid, if they're taking Bitcoin. And so I know vertical folks. Um, let's see if we can actually, uh, squeeze this in for. So vertical folks can see everything. Um, all right, that's good. So, uh, we go through this, um, and I'm learning some more things on this thing, uh, on this form, because we're going to be filtering a lot of people. Um, you know, I actually need to filter signal quality, um, because people book these and their signals garbage and I cannot hear them. It's. It's actually a huge problem. So, uh, what I've learned here, that's something I have to filter after I learned that I have to communicate. Hey, you actually need a desktop. Um, you know, that's just a video to say, hello, you know, I'm a human. This is who you're going to be talking to. Um, and then it basically sends an email. So we ask some questions, it goes through. And so this is kind of. We want to filter people and get the people we're looking for in these meetings. Uh, So this is a landing page that we built for ads and cold email outreach. And of course, it goes out to our, uh, email. Uh, that's how we, uh. Or our calendar link. Sorry, uh, they book a link on our calendar.
So during the product interviews, it was, it was pretty cool. It's pretty eye opening. Um, you know, because you get some information back, you see people's behavior, really, it's just you watch them and then you can ask them why they click something. This is some very important information. So we're collecting it on our landing, on our onboarding and how they interact with our product. And so this I'd have to recommend is some critical information that you absolutely have to do. Um, because what's really interesting is, you know, people are working on laptops, a lot of them, and we're seeing this a lot and it's like, oh, wait a minute, I have to actually make some UI adjustments because more people are on these screens and there it's kind of a frustrating experience if the product isn't really designed, uh, well, to, to fit in kind of like a very narrow laptop screen.
So a challenge we have is uh, an agent is pretty early in the prez cycle. So the technology hasn't hit turning point or uh, synergy and nobody actually knows what an agent is. This is actually quite fascinating. So there is quite the cognitive leap that I'm asking people to make to see the capability. So part of these videos is actually showing what an agent can do. So uh, what, what uh, here, here's an agent that's hooked up to my uh, calendar. I'm going to uh. Let's zoom in a little bit here. We can uh, cut uh, that out so folks can read. Um, so what it's doing is I'm having an agent hooked up to my calendar, my Google Docs and my Slack. And it can also Jira. So what it's doing is it is identifying the product interviews and then it is getting the uh, Google uh, documents and then it is reading them and then it is summarizing them and it is sending into uh, engineering. Uh, so you can see uh, you know, what we're building here. We take this lead. Uh, you know I went through this interview process, uh, have that documented and now you can see how fast the agent is actually at, at operating. Um, this section right here. So we get the information, we process it. Um, this is the human review team and then it goes out to the uh, Slack engineering and then we're going to make jurors for it and I'll show you that step next. So really this is an agent that's working across uh, four different um, uh, applications. Mhm. It. And uh, so I reviewed these and it's actually this is, this is really good actually. So what's really cool is the agent knows um, to uh, delegate these to the, the um, Jira, uh, manager and I'll say yeah, um, And so now I'm passing this information to another agent that is going to be responsible for making Jira's. And so uh, I'm just going to move over here for the. Right here you can see a chat underneath. This chat has been started. We've had an agent delegate to another agent. And so what it's going through is it is creating issues for each one of these things in our uh, product, uh, product review. And what's really cool is actually we have recordings. So the coding agent, uh, when these, when the coding agents pick these up and we'll show that in another video that they're going to be able to actually grab the video of and find out where this stuff is for reference. Right. So where we're talking about. That may not be necessary. I hope it isn't. But you can see here. That all these jurors are being made for tracking. So they can go out to engineers, and then we actually have a process where the coding agents will actually pick these up. So, actually, for an example, I think I could show Warp as another one.
Uh, so right here, uh, I'm in a coding agent, I'm just running haiku. Uh, this is Claude, uh, code. Uh, so I'm just going to actually uh, clear this. I'm in the landing project. I'm going to check that. Uh, okay, good. Uh, so what I can do here is I'm going to just show you example of you know, some of these bugs that we find in product interviews. Uh, we find in. Uh, we can quickly execute on them. So uh, here is uh, the landing page and what we have here is we uh, if we scroll at the bottom, uh, what I've heard is people want the prices first. That's the information that they're looking for. I've heard that multiple times. It's time to make a decision on that. Uh, so that was just a quick view of the landing page. Now we made those jiras. So now this is a human making a decision in the workflow. And what you can see here is. Haiku is a bit dumb. Wait, you gotta be kidding me. Haiku is actually really far behind. They have not updated it. Um, like uh, GLM Flash is like the. Okay, cool. So what's really cool is it put um, it pulled up uh, the Jira. Um, and so you can see on something that's as trivial as this is like, oh yeah, we got to move that, move that thing around. The agent's going to be good enough to actually do that of actually go to the landing page and adjust um, the layout here. All right, so there you go, it's moving up. Uh, you know this is a very easy thing to make here is you know, it moved savings and pricing. So when I go to my page now, oh I think, oh yeah, there it is. So it moved the agent did the thing, it moved the prices up. That's what is important to people. And so we've, we've made an adjustment here. Uh, so that is kind uh, of working with agents so you can see the entire flow. How this uh, change in from product review, uh, getting information, communicating it out to the team, creating uh, jira's and then executing in a code agent. Now also in part of this go, go to market strategy is the social media of actually communicating uh, what is possible with agents because that's the headwinds. Um, I'm actually fighting.
So I'm going to toggle show app. You saw the, the toggle of the screen here? Um, uh that is actually the, the region control of our custom social media uh platform. So I'm just going to turn uh off hide regions and I'm going to drag this down here so you can see it so. Oh nope, I didn't try it fully. Uh, so you can't see all of it because it's a little bit on screen, uh off a screen. But actually let's show those regions, make sure it can see itself. Uh yeah, I think, yeah I think this is good. We're gonna pull this in for the uh, for the vertical uh guys. So I'm gonna go hide uh regions and then what's really cool is I can you know because this is a custom app I can do stuff like um, remove the app from the screen recording which is really cool. So uh, what this has done is I'm recording this and you know this is actually an agentic product as well. Uh it is outputting horizontal and vertical formats and it's also chapter based. So I set a marker of what I want to chap, uh cut and it cuts horizontal, vertical and shorts. And if I want to make speaking notes to do another pass I can generate speaking notes. I can then send it up to uh YouTube. I can then send it. It automatically creates a blog. It goes out and it makes a blog and then it updates into all the social media and it broadcasts on social media uh there as well. And so that uh, communicates via buffer. Um and then we can also get analytics to actually change how our agents write uh, on when we get feedback, when the post has been uh, the post has been up. So you can see that is the entire like almost the entire flow of uh, kind of the building uh public uh loop.
Uh, so that's the whole flow from um, interviewing, um, somebody, a, uh, professional that I'd be selling to my ideal customer, uh, profile, making sure it's filtered, getting, uh, that feedback out to the team, getting it into Jira, getting it tracked, then making changes, uh, and then documenting it on social media. And um, you know, when this is done, I've done this in 20 minutes. Right, right. So, you know, you can really get a sign of. And as soon as I'm done, this can. This is going up to social media. This is going to be edited, animated, uh, uh, by AI as well, and all in 20 minutes. So this is actually what this technology does. And you can notice that there's still a fair amount of human work. Right. I'm interviewing, I'm making decisions on what Jira's goes, I'm reviewing information, I'm communicating with the team, I'm making decisions of what yours to execute. Um, and most importantly, um, this is not, I am m not an AI avatar. That I think is it. When we start putting AI in between the human relationships, I think that is, that will be our downfall. We absolutely cannot do it. We have a number of risks coming, um, that I'll talk about in AI literacy. But this is kind of, um, a showcase of what this technology can do. We work through Google Calendar docs, Jira coding agent, social media agent, uh, Slack. Right. We're, we're operating in all our tools so quickly now. Uh, so I hope this gives you guys, uh, a good intro of what's possible. Uh, and this is the end of the go to market update. Uh, two. Thanks so much guys. Bye bye.
Frequently Asked Questions
It identifies the product interviews, pulls the Google Docs, reads and summarizes them, sends the summaries into Slack engineering, then delegates to another agent that creates Jira issues for each finding.
Around 20 minutes from interviewing an ideal customer profile through to the changes being made and documented on social media, with the video edited and animated by AI.
No. Andrew emphasizes he is not an AI avatar and that putting AI between human relationships would be a downfall. Humans interview, review, and make the decisions on what gets executed.
The biggest early signal was that prospects want prices first, a request heard multiple times and acted on by moving the pricing section up the landing page.