The Blueprint Engine Explains Itself | Aug 24
Show notes
What the episode covers
The Blueprint Engine now explains its own outline decisions and suggested segments directly in the Mato dashboard, so you can see why your episode is structured the way it is.
- Blueprint Engine: new self-explaining outlines and segment rationales for each drafted episode.
- AI production workflow: how Mato uses AI for scripts, partial voice work, and editing on Made on Mato itself.
- Live AI host interviews: what it means to have an AI host interviewing a real human guest on Mato.
- Founders on-platform: why Alexander and Edward Benz are building and shipping features using their own shows.
This week’s guest is Alexander Benz, co-founder of Mato and host of their internal product feed, who explains the one constraint that makes AI-driven interviews actually listenable: always keeping a live human guest in the loop, on-mic, and in control of the final cut.
The team also mentions the companion blog post, “The Blueprint Engine Explains Itself”, which walks through examples of the new outline explanations.
Week 35, 2026
For the full changelog and product posts, visit heymato.com.
Timeline
In this episode
7 moments worth skipping to. The timecodes match the player above.
- 0:15Introduction
- 1:28What Mato Actually Is
- 4:39How Made on Mato Works Every Week
- 8:15Inside the Blueprint: How This Episode Got Made
- 11:56What It Costs and How Long It Takes
- 14:54Early Traction and What It Means for Listeners
- 17:48Outro
Quick answers
Straight from the episode
The questions this one settles, without the listen.
- What is the Made on Mato podcast about?
- Made on Mato is a weekly show where the hosts walk through what shipped on the Mato platform, why it shipped, and what it means for people running shows on Mato.
- Who hosts the Made on Mato podcast?
- The podcast is hosted by Elena and Blake, two commentators who both review the same changelog and openly disagree when they see features or updates differently.
- How is the Made on Mato podcast produced with AI?
- The show uses AI for scripting, some of the voice work, and a chunk of the editing, and the hosts are explicit about not hiding that AI is part of their production process.
- What is Mato as a platform?
- Mato is described as an “AI podcast engine” founded in 2024 in Los Angeles by Alexander and Edward Benz. It powers shows by using AI to drive core parts of podcast creation and production.
- How does Mato use AI in podcast interviews?
- On Mato, an AI host conducts a live interview with a real human guest—someone physically present and answering questions in real time.
- Why do the hosts emphasize their disagreements on Made on Mato?
- They see value in presenting both sides of a feature or update, so when they land in different places after reading the same changelog, they keep that tension in the show instead of smoothing it over.
Transcript
The full conversation
Every word of the episode, 3,326 of them, in the order they were said.
Read the transcriptHide the transcript
Elena ReyesOkay, so three episodes into planning this, and I still don't have a good answer for, wait, are we recording?
BlakeWe're recording.
Elena ReyesCool. Okay. So here's the thing. If you're brand new, you might be sitting there going, "What even is this?"
BlakeFair question.
Elena ReyesThis is Made on Mato. Every week we walk through what shipped on the platform, why it shipped, and what it actually means for you if you're running a show on it.
BlakeThat's the whole pitch. No fluff.
Elena ReyesI'm Elena.
BlakeI'm Blake. And, um, no anchor here, no sidekick there. We're just two people who look at the same change log and sometimes disagree about it.
Elena ReyesWhich happens more than you'd think.
BlakeIt does.
Elena ReyesAnd we're not gonna smooth that over for you. If we land somewhere different, you'll hear both sides.
BlakeThat's kind of the point of having two hosts instead of one.
Elena ReyesOne more thing before we go further. This show is produced with AI. We're not hiding that.
BlakeYeah. The scripts, some of the voice work, a chunk of the editing, that's the process.
Elena ReyesWhich is honestly kind of the perfect first topic.
BlakeBecause the platform we're talking about is the one that built the show that's talking about it.
Elena ReyesRight. It's a little recursive.
BlakeA little. So before we get lost in that loop-
Elena ReyesLet's just say what Mato actually is.
BlakeYeah, let's start there.
Elena ReyesSo Mato itself, founded in 2024 right here in LA by two guys, husbands actually, Alexander and Edward Benz.
BlakeFounders behind their own product on their own show. Kind of an odd flex for episode one.
Elena ReyesMaybe. Mato's press kit calls itself the AI podcast engine.
BlakeWhich means what exactly in practice?
Elena ReyesAn AI host does a live interview with a real human guest, someone actually sitting there answering questions.
BlakeReal guest, AI interviewer. Got it. That's still a strange sentence to say out loud, isn't it? An AI conducting the interview.
Elena ReyesIt is. But it's also kind of why we're doing this show, to actually watch it happen episode by episode.
BlakeFair enough.
Elena ReyesAnd after that, Mato handles the rest, editing, transcripts, clips, cover art, and pushes the finished episode out to more than 150 platforms.
Blake150. That's Spotify, Apple, YouTube, and then 147 places nobody's heard of.
Elena ReyesPretty much, yeah.
BlakeWhy does the platform count even matter for a show this size?
Elena ReyesBecause distribution is usually the boring manual part, submitting your show everywhere one at a time. Madeo apparently automates that.
BlakeSo is the interview the actual selling point here, or is it everything that happens after?
Elena ReyesBit of both, honestly. But the after part is where the time savings really show up.
BlakeOkay, give me a number I can hold onto.
Elena ReyesThe press kit says the whole thing, from recording to a finished, published episode normally takes eight to 12 hours of human work.
BlakeA full workday, basically.
Elena ReyesMore than one if you count editing on its own. And Madeo claims it does the same thing in about 30 minutes.
Blake30 minutes.
Elena ReyesYeah.
BlakeThat's the number I want to actually poke at later because that's a big claim to just take at face value.
Elena ReyesFair.
BlakeBut before that, I want to sit on live interview for a second because that phrase is doing a lot of work.
Elena ReyesGo for it.
BlakeLive to me means the AI is actually listening to what the guest says and picking the next question off that, not reading down a fixed list.
Elena ReyesThat's the claim. Yeah.
BlakePicture an actual episode. Guest says something like, "Well, it's complicated." Does the AI actually ask why, or does it just move to the next line on the sheet?
Elena ReyesThat's exactly the kind of thing I want to watch for as we keep doing this.
BlakeBecause if it's stitched together after the fact and just labeled live, that's a different show than the one being pitched.
Elena ReyesAnd if it's actually reacting in real time, that's a pretty different product than what most people picture when they hear AI podcast.
BlakeSo which is it for us specifically for what we're actually doing right now, recording this?
Elena ReyesKnowing the founders, the pitch, the platform count, none of that tells you what actually happens week to week on this show.
BlakeFounding story. Guest promise distribution numbers. None of it is the workflow.
Elena ReyesWe'll know it when we see it, a real follow-up on tape versus something that just sounds like one after editing.
BlakeThat's the test I want to hold this show to. Honestly, not just Madeo's guests.
Elena ReyesRight. So next, let's walk through what a Madeo week actually looks like, start to finish. Okay, so we know what Madeo is now. Founders, the guest thing, the platforms, got it.
BlakeGot it in theory.
Elena ReyesSure. But what does this show actually do with all that week to week? Like, what are you tuning in for?
BlakeThree things, basically. First one's the change log. Every time Mato ships something, a new feature, a fix, whatever, we talk about it here.
Elena ReyesAnd not just here's the new button. We want the why. Who was annoyed enough to ask for it? What it used to do before. What broke when they turned it on?
BlakeBecause something always breaks a little.
Elena ReyesAlways.
BlakeGive me an example, actually. Something concrete, not hypothetical.
Elena ReyesSure. Say a feature ships that lets you swap host voices mid-series. We talk about who asked for that, what limitation it papered over, and whether it actually holds up in practice or just looks good in a demo.
BlakeSecond thing is the blog. If Mato posts something, we're not gonna read it to you.
Elena ReyesGod, no.
BlakeWe're gonna talk about the stuff that got cut. The paragraph that didn't make it, the number that got rounded off, the internal argument behind a decision.
Elena ReyesRight. The stuff a blog post can't really carry because it has to sound clean.
BlakeExactly. A post has to commit to one version of the story. We don't.
Elena ReyesAnd then the third piece, most weeks we're not doing this alone.
BlakeA customer joins, somebody actually running a show on the platform.
Elena ReyesAnd we ask them the boring questions like, how long did booking a guest actually take you? What did your team do with the transcript afterward? Which clip actually got views versus which one you thought would?
BlakeThe stuff that never makes it into a testimonial.
Elena ReyesYeah. Nobody puts the clip I hated did the numbers on a landing page.
BlakeProbably should, though.
Elena ReyesI mean, think about it from a listener's side too. If we only ever hear the wins, we don't actually learn anything about running a show.
BlakeRight. The failures are where the actual lessons live. Anyone can post a win.
Elena ReyesOkay, so changelog, blog, guest. That's the shape of the show.
BlakeThat's the shape. But I wanna flag something because it's been sitting with me.
Elena ReyesGo for it.
BlakeMato makes this show about Mato using Mato.
Elena ReyesRight, the recursion thing again.
BlakeSo how does anyone listening know this isn't just marketing wearing a headset? Like, why would we ever say something's broken if the company making the podcast also makes the product?
Elena ReyesThat's fair. I don't think we resolve that today. I think we just keep it in view.
BlakeSame bucket as the Live interview question from before.
Elena ReyesTwo threads. Is the AI actually reacting in the moment? And is this show actually honest about its own company?
BlakeNeither one goes away because we said the word transparency once.
Elena ReyesNo, it doesn't work like that.
BlakeIt's easy to say we'll be honest in an INTRO episode. Harder three months in when something ships broken and it's embarrassing to talk about.
Elena ReyesThat's exactly the test. If we ever go quiet on a bad ship, call us on it.
BlakeSo here's how I'd answer it. At least partly, the show itself is supposed to be the proof. Today's episode is an example of the thing it's describing.
Elena ReyesWait, meaning what exactly?
BlakeMeaning the process that built this episode, the planning, the structure, that's not hidden. We can open it up and show you.
Elena ReyesLike, actually show the mechanics.
BlakeYeah. And it turns out there's a fresh one to point at. Something changed in how this got produced, and it shipped this week.
Elena ReyesWait, this week? As in this episode?
BlakeThis episode.
Elena ReyesOkay, now I'm actually curious. Don't just tell me, show me.
BlakeThat's the plan. Let's open it up. So the fresh example is Alexander's post. It went up on the site this week called Introducing Blueprint.
Elena ReyesWait, our own Blueprint, like this episode's Blueprint?
BlakeThis exact one.
Elena ReyesOkay, that's a little wild.
BlakeAlexander's post lays out that every episode now starts as a typed plan before a single line of script gets written. The plan, meaning what exactly?
Elena ReyesMeaning the structure gets locked first.
BlakeMeaning the structure gets locked first. Segments, points, sources, all of it before anything sounds like a podcast.
Elena ReyesSo the plan comes first, then the words get poured into it.
BlakeRight. And the post says the arithmetic in that plan, like segment lengths, gets checked in code, not eyeballed.
Elena ReyesChecked in code. So if a beat's supposed to run three minutes and it doesn't add up-
BlakeIt gets flagged before a generation even starts.
Elena ReyesHuh.
BlakeIt also means catching a smaller mistake, like a segment that reads long but actually times out short once it's read out loud.
Elena ReyesWhich is a pretty specific kind of bug to design for.
BlakeAnd after the script exists, it gets reviewed against the sources it's citing, so the claims have to trace back to something real.
Elena ReyesThat's the part I actually care about, not the timing math, the fact-checking part.
BlakeSame. According to the post, if the script can't back a claim, the system Refuses to run it.
Elena ReyesRefuses how? Like, it just stops?
BlakeThe post doesn't get into the mechanics, just says it won't ship what it can't support.
Elena ReyesFair enough. I'll take it at face value for now.
BlakeThere's also a before answer in there. Same show, once without the plan, once with it.
Elena ReyesOh, I wanna hear that.
BlakeI haven't listened side by side yet, but that's apparently the point of posting it.
Elena ReyesLet people judge instead of trusting the description.
BlakeExactly what it's there for.
Elena ReyesEven without listening side by side, just knowing that comparison exists changes how I read the rest of this.
BlakeSame. It's one thing to say trust the process, it's another to actually let people check your work.
Elena ReyesOkay, so this is the thing you flagged earlier. Blake, remember at the top you weren't sure if the AI host was reacting live or if it all got put together after?
BlakeRight. And I think this is the answer, or at least an answer. The plan is the mechanism that's supposed to keep the whole show honest. It decides what gets said before anything gets said.
Elena ReyesSee, I'm with you until the Refuses part. A refusal rule is only as good as what it's checking against.
BlakeMeaning?
Elena ReyesMeaning if the evidence it's citing is thin or wrong or just one blog post from the company itself.
BlakeLike the one we're literally citing right now.
Elena ReyesExactly like that one. Itll confidently Refuses to say something false, sure, but it can still confidently say something true and shallow.
BlakeThats fair. Discipline in the process doesn't guarantee depth in the sourcing.
Elena ReyesRight. It just guarantees the math checks out and the citations exist.
BlakeWhich honestly is still more than most shows can claim.
Elena ReyesLow bar, but sure, I'll give it that.
BlakeAnd to be fair, that- that's a low bar we're both climbing over too, citing one company blog post to explain a company blog post.
Elena ReyesSure, but at least we're saying that part out loud instead of pretending it's neutral.
BlakeSo where does that leave us on trust?
Elena ReyesGenuinely, I think it's better than nothing. It's not blind faith, and it's not full transparency either. It's somewhere in between.
BlakeI can live with somewhere in between.
Elena ReyesOkay, so process-wise, this actually sounds pretty Discipline. Plan first, math checked, sources reviewed.
BlakeSounds like it.
Elena ReyesBut none of that tells me what it costs or how long it takes to go from a blank plan to something we'd actually put out.
BlakeThat's a fair thing to ask about next.
Elena ReyesOkay, numbers. I'm ready.
BlakeMato's About page puts traditional podcast production at fifteen to twenty-five dollars a finished minute.
Elena ReyesPer minute?
BlakePer minute. So a 30-minute episode, they're saying eight hundred to two grand depending on who you hire.
Elena ReyesWait, that's just for editing and stuff, not even the mics?
BlakeThat's the range they cite. Studio time, editors, someone cutting clips, it adds up fast.
Elena ReyesOkay, so where does Mato land in that?
BlakeFifty bucks flat per published episode according to the How It Works page.
Elena ReyesFifty. Not fifty an hour.
BlakePer episode.
Elena ReyesThat's a rounding error compared to the other number.
BlakeAlthough fifty dollars a episode doesn't tell you what happens if you want more than the baseline. Extra revisions, a longer episode, more channels. The page doesn't spell that out.
Elena ReyesRight. Flat rate can still hide a bunch of tiers underneath it. We should ask about that too.
BlakeIt is. And that same About page calls the company a third way, sits between the expensive human crew and the cheap AI slob nobody trusts.
Elena ReyesRight. That's the pitch on the tin.
BlakeThe How It Works page also lays out the actual steps. Six of them, idea to published, and they say the whole thing takes about thirty minutes of your time.
Elena ReyesThirty minutes for the user, though, not thirty minutes total.
BlakeThat's the distinction I wanna sit on.
Elena ReyesBecause the machine's still chewing on it in the background.
BlakeRight. So what is a customer actually doing in those thirty minutes? Are they picking topics, approving a script, sitting through the interview live?
Elena ReyesOr are they just clicking six buttons and walking away while something else does the work for an hour?
BlakeBecause thirty minutes of your time and thirty minutes and the episode's done are two completely different products, even if the price tag is the same.
Elena ReyesOne of those is a genuinely fast workflow. The other one's just a fast checkout screen.
BlakeThat's what I wanna ask the next guest we bring on. Not the marketing version, the actual timestamped version. What did you do? When? For how long?
Elena ReyesBecause fifty dollars sounds great until you find out it took someone four hours of babysitting to get there.
BlakeExactly. Or it sounds great because it actually is thirty minutes and the rest runs itself.
Elena ReyesEither way, we should ask every single guest that.
BlakeEveryone. Same six questions, roughly.
Elena ReyesWe're gonna sound like we're doing an audit.
BlakeKind of the point.
Elena ReyesFair.
BlakeIt also matters for expectations. If someone signs up expecting thirty minutes total, and it's actually thirty minutes of clicking plus an hour of waiting, that's not a lie exactly, but it's not the full picture either.
Elena ReyesWhich is exactly the gap between a marketing page and an actual workflow log.
BlakeSo we've got the cost, we've got the claim time, and we've got the internal process from the blog post.
Elena ReyesRight. The whole documented refusal thing from before.
BlakeAll of that's fine on paper.
Elena ReyesBut paper isn't a show.
BlakeNo. So the process is fast. The reasoning's written down somewhere. Does any of it actually hold up when it's a real show with a real guest being made for real?
Elena ReyesThat's the only thing left to check. So does any of this actually hold up outside a blog post? Mato put out a press release recently.
BlakeOoh, receipts. What's it say?
Elena ReyesTheir founder, Alex Benz, is quoted saying they crossed ten thousand listeners across all their shows in thirty days.
BlakeTen thousand in a month? That's not nothing for a platform this new.
Elena ReyesRight, and it's the CEO's own number, so take it as a claim, not an audit.
BlakeFair. What else is in there?
Elena ReyesSame release says four enterprise customers, something like ten thousand dollars a month in recurring revenue.
BlakeOkay, so real money, but small money.
Elena ReyesSmall money. It also says over thirty active AI-hosted shows running right now, and a seed round they're calling pending for this year.
BlakePending meaning not closed.
Elena ReyesMeaning not closed. I'd want to see that actually land before I count it.
BlakeSame. Numbers that haven't happened yet don't count as numbers.
Elena ReyesExactly. But thirty shows and four enterprise deals, that's a company with actual customers, not just a demo.
BlakeThirty shows is also useful context for us specifically. It means we're not the only experiment running on this thing.
Elena ReyesRight. There's a whole portfolio out there we could eventually point to instead of just Mato's own numbers.
BlakeSpeaking of customers, their homepage has this one case study I keep thinking about.
Elena ReyesThe Wealth one?
BlakeYeah. It's called The Wealth Conversation. It's on their business tier, and the guests are local CPAs and estate attorneys.
Elena ReyesAnd?
BlakeThe site says those episodes convert two to three prospects into actual clients every month.
Elena ReyesTwo or three clients a month from a podcast.
BlakeFrom a podcast. For a CPA, one new client can be worth way more than fifty bucks an episode.
Elena ReyesSo the math isn't download counts at all.
BlakeNo. It's whether the guest walked away with business.
Elena ReyesThat's the number nobody puts on a slide, though. Nobody brags about downloads to their accountant.
BlakeTrue. Downloads don't pay your mortgage.
Elena ReyesAnd that's really the whole shift, isn't it? For a business show, the download count was always a proxy for something else. Turns out the something else is bookable.
BlakeWhich is a very different pitch than look how many people pressed play.
Elena ReyesSo if we're bringing customers on this show-
BlakeWe ask about that. Bookings, how long it took to land the guest, whether they got a client out of it, whether they came back for a second episode.
Elena ReyesNot just did the episode publish.
BlakeRight. Did it do anything? That's the standing question for every guest segment going forward.
Elena ReyesWe should probably ask the same thing about ourselves eventually. Honestly, whether this show is actually moving anything for Motto beyond just existing.
BlakeYeah. Add that to the pile of things we're supposed to keep checking on.
Elena ReyesAll right. I think that's the traction picture as it stands.
BlakeA young company, some real revenue, one strong case study, and a lot still to prove.
Elena ReyesWhich feels like a fair place to hand this to you for the wrap.
BlakeI'll take it from here. So if you want the actual change log, the blog posts, all of it written down, that's himotto.com.
Elena ReyesYeah. We're not gonna read you every line item every week. Go look.
BlakeAnd if you're a podcaster, or you're on a marketing team, and you're running a show on Motto or thinking about it, email hello@himotto.com.
Elena ReyesWe want you on here. Not as a favor. We want the boring details. Booking time, transcripts, what clip actually got watched.
BlakeReal numbers, not just ours.
Elena ReyesAnd if you've got a number that contradicts something we said today, send that too. We'd rather hear it now than keep repeating it.
BlakeSeriously, that inbox is not just for guest pitches.
Elena ReyesRight before we start grading our own homework every week.
BlakeSo that's the ask. Subscribe wherever you're listening and come back next week. There'll be another thing that shipped, another reason behind it.
Elena ReyesHonestly, I came in kind of skeptical an AI show could report on its own company straight.
BlakeAnd?
Elena ReyesI still want to see it hold up over a few more of these, but today it did.
BlakeThat's fair. That's all we're asking anyone to do. Watch it hold up.
Elena ReyesSee you next week.
More episodes
Keep listening
Other episodes of Made on Mato, newest first.
Sources
Where this came from
6 reports behind the episode. Every one of them opens where it was published.
- Introducing blueprint: every episode now starts as a planheymato.com
- Mato - The First Generation of AI Talent (Press Kit)heymato.com
- About Mato — The Future of Media is Liveheymato.com
- Mato Announces 10,000 Listener Milestone for AI-Produced Podcast Platformnews.marketersmedia.com
- Mato - The First Generation of AI Talent (How It Works)heymato.com
- Mato — 3× Higher Close Rates. 30 Minutes of Work.heymato.com
