How to Build a Calm Support Operation

[00:00:00] Alex Panagis: It’s a human experience and I think the same thing in support with software, ironically, which we try to dehumanize with AI and everything. It is at the end of the day, human experience.

[00:00:08] Antoine Minoux: Hey, we can take things slow. We can accept that we are probably not going to change the world, but we can do our little thing and do it really well.

Or like to give is we are closer to a neighborhood bakery that has the sense of craft rather than like trying to build a bread factory or something. Certain

[00:00:26] Alex Panagis: companies in the space and no, no names will be mentioned, but. You could probably argue that their own support experience is extremely subpar despite their AI solutions, you know, based on their own benchmarks being exceptional, supposedly compared to other AI solutions that it’s been tested

[00:00:42] Antoine Minoux: against.

I hear a lot of it paying for 5% of the product, but we’re paying a hundred percent for 5% of usage, right? And so this end up being customers feeling like, oh my God, I might either not getting the value that I’m paying for it, or also feeling overwhelmed by the sheer amount of features and settings. If you’re not even able to deliver on what you promised, what’s the point to just keep adding to that stack?

[00:01:04] Alex Panagis: Most founders, dread customer support. It feels messy, reactive, and expensive. But what if your support setup was actually simple, calm, and even a driver of growth? Today we’re diving into how to make that happen with Antoine Minu, the founder of a fernand, a support platform built for SaaS companies with multiple bootstrapped exits, including the sale of Improv MX last year to focus on this mission.

Voila, Norbert before that, Antoine has a battle tested playbook for what he calls calm support. We’ll unpack why SaaS support gets over complicated, how to design systems that scale without chaos and where AI actually fits in without introducing unnecessary complexity.

Antoine, welcome to the show, Alex. Thanks for having me. Very good. I hope I’m pronouncing the name of the company correctly. I did hear in a couple of other interviews you mentioned it was, I mean, obviously it comes from a, a, a

[00:01:56] Antoine Minoux: French name originally. That’s right. So it’s Nu, but I won’t mind if you say Fernando, something like this, so people give their own interpretation.

I think it, it’s actually a fun story. Behind the name comes from my grandfather, which was called, he was a Postman, and so I wanted to honor his memory also. For the fact that I’m working with my brother-in-law, so married my sister. So there is this family aspect and yeah, we thought that Postman was the nice little metaphor to a customer support help desk software, like delivering messages to your customers.

So that’s why we chose this.

[00:02:30] Alex Panagis: I love it. I love it. I do have a question about what it’s like working with family for later in the interview, because I’m curious about that as well in typical true operators fashion. Let’s dive straight in. Starting with your core philosophy of comm support, could you set a bit more context for us and explain what comm support means in practice to you?

Sure.

[00:02:47] Antoine Minoux: So there is this whole concept of calm companies and comm software, which is taking the counterpoint to everything you’re being taught in the, especially in the VC world and Y Combinator. And there is this new thing right now where it says you need to work 9, 9, 6, like from 9:00 AM to 9:00 PM six days a weeks.

This would be going against everything we’re trying to do in, in. And with my company in general, with my life in general, which is, hey, we can take things slow. We can accept that we are probably not going to change the world, but we can do our little thing and do it really well. I like to give is we are closer to a neighborhood bakery that has this sense of craft rather than like trying to build a bread factory or something.

So with friendship, this speaks to us really well. And so taking this metaphor closer to the customer support thing, which is what we do at Fanon, I think for many years the industry has been set in a way where the people that are managing or that are buying the software are not the people that are actually doing the work.

And so what tends to happen is that those manager, head of operation or whatever, they will spend their time in reporting decks and like trying to optimize every aspect of. The time to response average and satisfaction scores and stuff, which ended up completely, I think, in my opinion, breaking the picture so they’re not getting the full picture of what actually delivering good service means or is.

And it’s also stressing out, um, agents so that, oh my God, I need to improve my time to response score. So I’ll just respond anything. And the, what the philosophy of our product is to break away from this and to allow agents to do their best work to feel supported by opinionated product design that just helps them stay in flow, stay focused, and reach inbox zero at the end of every session.

So yeah, this is what we are trying to do. I think we’re the only one doing this in the customer support space at this stage, but you can know similar products may be like superhuman for email. There have been other software in the industry that have tried to take this aspect. But yeah, again, I think this is fairly novel, something we’re trying to.

Push little bit more because this is what we’ve experienced. Like we’ve experienced the pain ourselves as funders and that’s what we wanna bring to market.

[00:05:16] Alex Panagis: Awesome. I think it’s very much in sort of, I mean the whole calm philosophy and one of my favorite books is It doesn’t Have To Be Crazy at Work from Jason Fried and David from 37 Signals.

So do you think a lot of the product thinking, and I, I think you’ve also cited that in, in certain places, comes from building your product, Fernand for the company that you want to run. So you want a support tool that works the way you want it to, and it’s built on the premise that if you build it and you like it yourself and it works for Fernand or in, in your case, also would have worked in the past companies that therefore there are other companies like yours that don’t need what, you know, enterprise grade software has, where as you put it, you’ll have like six column layouts with panels and widgets that are all fighting and competing for your attention all at once.

And I it like, I guess. The short version of this question is how do you balance that vision and staying on that vision with ultimately wanting to capture more and more market share and grow as well?

[00:06:11] Antoine Minoux: Yeah, that’s something that definitely came to us because I was so optimistic about this product. I’m like, oh my God, we’re gonna be right.

The market, it’s so good. I think it’s reasonably priced as well. And when we entered, when we launched, it was like quiet compared to other products that we launched that had this initial traction and almost like any instant product market fit, this was different and it, it felt a little bit like, huh, was this problem only applying to us or just a very subset of, of that niche?

And I think part of it was, yes. So this is a little bit of what I told you about like the people buying the software, making the purchase decision of migrating to something. I’m not the one that are necessarily experiencing the pain. So we needed to refine our ICP towards more like smaller team solo founders, this kind of people that actually do customer support and care about getting great tools.

Um, but in general, I would say this is a more like a bigger problem in terms of the industry size. So it’s a very competitive market and so it’s more like driven by sales and stuff like this. We have bootstrap company and we don’t really have any marketing strategy or sales strategy, so we were just hoping build it and they’ll come, which people make fun of, but in my experience, it works.

Still do something great and it’ll eventually work. What we realize over the time is that for a very competitive market, this is not true. You have the barrier to entry to first have people know that you actually exist, which is a big one. And the second one is that if they even know that you exist, how do you communicate your value proposition in a way that it makes it so appealing that they.

Do, I don’t know, like months of change management to adopt your solution. So yes, this was coming in, but the idea in the beginning was yes, like we are annoyed with the solutions we’ve tried and we want to have something better cut the bullshit. Like just do what we would love to have and we sub, okay, let’s just build it and we’ll see.

So that, that’s how we came to this.

[00:08:20] Alex Panagis: Nice. Yeah. I think the build it and they will come is a, is a very like funny concept, but I think your take to it is not the typical, like, we saw this market opportunity and we’re gonna build it and we’re gonna have some marginal improvement or build, you know, one or two extra features.

I think you do have, at least from, in my view and from my perspective, you’ve approached it with a, a unique take on the problem and also that has reflected itself in the product. So the philosophy of support being a calm function in the company and in general also the wider company being a calm company as opposed to.

A metrics driven company or a company that doesn’t operate in that way. And not to say that com companies shouldn’t look at metrics, I think it has manifested itself in product and design decisions more heavily than most other founders ever have the luxury of paying attention to. So if you could talk perhaps a little bit more about that, because I know you’ve, you’ve spent a lot of time reflecting on details like the choice of colors, the positioning of elements in the UI and what, how much information you’re showing at any given time, which I think is not like in a typical support software, it’s probably if someone is trying to compete from a business point of view, they’ll just be like, okay, this is what the competitors do.

This is the data that they show. Let’s show that data and call it a day. Yeah.

[00:09:30] Antoine Minoux: Yeah. And I think what makes us more appealing or at least appealing to the people that that use us and recognize the value of what we’ve tried to do is the things that we didn’t do. So there are lot of competitors in that market, as I said, and there is always this race to feature bloat.

So how many integrations do you have? How many reporting metrics you offer, like now, especially with ai, like how much AI can you pack in every corner of the product? Some of it is great and some of it we do as well, but I think from the demos that I’m doing with our customers, I hear a lot of, we’re paying for 5% of the product, but we’re paying a hundred percent for 5% of usage, right?

And so this end up being customers feeling like, oh my God, I’m like either not getting the value that I’m paying for it, or also feeling overwhelmed by the. Sheer amount of features and settings that you have in the thing and feeling overwhelmed that maybe they might not be using the software in the right way or like harnessing the power of it fully.

Or even like as software company goes, things tend to break more as your product surface increases. And yeah, you have companies that release something daily, uh, but where the actual experience for the customer that is on the platform day in, day out, which is what we are optimizing for, is not that great.

So you have a new bug to chase every other day. So that’s also another below or really core value of Comm software is that we’re trying to release things in a way that’s gonna be very smooth. And even though we are fast and we release stuff all the time, we’re really going to try to make it like an appeasing thing.

A linear has this idea of if there is a bug, we should fix it right away. And I think it’s definitely. I think that I would abide for if you’re not even able to deliver on what you promised, what’s the point to just keep adding to that stack

[00:11:33] Alex Panagis: yet? It’s what most software companies, I think, do, they have like a growing list of complaints and bugs, but because they want the marketing hype, they chase the next big release.

So o obviously AI was going to be a bigger part of the conversation. I think it’s hard to avoid probably in any conversation. I think it’s, I, I have a, probably a similar take to you. I think it’s, especially in support, it has made waves so early at the advent of ai. I think now a couple of years ago, it’s crazy how things have passed.

Klarna announced like massive layoffs of support staff and there was similar companies that followed suit and then some decisions were eventually reversed. And I know you have an opinion, so it’s a multilayered question. You have an opinion of Overstaffing on support because it’s a particularly important function and you want to be able to spend time ultimately talking to the people that pay you when it comes.

Ai. From your experience, how do you decide, you know, in what way does that manifest itself? So I think right now you have smart replies, which is drafting, but it’s still largely human initiated and led. And you have obviously the context and the library that goes along with that so you can feed into the model.

Do you think that the more automated tools don’t really serve much of a place? Or how, what, what is your view on that?

[00:12:45] Antoine Minoux: Uh, we also released about three months ago, the ability to have an AI agent that takes over the first level support in your chat. We also let you draft pre-made responses so that you can just review something, touch up a sentence or two and do it.

You have, write an answer with smart replies based on a system prompt, and also for your knowledge base, help you write, uh, knowledge base articles. You see, the common thing with all those four main features is that you are always in control as a human. So there’s this idea of human in the loop, right?

There is nowhere where the AI is really fully taking over. The closest to this would be the AI agent in the chat, but we’ve made it so clear that you can always escape this by talking to a human. We have this huge button on the chat that says, talk to a human. If you chat and you write this and express interest to like being transferred to a human rep, you’ll be getting this kind of behavior as well.

And I think this was really important for us because AI at the moment for customer support looks great for, I’d say e-commerce. Anything in retail, anything that’s very transactional. I don’t know, airlines, flights or whatever, train, any kind of ticketing, anything that is easy with scheduling. For example, booking calls, cetera.

Arish software is SaaS in particular. In SaaS, you’d get message around pre-sales question, which AI would be terrible at ending because they just don’t have the capacity to like make any kind of decisions around. Maybe they’d ask a little discount or they’d want more clarity on certain feature. Usually the answer would be ready generic, even though you’ve trained it on a lot of knowledge.

And usually when people reach out it’s because there is an edge case that is not easily like explained by your product side or your marketing or your knowledge base or whatever. And the second thing is usually bugs. Things that don’t work, things that are a little bit hard to get from your product, and those things are way harder to automate.

So in some sense, like the one we had before improve mx, it was much more infrastructure based and we had a lot of recurring questions about how do I set my DNS, how do I set my MX record? Um. What are the steps to go and change something on GoDaddy? Right. Or this thing doesn’t work. And with ai we were able to automate a lot of those questions and because the support flows are fairly big, then you’d had, um, pretty big gain for, we have that first level support, but I’d say right now we are getting a 70% hit on those things and the rest is fairly meh.

So I think a lot of the people that are in my position are like founder of a customer support thing. They would try to hype it like, oh my God, this is so cool. This is like absolutely the future. This is amazing. Yes. But I’m always thinking there is a reason people want to have a human interaction. And again, this whole business that we are in, this whole niche is sales driven.

Why? Because people want to deal with humans when they’re making like purchasing decisions. So it’s just. It tells me a ton, although I wouldn’t also not be wanted to fight it or not fight it as much as I was fighting chatbots in the past. ’cause chatbots were really like lower level kinda decision three things.

They had no logic, no reasoning not what you can achieve with an lambs and yeah, rag to people like knowledge from Vector database is infinitely better. But I think you need to apply it in a way that makes sense for your business and it’s not a one size fits all kind of thing. Um, one thing we also took in mind, and I think that’s very important, is the ability to customize your brand voice and tone.

So I’m not speaking. With my customers the same way that a, that a bank would speak with their customers and they might not be the same compliance and like negative romping kind of things that you’d want to block as, as guardrails for a conversation with a startup and a conversation with whatever government.

So it’s really in the edge case and the details, and I’m not sure if we are really educated properly at the moment to really implement that. Right? So there might need to be a few years for us to go through all these mistakes and so that we build a muscle both to set it up and to use as a consumer.

[00:17:08] Alex Panagis: Yeah, I think from a consumer point of view, I think the, you’ve kind of hit the nail on the head. Certain companies in this space and no, no names will be mentioned, but you could probably argue that their own support experience is extremely subpar despite their AI solutions, you know, based on their own benchmarks being exceptional, supposedly compared to other AI solutions that it’s been tested against.

And it’s. It’s not so much necessarily that the AI responses might not be actually objectively superior compared to what other tools would give it, but it is that, I think the lack of the escalation path and the understaffing of support that people have been able to do because they think, okay, now we can just deflect a whole bunch of queries.

So what ends up happening is you, let’s say previous before ai, you’d have a customer who reaches out and yes, they have to wait 20 minutes for an initial response, but they get the initial response in 20 minutes instead. Now, what I’m finding more often in companies where we look at the support function and whether that’s playing a role in stunting their growth, it’s actually, oh, they’re using ai.

The, the, the customer is getting a response instantly within two, three minutes from once the AI has searched and done all the tool calling that it needs to. Then if they have a follow up question that they want to escalate, it takes like an hour or two hours to hear back from support. So the end user experience is like, okay, well I’ve sat here and I’ve gotten the response might get you 50% of the way there, but it’s ultimately still not the connection to someone in the company, which is what most of the time in, at least in SaaS, I would say people do want.

[00:18:32] Antoine Minoux: What I’ve also noticed with our customers is that before ai, they were the support function really tightly linked with customer success and then they eliminated a bunch of roles, meaning that the more like the proactive reaching out of, oh, why are you not using this feature that would be beneficial for you in such a way.

Uh, I don’t have any data, but I’m pretty sure this closes a fair amount of churn. And again, like missed realized potential just because the people you needed were not the cost center you thought it was, but were more like, like having real genuine connection with customers. Usually also you have customers that are pretty, um, silent and you never hear from them and maybe they are like thousands of dollars of annual contracts running, never hear from them.

But then you have those smaller ones that keep growing and you have the relationship from the start. They report a ton of stuff. You help them, you build that thing and they will never churn anywhere because they feel like, oh my God, I’m, I will not be able to talk with this person or this person about this.

You get a pitch from a competitor, but no, sorry, we are happy with X and I think this is what AI makes super impersonal, right? Because I don’t care if I’m missing this relationship ’cause I don’t have any. And I think it’s also a much bigger question about like your attitude towards bugs, for example. So how do you react to a bug when you’re a customer?

You fill that into the chat bot, hey, this doesn’t work. And you get an a nice answer from the bot like three seconds later saying, oh, I’m so sorry. Yeah, I’ll pass it on to the team. Okay, cool. You are a voicemail like responder, which. I guess like just confirmation that my email was sent. It’s a similar experience than what I had before.

Right. And as a consumer, the difference is I’m not even sure that it’s reaching anyone ’cause you know that it’s an automated thing so it can cause like ton of different frustrating experiences but I’m not really sure we are there yet. You mentioned Kna, they’ve rehired uh, a bunch of people. Mostly for this like higher enterprise client I think and customer success.

I think we will need humans as long as software exists. Maybe software will not be software as we know it in a few years. You might prompt, I want to count and outdesk for my thing with an integration to at heel and whatever and we’ll just like magically vibe code it and you’ll be able to use that maybe.

But then I think we are still a better away from that vision. I,

[00:20:57] Alex Panagis: I, I think that they’re still quite far. I think like with most things, the, the, the presence of AI has just raised the bar because previously getting a quick response from a human to acknowledge a query was something that would impress someone, and it would largely impress a consumer because they knew it was an actual human who had read the request and said, oh, I’m really sorry that you ran into that bug, explained, verified that they understood what bug they’re talking about and got back to the user.

When AI does that, it has no meaning to the end consumer largely. It is, like you said, just basically saying, thank you for submitting a ticket, thank you for filing a bug report, because that the, the actual person processing the information and getting back to the other person, it’s completely lost in translation and I think that, like you say, support is probably in hyperscalers looked at as a cost center most of the time, or almost definitely it would be looked at a cost center as opposed to a profit center, whereas those micro interactions where you have a good experience with support.

They have a really big impact. And I think one of the only massive companies that probably does this incredibly well at their scale is Amazon, where they have somehow, and they have implemented AI for parts of it now as well, but their support experience is what actually makes people feel confident in ordering from Amazon versus your, you know, mom and pop shop retail brand that is local.

And that’s unfortunate for the local retailers. I’m not saying I necessarily su don’t want to support local smaller brands directly, but that’s obviously the mode that they’ve built over the years. Which brings me to my, my question on related to improv MX more so because you had a free tier for that product, I think you’d probably agree that the support function and the way you build it is heavily dependent on the business model.

So if you have a free tier, it affects things quite a bit. How does that differ? So if you’re now, you don’t run improv X anymore, so you’re not running a free tier product, how would that, your mindset to, you know, using AI deflecting support queries and also the level of. Support and attention you provide vary if you have a free tier versus if you don’t have a free tier.

[00:22:56] Antoine Minoux: Actually having the free tier was one of the reason why we’ve built funnel because none of the customer support tools out there. Intercom grew Val as well. Sout whatever the pricing was built around the ID that your users get you certain value, which was not the case for us. Well, you had a hundred thousand users free and it was making no sense to pay for them as like you, I think you had a account, they might have changed our pricing since then, but like this was the case.

We made this actually moat for improve mix, so we wanted to offer this same or an at least agreeable, good level of support for free users to be able to create that like nice and best ship like experience to the brand. And then usually what we would do is we’d have an automated follow up after a good support interaction all set up with internal and our foreclosed thing to say, Hey, it seems like your thing was solved.

We’d really appreciate if you could leave us a review on G two or capa, whatever. And this has fueled a lot of our organic growth and obviously bring also paying customers. And so I think we saw it as a way to grow, which is going against the idea of cost centers. We also saw it as a way to do the right thing.

’cause we just wanna be like nice people and be nice to people that are using our solution and actually solve problems. And also we used it as a way to make the product better for everyone because the bug then the the edge cases that gets submitted by the scale of people that you have under the free plan.

Makes the product better for the ones that are paying as well. So they’re like different angles to to this. You’ve mentioned it before, but we also made the decisions to completely overstaff support back in the day. So I think we had at the top of the, the growth thing, we had, I think three support reps, plus myself and my co-founder were also looking at customer support.

So you might think five, five people for maybe around like 30 conversations a day, which is insane. No one would have that many people. But what it would do or what it did to us is that we could really get to the bottom of everything. So if out of the 30 you have 10 easy ones that are just things that you say, yeah, sure, I’ll always reset your password or whatever.

But then you have three or even two that are long investigations where you need to involve a software engineer and you need to go really dig deep down into, there was emails or deliverability stuff and like IP reputation and whatever. Then this can usually go and take 2, 3, 4, 6 hours. Do you really wanna spend six hours on a ticket for a free customer?

And the answer for us was yes, because this problem is usually going to either come back or is bigger than it looks, or it’s not isolated to that customer over the time, of course we made up some guidelines to say if you have an issue with your domain registrar or your DNS or whatever, we’ll push you towards getting them to support you.

’cause we cannot offer support for Gmail, for example. But we were still very nice about it and very guiding. Um, so yeah, I think you can turn support into a mode, like an unfair advantage almost. ’cause no one does this well. Like I, Amazon is a really good example where I felt like I had really over the top support interaction in the past.

I think linear was one where I felt really good as well and I asked something, we built an integration and they were like, yeah, sure, we’ll hook you up with a software engineer to answer your questions. Like literally in days AEO is one as well. It’s this CRM tool that again, I had access to actual engineering the company to, to the stuff like, that’s great and no one really does this, but you’d be surprised how the bar is actually so low because no one’s want to really deal with this and don’t understand the long-term value of providing really good support.

Personally, I know every of our customer almost by, by name, um, or at least I would know vaguely who they are when they reach out. That’s something.

[00:26:58] Alex Panagis: It’s, I think it’s like a beautiful scale to run a company at, and it kind of, I think it, it, it’s a unfortunate thing, but also a luxury in a different sense when the company goes past that.

So I remember early days, well, early, not that early, because Basecamp has been around for over 20 years now, if I’m not mistaken, where we were Basecamp users, and that those interactions are similar and they still go above and beyond. We’re no longer a Basecamp user. But I, I do remember before leaving and switching to Asana, now that they still went far above and beyond.

Now with our current project management system, if we can even get a hold of someone, it’s, it’s like, it’s not, not even a matter of reaching out, we don’t even try to reach out anymore because we have sent in tickets and not heard back. So I think that like element of, you have the luxury when your product is so strong and has such a moat that the support can actually be quite frankly, awful.

And people still don’t leave. And I hope that most people watching this wouldn’t strive to build such a company. But I think there’s then the converse, which is that the product. The engineering team isn’t at the expected level, which actually makes support far more complicated. So for example, for example, you mentioned the 30 tickets a day.

That’s to somebody who’s worked support, which I have as well back in the day, and I, I actually find that everybody who works in a software company should spend time working. Support can be quite a high number because it’s the nature of the requests. So if you’re a support engineer that’s dealing with, or just a person in support dealing with 30 questions, but they all relate to stuff that you have no control over and requires somebody on the product or on the engineering team to actually ship a change or make an improvement, you just, it, it’s incredibly anxiety inducing because you have, I mean, if you care enough about your job, you end up being so attached to wanting to get the customer the outcome that they want.

So if you can to do that, it, it’s really frustrating in my opinion. Do you think there’s a point or a type of product maybe where running a free tier and offering this level of support works and then maybe at one point it just falls apart? Or do you think that it only doesn’t work if the engineering function basically can’t ship stuff the way they need to?

[00:29:02] Antoine Minoux: Yeah, both. I think both. So like with Improve mx, we were really starting to question this, like this obvious mode that we created and oh my God, this is a ton of resources. We are playing this and we are not really sure what we are getting on the other end. So for us it was like, it’s a nice thing to do, but is there a reason why literally no other company does it this way?

And so it was like, you keep mentioning Basecamp and I think Jason Free and David and my Anson does this contrarian thing a lot. So like everybody does something, they’ll just do the opposite. And then. For sometimes like the opposite that they are leading with becomes the norm. I’ve seen this happen a few times because it’s just a logical thing to do.

Why should we get our business advice from all the VCs and stuff? And like the playbook that works for them is not the same. That will work for a two person bootstrap company most of the times. Also, just because you cannot push that much like dollars towards the problem, which is what they do. Oh my God.

We want to be the Amazon of customer support. Yeah. Okay. Just hire a bunch of people, create a hundred percent team, have the nicest offices and whatever. And maybe you’ll get to this although matter, but yeah, in terms of having a free tier, I am on X or Twitter and like people keep saying all the time, get rid of your free plan.

You don’t need this. This is the best thing you’ll ever do. And I think, I don’t know, I think it’s all a matter of experience design. So for Fanon. You are into a product that needs a lot of change management. So you need to change, usually you already have something, so you need to migrate your data, you’ll need to train your employees, et cetera to use the new tool.

Um, so this comes to the fact that I’m doing personal one-on-one demos with people. I’m trying to explain the value and what we do best and stuff like that. That’s why also our marketing side is super outdated, need to redo this for a year, but it’s not that important ’cause spending more time with people on a one-on-one basis.

And then you have people that are more self-serve where like for example, my co-founder has A-H-T-M-L to P-D-F-A-P-I. So it’s like you feed it HTML and spits out A PDF. Great for things like invoice generation or ticketing or whatever. And so for him it’s super easy. He advertise on Google ads and people come to his website and they buy an API key or a subscription and then boom, they start to generate PDF.

And I think for him, having the free plan is just a way to reduce that friction. To actually generating the PDF and making sure it works. If you take a software engineer, they just want to try a quick proof of concept, how quick it is, how easy it was to implement with their library or whatever. And so if they wouldn’t have the free plan, there is a case in the journey where the software engineer comes to the fact that they have to put a credit card.

I was like, I’m not ready. I don’t know if it works yet. So you’re chain breaking that chain of value. And I think for me, this is what the free plan should, should get you, is to really get to the wow, the aha moment, the value thing. Then you could argue, yeah, cool, let’s just have a trial, right? Which is what we are doing with Funnel, have a 14 day trial, so trying to replicate this.

But I think for Serial, the costs are fairly low in terms of infrastructure to offer a plan with, I don’t know how many, I think it’s a hundred, maybe 50. Conversions per month. What allows him to have like a pool of customers that are using the thing that are maybe also ambassadors. So ask them the same play call and write a testimonial please.

And you will, I’ll hook your account with double the credits for this month. Um, and just to appear nice because you might have a person that is using this for their solo like side project for their soccer club, but then they work at a big software agency or whatever and they will use that in production for a big enterprise account.

So I think it just grinds the gears and like always the gear actually to put you in the position for later. So I would say it depends and it can work. And in regards to support with ai, you can do a lot more, much faster than not automating everything. But what we do is pre drafting answers like smart replies.

Prompted a few words like decline politely, and you get like this nice message, you can do much more than what you could have done four years, five years ago. Um, so I think I would still advise, read here as something that you should not dismiss as quickly, just because a few gurus on a social network have said that it was terrible.

[00:33:32] Alex Panagis: Yeah. I think I completely agree with your take. Like it’s, it’s such a product question. It’s not just a matter of how do we, like, what do we want to do? What type of company do we want to run? Because Fernand is exactly like you say, it’s an alternative. The people who are buying it. Are likely going to be setting it up based on your target.

You’re not trying to sell to a developer in a, you know, 2000 person organization who’s eventually gonna pitch it to somebody else. So there needs to be a, you know, forever free tier. So I completely agree. I actually have a very specific example which I think is worth bringing up because it’s a product which has a free tier and the benefits of the free tier are very big.

I would argue it meets probably almost all of the criteria that I have in my list of like things I would use when validating whether a company or doesn’t need a free tier, which is senia.io. It’s a social and testimonial platform. The founder, Ali, who’s actually also based in London, is quite active on X as well.

I think they’re in like a very interesting position because I think a, any founder and any consumer can say, yeah, that’s a product where the, so the, the presence of a free tier makes a ton of sense ’cause you have the viral loop built in. But there is also that, and I think he admits to it as well, that it is more so right now a pro temporary product problem.

That having the free tier, despite it probably being the biggest source of growth that they have, is causing a huge burden on support. So there’s that element of like, okay, we can shut off the free tier, which is basically shutting off our growth and our viral loop and shutting off signups, which could kill conversions.

Or we keep the free tier, but then we possibly struggle to deliver expectations for the entire user base because we have too many free users that are also reporting bugs, and we’re just not really able to, to keep it out. I, I think the, i I, not to speak for you, but I think the unfortunate reality in that situation is that there’s no easy way out.

Like both, both things are a, are a bad choice, and the hard truth is that it just has to be like, how quickly can you solve the product stuff and get back to providing like a really good experience? But I’m curious if you have different thoughts.

[00:35:32] Antoine Minoux: Yeah, so I’m actually close to this because Senia is using Fennel as a customer support platform.

So I’ve been having quite a few conversations with, with Ollie and also Wilson back in the day when he was still working there. I think actually this is one thing that doesn’t play in their favor. So the technical co-founder Wilson has left the initial like partnership. So it means Oli, which is more like Marketing Master is known, the sole proprietor and like kind of solo founder trying to get his way into building a technical product.

’cause it’s still a very technical product. So I think this is a very difficult thing for him and for anyone honestly, to navigate. It might just be temporary until he finds a way to bridge that gap. But it’s just really hard to operate software company when you’re not technical. So what will happen is that bugs will pile up things.

And decisions that I’ve made by the previous engineer might not be understood by the person that you recruit next, or like even the co-founder that you might bring in later. So the people will be tempted to refactor, and a refactor is not something that usually grow makes you grow. Or it could even, in my experience, cause more issues then provide answers.

Mm-hmm. Um, so I think they’re just in this middle ground that’s very specific to their situation. If we take the metaphor a little bit higher and imagine it’s any kind of company, I would say that you need to look at the gross curve and how fast they’re scaling. So taking like the example of Sandra, they’ve grown massively quickly compared to the few examples that I have.

Meaning that they will be always growing pains, right? This is just almost a normal human thing. We’ve experienced the same recently with Phenomenal where we started. Getting into database logs and stuff because the sheer amount of data that we are producing is getting so intense for the resources or the infrastructure map that we had planned or the architecture.

And so we needed to start thinking about ways to reduce that and load the data differently, et cetera. And I think it’s okay if we would’ve optimized this from the get go, we would’ve never released the product or we might have optimized it anyway in a way that’s not efficient because it was just like theoretical.

So you just need to get through these phases, in my opinion, and that this will happen and it will get resolved in real time, which is all also why I am skeptical about like AI completely being able to vibe code a purpose made tool for you and be able to use this. Obviously you would not really have the scaling issue ’cause it would be restrained to your own environment, but so many things that you need to keep in mind.

To have something run and run well and does really what you want it to do. I think we’ve diverted a bit from the original question, but yeah. This is a topic that keeps going back to my head because you could be the best operator ever if, if the industry is changing as quickly as a customer can just replace your solution with typing three sentence into your chat box, then maybe we need to rethink what makes us unique and how can we help company deliver the mission that we have, which is for us like communicating with your customers in a way that is anxiety reducing or at least more focus, and so that help you deliver more value to your customers by being more in sync.

And maybe there are other reasons or other ways to achieve that than providing the software that we have right now.

[00:39:06] Alex Panagis: Yeah, I, I always think, I mean I’ve obviously we’re, as someone who works in with software companies as well, that’s our entire. Business model is largely dependent on the software industry.

I spend a lot of time thinking about, okay, what if AI and my brother, who’s a software engineer as well, or head of engineering at a startup in London also is, okay, well where is this all going? And I think, I don’t necessarily think this is like a, you know, one liner that answers the entire question, but if you take any category of product you have in general, you have a big incumbent that is commercial and then you’ll have some open source, sometimes open source with a commercial angle, but often not even a commercial angle to it.

Alternative. So like for example, in scheduling you have Calendly and then you also have kyle.com, which is com commercial open source software. The number of people that actually choose to self host kyle.com has to be fraction. And so nominal in the total like number of people that use Kyle and eventually like sign up for the generous, free tier of Kyle and then therefore enter the conversion path.

And it’s also the element of. The maintenance of the tool. Like even if you have the whole existing architecture scaffolded like you do for Kyle, it’s not really like that easy to come in and say, I want to ship an extra feature purely with ai. At least not yet. I think that element of it, like building widgets is where things will probably get a lot easier, which is why, you know, let’s say for Fernand the, the way I could see this manifesting is if I have my own billing system that isn’t Stripe.

Let’s say I want to integrate with Paddle, which I think you actually do integrate with. Yeah, yeah. But hypothetically you have a user who has their own billing system that isn’t Paddle or isn’t Stripe, which you don’t already integrate with. They should be able to vibe code or using AI very quickly, scaffold the basic integration that lets them pull in the customer details into the rest of the product, which is built and carefully maintained, not just some sort of AI vibe coded solution, in my opinion.

[00:40:56] Antoine Minoux: Yeah. Yeah, that makes a lot of sense. I think this is a future or a political future that we are going into, but I’m still. Wondering about this whole infrastructure, compliance, security, maybe even a little bit about legal. Like we’ve sold a company to a US company and there was a lot of GDPR and data privacy, US privacy shield kinda thing going and like literally selling our data to a US entity.

So I’ve seen, I think I have a bit of a trauma after this and I think AI would just not do a really good job at, um, at this in the time being, although you could say this is just a skill issue, it is just like better prompt engineering. You could do it better maybe. But again, I think we not, we just don’t know how to use the tools really well and it’s not even our fault.

There is a new tool coming out every month, every week, even every day. So, or even the tool that are using the model is changing. So how are you supposed to be like great at something and mastering something when I think the common perception, I think 10 years to master something or master a new skill.

We are very much in the infancy of this, and so it’s fine. I’m looking with a curious mind and the same with support, right? Like we are trying not to jump onto something too quickly, but if we see something that is really helping and helpful, then I will look a little bit closer. So right now I’m writing an entire integration guide with open AI agent kit so that our customers could build a purposefully built agent for them that can just ask qualifying questions, check it under their database, then come back with something like build bigger, more complex agents.

And it’s something that is not at the core of our mission. So if they can offer it for us, then so be it. I think an intercom for example, they will try to build that in their offering, which is great, but maybe what people want is really just to use their open AI key where, which is already trained with their model and whatever.

So I think there will be different paths and we might see where this goes.

[00:42:56] Alex Panagis: I’m, yeah, I’m super curious to see that feature and try to think that we, we’ve also from a security standpoint, generally steered clear of giving too much tool access because I think there’s, it’s just too much of an unknown and at least I think most mature companies are at that stage.

So I’m curious to see how your guide addresses like, okay, what is the actual limits? Because even with the agent kit release, I think there was, I mean, they have built in ways to create guardrails, but I remember like the day after some researcher managed to like break through the guardrails. So I definitely think we’re at like a stage where every, almost every day, like the, the reality of what’s possible, people are just finding out because the tools haven’t been around long enough for us to actually explore.

Not only that, the tools also happen to be non-deterministic, so that you can’t just try like a finite number of things with them, which certainly doesn’t help. I have a very specific path to go down, which is kind of related to the freemium discussion, which is the scope of support. So I think with improv X, it was a, perhaps a simpler product.

So. I have a more artificial question to ask because scope of support is something I find manifests itself in very different ways in very different companies. So if you’re like a cloud server management platform, how do you actually go in and diagnose performance issues or that’s considered application level support?

So no, we wouldn’t touch that. I think in pronouns, example, the like simplest one I could think of is if you were to allow customers to, you know, customize their email header and footer to add a custom design so that all support replies are sent from a branded email instead of from a plain text email.

Not that that’s something that I recommend because I think the plain text is perfect and it should be conversational, but in the hypothetical sense that this is a feature, you then obviously will have customers that don’t know how to create A-H-T-M-L template. So they will do it and then the formatting will break and then they’ll try it on like five different email clients and they’ll realize that the logo doesn’t appear the same size.

Like that’s just one example of how one feature could lead to scope creep in support. I’m curious how you draw the, the line in Fernan I, I think product wise you’ve designed it so intentionally that you probably don’t have that issue as much, but maybe in improv X you have a story to share.

[00:45:03] Antoine Minoux: So actually I have a story with Fernan, so I think this is your biggest superpower, especially as you get started.

So we are still small enough that I know most of our customers and we don’t really have like huge scaling issue as in like we get a hundred new sign up of paying customer to support per day. If we get into this, I would probably change the approach, but right now what we are doing is like super wide globe support migration, like free data migration in the case where it makes sense and we can support it.

And for your case, I go as far as like when we’ve migrated the knowledge base, for example, I’ll review manually. That the knowledge base has been imported properly, I’ll set their brand colors I’ll like even resize their logo and upload it in the best format. And sometime I even see that on their marketing landing page they have some typos or they have a thousand pixel logo that could be 99% smaller in size and that probably is causing them SEO performance issues.

And so, so like I would go as far as writing them, Hey, I’ve optimized that asset for you. You should also replace it on your page because reason A and reason B, so we are using it again as a mode, but also because the market that we are in demand Sta building that relationship and establishing yourself as a trusted partner, especially for us as we are not like a powerful powerhouse of like hundreds of people we have to do.

10 times more. The effort to show the transparency, the respectability, the fact that yes, if our platform goes down on Christmas Day at 3:00 AM we will still be there and show up and open an incident and investigate. And I think this is what your customers want to know. Do you have my back? Or are you just a fun little toy?

Um, that can break easily. I think that’s what I felt as a bootstrap company in a sea of VC funded or bigger options.

[00:47:12] Alex Panagis: Yeah, I, I think in the, the converse in the VC world, it probably, you know, how much can we get away with providing a lower level of support because our product is sticky enough. People have built so much in terms of workflows, so they’re so tied into not migrating away.

So we know we’re not gonna lose them. Even if they don’t, we are not able to like, help them build the workflow one-on-one.

[00:47:32] Antoine Minoux: I agree. I also think it, ’cause it definitely would scale in my opinion. Like I would rather hire people to do it at scale. Than to lose it completely. So I don’t know what the future is made of, but that’s probably the route we would take.

And we already had someone to offload me doing demos. We had two, two agents doing, uh, as contractors to take off my plate and also for time zone issues and stuff. So some people could have more open slots to book demos. And I think that’s really one thing that I would like to retain because just, just the nicer thing to, I don’t know how to explain this.

This is, this is the kind of company that I want to build and I don’t, again, like I want to be the friendly bakery down the street where you pass by and the bread smells nice and they’ve used the ingredients from um, I don’t know, like a hundred kilometer round and maybe when everything is sold out, they’re not baking again quickly.

They’re like, sorry, we’re closed for the day. We cannot serve you today. Come back tomorrow. This is the kind of business that I want to

[00:48:29] Alex Panagis: build. Yeah. And instead of the one which I guess that the counter in the bakery analogy would be the one that plasters advertisements everywhere and just like, looks super.

Good from the outside, but then ultimately they just, you know, are using like not the same quality of ingredients. So I think the exactly,

[00:48:46] Antoine Minoux: like I have a perfect end of this metaphor. Like my wife said, Hey, let’s try this service. We live in Hamburg, in Germany, and there is this thing one A them. It’s still great.

They can deliver bread to your door every morning at 6:00 AM And she’s like, this is great. We could get bread every day. So I try it. Actually, they have a free trial, which is pretty fun. Pretty bold. Yeah, bold for something that’s physical and that incurs a lot of costs for them. But okay. Free trial, at least one day you have one free delivery.

And, um, we try it the next day. It’s there on top of my, on my doorstep. I’m like, oh my God, it worked, but it’s really interesting. But then I opened the thing and of course the bread got a little bit saggy, like from the outside humidity and stuff. It’s not really supposed to be like this. And I think what they’ve done is they’ve optimized the service, the operation, everything perfectly.

And they have a very unique positioning. I think for example, for older people or people that have some kind of disability or maybe people that just don’t care. This is like the perfect thing. Like a godsend, oh my God, this service, where have you been of my life? But for me as a French person, like any brand that I’m paying for, and that is not like at least a level of whatever crispiness or what I felt like, ah, such a wasted opportunity.

Right? And so I feel like I would rather drive 20 minutes than going to the bakery that I know will have the right thing, uh, rather than getting this humid bread.

[00:50:15] Alex Panagis: Love it. I love the analogy. I love it. I think the, bringing it back to the support discussion, it’s the comparison of using support at every touch point to turn it from a cost center into a profit center by actually delivering or over-delivering on customer expectations and doing everything you possibly can and for as long as you can.

And I think the, the counter that can probably let companies extend this a lot longer is what would you instead be investing in, if not customer support and customer success, which most of the time, at least in B2B, would probably be some form of marketing or some form of sales. And I think that what you’re describing, for example, migrations one-on-one, migrations for Customers is something that is only offered at the enterprise tier of alternative products in a, a super competitive and funded space.

So for a new company, it’s a fantastic way to break into, like, to not only compete with the others, but actually deliver an experience that. For an unfortunate reason, the others just have decided isn’t important to them anymore. Which I wonder if perhaps from improv mx, my logic, and I do see it manifest in, it depends on the product again, but in some companies it’s quite just not feasible.

For example, in like a, like for, let’s take Versal as a example or a any cloud hosting provider that lets people deploy. If somebody has a next JS performance issue, now they go quite above and beyond I would say, in terms of what their team will try to do to investigate it. But it’s still at the certain point they need to draw the line of like, okay, this customer is like on our free tier, can we really spend this much time?

And it really is frankly that they’re not a very good developer. I wonder if there’s like an element of a scale where like you just decide it’s not worth it from cost point of view, which is, again, I, I understand and I agree with your point that you would always choose to do it as opposed to not do it.

But then the secondary point that I think comes with the territory is. If you choose to do it, how do you avoid not being inconsistent across different customers? Because there might be a day where you have three support tickets, so you can give those three the like, like all the attention in the world and going above and beyond.

But then if the next day you had an outage and the bug, and then you also have 30 tickets not related to the outage, those 30 tickets not related to the outage probably are gonna be, you know, resolved just as quickly as possible so that you can focus on dealing with the infrastructure issue at hand.

[00:52:34] Antoine Minoux: Yeah, yeah, that’s right. That’s a very difficult problem to solve. I think we’re always trying to strive for transparency and saying, for example, in the case of an outage, we would say, Hey, we have this thing going on right now. You might be impacted. We’ll get back to you within. Three days or whatever. And then in fentanyl we have this news function where you can just type any human thing, remind me on Friday, and it will just translate that into data and it will reopen that conversation on Friday.

But we’ll always communicate to that customer saying, this is why you’re not getting an answer. The mistake I’ve seen our customers do the most, our customers, as in like clients using our product to serve their customers, is the typical developer funder that just wants to solve the thing directly and be going to respond to the person that raised the bug or the issue saying, ta-da, it’s fixed.

And now you can try. And what happens is that usually people underestimate the time it will take to fix so they don’t answer anything and then they look into the bug, maybe you gonna get distracted and it takes a day and then another, and then it piles up and the customers still haven’t had any answer because the developers thought it would take me two hours and then I come as a savior.

And so I always try to tell our customers like reply as. Soon as possible trying to acknowledge the problem. I don’t think AI should do this job. It should be coming from you. And yes, sure, you can automate the form factor. Like again, use smart replies if you want. Customize the thing, use your own system prompt to make it feel like you, but it has to come from you.

Like you have seen this, you’ve acknowledged this. And also based on our product philosophy, we have this entire thing that your inbox is a little bit like a to-do list and you need to reach inbox zero at the end. You need, yeah, it’s included in the entire product design philosophy. But yeah, that’s one thing I would do to solve that kind of quality gap is just trying to be super honest and communicate the thing.

And most of the times people would be like, I get it. So hey, I’m sorry we are really underwater right now. I just cannot help you diagnose this as much as we would’ve wanted to. Here is a whatever, an article on next jazz’s performance, like troubleshooting. I hope it’s okay. And most of the time it would be like, yeah, cool.

They treated me as a human, as an adult. I get it. I’m trying to accommodate them. You probably also created a bit of a nice loop of creating this relationship again, like the thing I was trying to say earlier. Um, yeah, I think this is just like all about that concept, I think of transparency and setting expectations properly, um, that I would try to do.

But again, I would always try to say, go crazy. Try to replicate the VIP experience as much as possible, as long as you can afford it. Again, like if you’re hammered by a flow of conversation, maybe probably similar to what Sandra has been saying, then you need to take other actions.

[00:55:23] Alex Panagis: But I love, yeah, go ahead.

Yeah, sorry. I was gonna say, I, I, that, I think that’s a perfect answer. I couldn’t have put it better, but basically goes back to the bakery analogy, which is the friendly neighborhood bakery. The bread is not necessarily exactly the same every time because it has imperfections, it’s not mass produced, and maybe they run out of stock on the one that you always get, so you have to buy a different one, for example.

It’s a human experience, and I think the same thing in support with software, ironically, which we try to dehumanize in the, with AI and everything. It is at the end of the day, human experience. So the reality for Nia, it might be like a, as we said, it’s probably a more extreme scenario, but is as long as the customer has the confidence that you are doing what you can and you care about their issue, you’re actually not just writing off the issue because they’re, because you see them as less important because they’re not, you know?

Mm-hmm. A customer with a hundred, 200 seats because they have only two team members or one team member in your product. If it’s seat based, then I think you generally can still with less. Again, not to suggest that therefore put less effort, but if you’re in a situation where you can’t afford for a temporary period of time to put the same amount of effort, I think it lets you still create that experience and basically a.

You’re not abandoning the expectation that the company, you themselves should have to provide this VIP experience for everybody. You’re doing so temporarily so that you can at least provide a baseline experience and then as soon as you can, you get back to providing the VIP experience. I love that. I think that’s, in my opinion, how most should approach it and how if companies were to do it intentionally, they actually could, I think at VC and hyperscalers it, like unfortunately, probably gets into the territory of, you know, looking at ROI and seeing like, oh, this is the revenue that we’re getting from this specific account.

So the support person isn’t allowed to spend more time on, on a, on a specific ticket, which is, yeah, I think that’s when you get into the territory of a business that both you and I find less enjoyable to run. That’s right. The next question, which I think ties into a business that’s enjoyable to run is unlike, I think it’s pretty uncommon, I don’t actually know many people in the SaaS space that basically have what you could call a family business.

I think it meets the criteria, which is that you work with your brother-in-law. So I’m curious to hear a bit more about how that is, how it came to be as well.

[00:57:36] Antoine Minoux: Yeah, yeah, that’s, so it was the other way around. So we started working together, ended up marrying my sister. It became a family business for, for the definition.

But we are started really early, so I was maybe, I think 13 or 14 and he’s, I think four years older than me. So I was probably 17, 18, and I was really wanting to go into, not software at the time, like building website was, it was cool. I don’t even know which year this is in, but it’s a long time ago in the Times of Dream Weaver and HTML three or whatever.

I dunno.

[00:58:11] Alex Panagis: So

[00:58:12] Antoine Minoux: he was fairly technical. It was really into cybersecurity as well, and. I think we bonded a little bit over this and he showed me how to build my first website with Dreamweaver. It was this UI kind of framer like thing where you didn’t have to write code, but it was also producing layouts with HTML tables and stuff.

Like a, a long time ago you didn’t even have RDA corners. You needed p and g stripes with anyway, um, a dinosaur. But like we ended up seeing that, oh my god, maybe we could build something together. And I was going more and more into product design and it was very clear that it would be more like a backend developer.

Yeah, it was the perfect symbiosis. Right, because like at the end of the day, what kind of skills do you need? We were talking about Sandra, right? It’s really hard when you don’t have a technical co-founder. It’s also really hard if you don’t have a good marketing person. ’cause you won’t get anywhere. But I think marketing is a little bit more softer.

So we have terrible marketers, but doing a bit, or sometime even not doing any is a form of marketing. When you don’t have a product designer or a frontend engineer or a backend engineer, then you just can’t build anything. Arguably, maybe not with vibe coding, you could get somewhere with like lable and whatever, but still in its infancy.

So we thought, okay, let’s start building something. We’ve had a lot of failures. We wanted to build like A PHP, my admin, like a SQL interface, replacement open source, no market for this probably was also way too hard for us at the time to just have the pretention to, to replace that. And then we wanted to build kind A CRM.

Um, I was really good. We went pretty far, but we never released it. Mistake number one. ’cause we just got discouraged by the size of the industry and stuff. And I think in Insight, what we had back then was maybe a bit similar to what HubSpot did. So if we would have followed through, maybe we were the next HubSpot.

But what worked for us was a tool called Wall Norbert, which is a way to identify. Emails of of people. So it was really used by recruiters, by marketers, and especially as cold emailing started to be a more adopted or a more famous method to grow your B2B SaaS. And yeah, this was really our first success.

We ended up selling that through F International to a company called Ram Ventures Popoff and to Jean Patel, which I think was for them a nice addition to another product they had called Mailshake. And with the money we got from the exit, we ended up building our own broker, this thing called transfer slot so you could buy and sell smaller side projects.

Um, I think it’s very similar to a company called acquire.com right now. Um, we realized running trans lot that we ended up being brokers being in calls all day. And this was not really fun for us because what is our mode is building challenging technical products. And so we ended up selling transfer to the broker at the international.

I think they killed it a few months later. Um, but again, a nice little exit. Use the menu for something else. Both improve MX grew it to, um, many hundreds of thousands in a RR and end up selling it for money that we are just injecting also into for Loop. As you see, we have already three exits under the belt.

I think this is because we are really mostly creators. We like to create and build, which is funny because the name of this thing is True Operators. Right? But I don’t really consider myself a really good operator per se. Like with improvements. We had to hire and we had to create SOPs and stuff like this, which is like fun for a moment.

Yeah, really love working with my brother-in-law. I think it’s also studying the stage into a lot of. Transparency and there is just no option where we’ll fight. Like we cannot fight. There’s just no way. Right? ’cause this is bigger than us and I think our personalities work really well together. So this is, this is also a mode that we have.

’cause I’ve talked with other founders and they are almost always fighting with our co-founder about something that’s never happening to us.

[01:02:27] Alex Panagis: Yeah, I think especially when you have technical and non-technical, I find that the co-founder who is technical and perhaps completely rightfully so, as we said, it’s foundational to being able to run a software company has this not superiority complex ’cause that suggests that I’m saying they shouldn’t, but they have this view that they’re the one doing like more of the work and then the marketing person is like more soft skills.

Like we even naturally when talking about it, we would say the same things. And I think it is kind of the case that therefore those types of co-founders who are paired, you know, non-technical with a technical. I think generally ends up not working unless the personality like really, really clicks. I’m curious, I dunno if it’s public, so we can cut this bit, but you, for part of your journey, at least I, I don’t know if it’s still the case, you are also maintain your like principal designer role at Yelp.

Is that still the case? Yeah.

[01:03:17] Antoine Minoux: Yeah. That’s still public. I’ve tweet about this. I’ve also appears that I wanna write on LinkedIn. I had a public talk, maybe you can link to this then I done in Berlin about, about how you can still work a full-time job or fellow part-time and why to do it. Because I get this question most of the time and I think it’s a kind of about like getting the best out of both worlds.

So I get the upside of running my own businesses and doing the things that I want. And for me it’s a little bit like a creative outlet as well. And for the operational things that I don’t like to do that require 24 7 support, we might just hire people to take that. Um, but then my job at OP as I.

Part-time principal designer is much more strategic. I get to juggle with like millions of users on my ads to dealership. I just came back from a shred where I’ve met the entire design team in, in Montreal, in Canada last week. So it’s just like opening a world that you would just not have running your business that’s very lonely, at least as a bootstrap business.

Um, and also quite frankly, it gives me the financial stability, the benefits and whatever to be able to run my startup in a different way that I would first potentially not be able to. So especially the calm aspect, right? It’s also calm because I cannot afford it to be, if I had three months of runway, it would be really hard for me to say, yeah, it’s calm.

We’ll just read as a future in six months. So yeah, I think this works for me. I. Don’t want to say this is the new working model for anyone, but I would say like a lot of product designers, PMs or software engineers working at like the final companies, they could benefit from trying to launch a side project, running a business on their own, even if it’s super small.

But trying to understand like other pieces, like what is marketing really doing? Or how can I design this in a better way? And then they would be much better at their job. They would also have much more ammunition to go and negotiate compensation and stuff. So like, look how engaged I am in my role and stuff.

And I think you’d just be a better designer, be at a pm a better engineer if you’re trying to expand your world a little bit. And I think to finish on this point, vice versa, right? Like a lot of things that I’ve seen at Yelp has taught me, um, about my job and how to be a better funder. Like a lot about the business side, like the kind of tooling that they all say on top of compliance.

Goal and like also the way they approach customer support. Like I get to speak to so many smart people that I would never be able to access otherwise. My manager was before is not working yet. My ex manager is not working at Google. The previous one before that was at LinkedIn, etc. So you get access to like of the network of people that you can tap into yours, very quote.

Again, I’m not sure if it’s a model worth for everyone. And sometime I get distracting from customers like, Hey, it puts, you’re not working full time on this thing. It’s like a really big red flag. But then I explain and they understand again, like it’s like the nice popup bakery down the street. Good big, still delivering the best brand even though the funder has like a, a it somewhere and has a, uh, one-man show every Thursday may and whatever.

So, but it’s more of a personal thing.

[01:06:40] Alex Panagis: Yeah. I also think it probably makes you build it much more intentionally because you’re, you know, you can’t be the on call engineer twenty four seven, whereas if you have a technical founder who’s solo and goes all in. He doesn’t, definitely wouldn’t have initially the money to say, just hire somebody to cover the hours where he’s asleep.

So then it’s basically never going to be 24 7 coverage initially. So I think that’s, yeah, in the bootstrapped world at least, like the, the, the fact that there are companies like I was at, I honestly, I taken back by the fact that, uh, big tech company, which I believe is what Yelp would fall into. It’s like a bigger late stage company, the internet company, as opposed to like the new, there are newer companies like Versal, which I, from outside looking in really, really does support people starting and running products, especially because they’re also dog fooding, versuse, and therefore getting to know the product more, how to scale like a service as opposed to when you’re already in the versal ecosystem using Versal, obviously it’s like really predefined how things are done.

Whereas if you’re building a product from scratch, you see like, okay, well now I got like a bill that’s way higher than what it was supposed to be. I think it’s really smart. That companies do that because it probably also makes you far more excited to work at like vice versa. Like the work at like Fernand makes you more excited to go in and have that change at Yelp.

And then the work at Yelp, like when you come into Fernand is like a really nice change of pace as well.

[01:08:01] Antoine Minoux: Yelp just released an a product called Yelp Perceptionist, which allows small local businesses to set up an AI receptionist on their phone and data will respond to them. And I work on this as a product designer, especially on the onboarding, purchasing experience.

And a lot of it I was able to do faster or more, let’s say educated because I worked on that feature, not voice, but the AI agent for And at which company would you have someone that has done something so recent? Very unlikely. You even take the people at Google, they would need to figure it out and they have a lot of smart people, engineer and designers, and they would figure it out.

It’s fine, but I still think there’s a bit of an edge, right, like that you. Use, but the virtual example, I think it’s right. ’cause they’re, especially, I think hiring based on this, I think they see your kind of small entrepreneurship endeavors as a green flag signal to make you an offer and try to get you to work at versa.

I think though I’m able to do what I’m doing because the work that I’m doing is much more strategical and is less of always our 9:00 AM to 5:00 PM kind of job. And so sometime I’ll have weeks there where I’m working 60 hours and sometime I have weeks where I’ll spend maybe more like 30. And at the end, no one cares.

My manager wants me to compete the job that I signed up for and then produce the value that I have to, and so that’s why I’m recording this podcast during my lunch break right now. It’s totally fine. No one cares as long as at the end of the day I’ve done what I needed to do. I think if I would work at a company like Versal, which is more still in expansion, more like hypergrowth.

I would not have the luxury to be able to take so much more freedom from the way I run my calendar. Right? That’s my expectation. But I think I would wanna be more all in and the fact that Yelp is a publicly traded company that needs more like clear thinking. You could do in one hour what a junior designer in my case would do in one week in terms of like just looking at the value you produce, right?

The junior would do a crazy big thing and he would spend like hundred hours, but in one hour you could have done the same. I think this is what I can bring and somehow how this manages

[01:10:11] Alex Panagis: to

[01:10:12] Antoine Minoux: be working.

[01:10:13] Alex Panagis: Yeah, that makes a lot of sense. Well thank you for coming on. It was really great conversation. Perhaps the future one, we could probably talk about the stuff that you’re doing at Yelp.

’cause it does sound super interesting if you’re open to coming on again at some point. Sure. Nice. Thank you.