Aydin Mirzaee, Fellow CEO on Pivoting Meetings to AI

Michael Koenig: Hey, it's Michael. If you've been on a

Aydin Mirzaee: Ooh, excited to be back. I can't believe it was only a year ago, but you know, heck, even if it was like 3 weeks ago, so much has happened in the last 3 weeks even that there's just so much to talk about.

Michael Koenig: I think what we're seeing with Fellow in particular and all of the progress you all are making is like years of development happening in months. And let's start there. So much has changed. It's been some time since your, your first appearance. What's the biggest shift over the past year that you've seen at Fellow?

Aydin Mirzaee: Yeah, I mean, a lot has happened, but one of the things which is really interesting, Michael, is just our transformation. So Fellow has, we launched a product in 2019. We started as this meeting productivity platform, and over the last, I would say, 18 months, we pivoted the company, and today we're, I think, the best AI meeting assistant out there. A little biased, but we can talk about why I believe that. And the biggest change has been just the transformation of the company into this new AI-first company. And I think this is a really important, uh, you know, point to talk about. Part, part of the reason is that we transformed and we had to, because I fundamentally believe that had we not completely transformed the company and the product, that we would've gotten extinct as a, as a company. AI impacts every company, but I think that impact comes in waves, but I think it's coming for every company. So meetings and AI, like the use case is really obvious, right? So it's kind of, you know, crazy not to have an AI meeting assistant in a meeting today. And we could talk about all the reasons why, but that's kind of like one of the first waves of what AI impacts. But I think AI is going to impact every single company. So I do think that the transformation that we went through is something that every single company also needs to go through. And it's a very difficult thing to do, especially when you have a business, you already have customers, you have processes, you have employees, org structures, metrics, everything around like the old way that you used to operate. And just understanding that like, you know, all of that needs to needs to change. You know, it's interesting, you know, we talked about this last time, but my last company, you know, we sold to SurveyMonkey and we worked at SurveyMonkey for a few years. But even if you think about just, you know, something as basic as surveys, like even this is going to get disrupted by AI, right? Like there's this new YC company, I think it's called Outset, and it's just basically building these like AI interviewers. At scale and they're working with companies like Weight Watchers and so on and so forth. And literally like there's humans talking to AIs, like voice talking to AIs and having, you know, long conversations that go very deep to ask about like, why do you wanna lose weight and what would it mean to you? And just going like many levels deep, but being able to do that at scale. So my point is that it's going to impact every single company, you know, as time progresses. I would say that we were almost like lucky that, you know, AI and meetings make, you know, jive very well together. So we could be at the forefront of the transformation and kind of get ahead of it. But yeah, I mean, that has been a major learning for us is like, how do you transform your company that's been around for a few years, has thousands of customers, you know, that love the product in its traditional sense and just, you know, pivoted and say like, hey, we're no longer that company, we're this new company. And all the things that it kind of entails.

Michael Koenig: Well, there's a distinction that I want to draw here. That word pivot in tech is generally thrown around when something's not working, right? A startup has a new idea, they test it, there's no need for it in the market, no demand, so they pivot. You guys had a product and, and still do have that product that has incredible demand, has product-market fit, but you as the CEO and probably the rest of your team contributed to having this concept that, okay, this is where the puck is moving to and let's just do this. How do you go all in on that? That's huge.

Aydin Mirzaee: It's a difficult thing to do, and I'll kind of explain why it's difficult because you have to think about all the things that need to change. One is the easy one is the product needs to change. And I say that's the easy one cuz it, it's pretty obvious that that needs to change. But once a product changes, then your onboarding flows need to change. Then your pricing and packaging needs to change. Then the way that you service your customers need to change. Your sales team structure needs to change. The way that you describe yourself and your website needs to change. The way that your customer success teams work need to change. The way that you organize your teams need to change. And you need to do all of this. And then the hard part of how do you go back to your original customers and say, hey, I know that you bought this product because You like these sorts of things and, but you know, you should, you should change to using this new thing that we're doing and it's better, but we have to convince you about that. And it's priced differently, so it'll actually even cost you more than what we used to charge because now it's, it's powered by AI and it does all this excellent stuff for you. But you know, these are additional things, so you kind of have to buy into this new vision of the product. And so it's just a combination of all those things. And I think, you know, part of it was that we had this pivotal moment where we brought everybody together and we kind of talked about, this is where the puck is going, like you said. And if we don't change, then we will go extinct because this is where kind of the market is heading. It's kind of like being in a world of, you know, the phone book when the internet is going to come out, right? Like there's like massive shifts that are going to happen. And so once we kind of put that there, we had a hackathon and we started by getting everybody in our development teams focusing on building with AI and building like the, you know, what are the things that we could do with AI in the product? And so we started there and, you know, we had a beta, but it was, it was kind of like that moment where we said, no, this is the new future of the company. And we started using very strong language internally. Things like, "For the new fellow to be born, the old fellow must die." And using this very strong language, but it was really just to really jolt people from this idea of it's not about the way that we used to do things or the use cases that we used to support or all of these things. It's more about what are the things that AI makes possible? And it's a new set of building blocks. And so now that this exists, We're still gonna solve the same problems, but we have brand new building blocks. So it's not to use these building blocks to solve these problems is, is, is almost crazy. And if we don't do it, the next company that comes around that just started will. And so there is a little bit of like disrupting yourself that needs to happen and everybody kind of needs to get on board with that.

Michael Koenig: Disrupting yourself is an understatement, but what you guys did, I mean, it's nothing short of remarkable and there's so much that we need to tease out of what this transformation looked like. Like. Before we do though, you had a product that had product market fit, paying customers, renewing customers, great figures, everything going on. How did you test this new AI meeting assistant? How did you go from, hey, wouldn't it be cool if it did this, to, hey, uh, actually there's a need in the market and our customers actually want this?

Aydin Mirzaee: You know, we did talk to a bunch of customers, and a bunch of customers were starting to talk about basically how it would make sense for Fellow to have a meeting assistant that could do some things like take notes automatically, track actions and decisions and everything else. And so we definitely did get some market pull on those things, but the challenge with only talking to your current customers is that they're waiting and the information that you get from them is going to be weighted towards what they know about you already. So it kind of starts from this base level. So we definitely did talk to customers and so our forward-thinking customers were thinking about, you know, wouldn't it be great if, you know, today I have to manually prepare for my meetings, but if Fellow used AI and knew everything that we talked about before, it would be in a better place to suggest what we should talk about during the next meeting than, you know, than I could. Or wouldn't it be great if Fellow had all of the conversations that, you know, that, that happened and at the end of the year, you know, after a one-on-one meeting, I can say, oh, like what feedback should I give this direct report of mine? I mean, you have one year full of one-on-one meeting information here. Like Fellow can probably come up with this much better. Or, you know, things like, oh, we had all these customer conversations. What are the top 3 feature requests that our customers have been talking about? And so, We started to hear things like that. You know, we built MVPs, we had our customers try our MVPs, but the other thing that we did was that we made it so that every new customer that came into the product was getting the new Fellow. And so this was another part of it. And then slowly, like we made that the priority of like every new person that comes in is gonna get the new Fellow that has the AI meeting assistant that, acts like a chief of staff for your company. And then for our legacy customers, we had to do a lot of manual work. Some of them came on board themselves, others we had to use our customer success teams to engage with them and show them, hey, like this is how you're using the product today, but this is how you could do things. Why would you ever write something down yourself? Why would you ever try and hold someone else accountable? Track their action items when Fellow can do all that stuff for you. But it required more work. So oddly enough, when someone starts using a product, it's actually difficult to get their attention and to say like, hey, you should use this other, like you should use this thing too. Cuz people get tunnel vision and they just start using a product a certain way and they don't really wanna change from that. I think there's this like famous anecdote where, you know, for, I think it was Microsoft Word where they ask customers like, what are the top features that you want? And. Customers came up with this whole list and I don't know, like 90% of them already existed in the product, but nobody knew about them. So there is this thing that happens for products that have been around for a long time. So it is actually like, it does require effort to get over the inertia of like people already using a product in a certain way.

Michael Koenig: Well, so now we're talking about reeducating customers, existing customers, and teaching new customers. So let's set that aside. One of the things that I'm really quite interested in here is let's get back to that stark language that you used. In order for the new product to live, the old product has to die. How did you approach change management? Because that can be quite daunting to a team, whether you're in product and like, oh my gosh, I've built this for so long and now we're just killing it, or customer service who like really enjoy— whatever it may be, how how did you get to that change management to do this so successfully? How did you get to excite the team and get them to embrace innovation and embrace this next wave?

Aydin Mirzaee: Yeah, so, so these are good questions and, and obviously like there's a bunch of tactical stuff, but I think a lot of it just comes down to culture. So we had to change cultural elements within the company too. So I'll, I'll give you one kind of tactical example. We had this cultural value which was called, um, pace quickly. Speed matters in companies and startups, urgency matters. And so we've always had this, but we started to change some of the values in the company. So we archived Pace Quickly and we replaced it with Fast and Fearless. And the reason for this was that it wasn't about pacing anymore and quickly wasn't fast enough because the world of AI is changing so quickly. That we had to up the pace. But the fearless part is really, really important because a lot of the slowness came from being fearful of making these kinds of changes. Oh, what if we do this thing and it pisses off a whole bunch of our legacy customers? Or what if we do this thing and it doesn't work? So there was all these fear elements of making changes. And so every time we made a change that was, oh my God, we just did that thing and we, you know, we deleted that feature and we added this thing, or we went in and we said, oh, guess what? Everybody, you know, on our paid plans is going to get unlimited AI, which is like a big thing to do because you don't know how much people are gonna use and like what that's gonna cost. And, you know, there's this tendency of, well, let's just approach it slowly. Let's like slowly, maybe put it on our highest level plan first and see what happens. But once you kind of add this fearlessness nature of it, it allows you to move faster to be able to do these things. So a lot of it was just cultural elements and kind of emphasizing that people had the opportunity to make bold decisions and not be fearful that it might have consequences. Because again, in most cases, most decisions are reversible. Most of the things can be backtracked. But once you've had an established business, this starts to creep in, right? Which is you start to be very methodical. Every decision has multiple stakeholders and you try and do things incrementally. But when there's drastic shifts in the world happening and technology, you can't be incremental anymore. You kind of need to make big bets and be able to move fast. And so those are, yeah, culturally very difficult. But this was, for example, one of the things that we did and we kept emphasizing it at every town hall. Every time we, we did a decision which was kind of bold in that way, you know, we shouted it out by, you know, this is the Fast and Fearless value in play. And so yeah, this was kind of one of, uh, one of the elements there.

Michael Koenig: And it underscores just how vital internal communications are. And it's something that's often overlooked, especially in maybe companies that haven't reached the operational maturity that Fellow has, for instance. What is your communication cadence? How do you all go about it?

Aydin Mirzaee: Yeah, so we have a bunch of things in place, so, and I think all of them kind of go hand in hand. We, for example, have weekly town halls and the town halls are about, you know, part communicating what's going on, wins across the company, new product demos. And again, there's huge velocity in the teams as well. So we're constantly shipping product. And so there's this element of, you know, see what's going live this week. Okay, this is going live. You know, marketing needs to update all of the right things. Our help center articles need to be updated. Our comparison pages need to be updated. The sales team needs to know how to speak the new language and be able to talk about this stuff. So town halls start to become really important because we're just moving so incredibly fast that we have to keep everybody on the same page. So this is really like a highlight of the week. The other thing that we do is every week I get to send a company email to everybody, which just emphasizes the things that are on my mind, how I'm thinking about things, important anecdotes. We have monthly business reviews, which are really fun to run. And it's really cool because you get people from all across the company that come in and talk about their OKRs, where they're struggling, where they're succeeding, and then giving the executive team to be able to communicate and give coaching and advice on how to tackle those things. And it's not just the executive team, right? It's people throughout the company that participate in these business reviews. We also have our quarterly business reviews, which are more about the whole quarter. And obviously we have our in-person meeting. We're a fully remote company, but we also do twice a year, everybody in-person gatherings where we talk about the vision and like where we're going and get everybody on the same page. So you're right, communication is always important, but particularly when you, when you're moving incredibly fast and you want to give people the ability to just execute at the ground level without having to get approvals from 100 other people and that kind of thing. But if you're going to move at that pace, then you really need strong communication structure so that All those things can be communicated with everyone.

Michael Koenig: And you talked about the vision. Last time we talked, it wasn't on this podcast, but we talked about how you were going for the chief of staff for all of your meetings, and it kind of feels like you're there. So what next? And the reason I ask this is it seemed like a really big audacious goal and it seemed years away, I feel like. And now if you're not there yet, you will be very, very soon. How do you think about that as a leader? How do you change kind of your goals and your expectations and forward thinking?

Aydin Mirzaee: Yeah, so it, it, it's really interesting. I, I, I see this as, uh, you know, 3 phases for the company. The first phase was we focused on meeting productivity. The second phase was we morphed into an AI meeting assistant, which is, uh, the way we refer to ourselves today. And where I think most software is going to is this just agentic phase. Right? So we have been traditionally taught that we are users of software and like we're doing the work and software's helping us do the work, but we're going to enter this phase where software just goes and does things by itself. Right? And so this is the agentic phase where workflows just happen automatically and there doesn't need to be human interfaces all along the way. You can just have these agents that will go off and do things for you. And I also think that we're heading into this world where you will have, you know, human workers and AI workers working side by side, and each one will do different functions. And so I think that the future that we're building towards is Fellow will be like a digital employee that you hire. You'll hire a Fellow, it'll come into your company, it will attend all the meetings in your company, and do all the basics like take notes for you, track actions, hold everybody accountable, remind you of things, help you prepare for things, but it'll also update all the systems of record within your company. So, you know, it'll update your CRM. If you talk about something or you teach someone something, it'll update your wiki automatically. You don't have to tell it to do that. Or, you know, if you have a task management or project management system, you talk about the status of something, it'll just Go automatically update that. And that's a powerful thing to have. But I also think that the next stage of this is, we believe that the biggest source of data within any company is the conversations that people have within that company. And so the next part is, can you use Fellow to deliver insights to you? Can you say, what are the bottlenecks in my company right now? And get a really great answer that would actually far exceed an answer that any one person could give because AI can consume much more information than any one person can and be able to answer questions like, what are the top objections that my customers are giving my sales team? And so these sorts of things. So I think that stage is the most valuable stage that a chief of staff can bring. Is can it bring you these insights? And we definitely have a lot of elements of this today, but there's still a lot more to build. But this is what we're building towards. This isn't about recording your meetings. This is more about hiring a chief of staff for your company. And that's really what Fellow is all about.

Michael Koenig: That's amazing. When we first talked and you gave us a tour of Fellow, one of the things that you said is it's essentially the operating system from your company, and that's sort of the evolution. You built a previous startup, you exited it, you went into this company, and you found all of these huge operational elements that were just in place, but you didn't even realize. And now Fellow became that next instantiation of those operations for any company. And so now you're talking about the DNA of Fellow is still tied in with the vision you're talking about right now. Of being that operational force within the company. I love that. Now let's talk about in terms of how customers are using the AI-powered meeting tools. What has surprised you? I'm assuming you all are paying attention to analytics, seeing different things, getting anecdotal stuff from customers.

Aydin Mirzaee: It's very interesting. Like some of the really cool stuff that I've seen is, I'll give you some like examples. Uh, so there's this one recruiting company. They, they're a recruiting agency, they have a bunch of recruiters and it's interesting. So obviously they invite Fellow to all of their meetings and, and Fellow's there, you know, you know, for all the meetings. But then after like they've chosen a handful of candidates that they want to then refer back to their customer, they will ask Fellow that, hey, based on this meeting and based on, you know, this is the, the job description and they'll kind of paste that in. Please write an email to our customer explaining why this candidate is a good candidate for this job. Right. So, and then like Fellow will just, you know, take all that context and like spit out the email and then they just send that to the�� so they've made the decision that like, this is a, this is the right candidate, but then the work of selling their customer and giving them the explanation why, that's all produced by Fellow, which is just like really cool. Another one is just like internal, uh, internal communication stuff. So we, you know, customers will, uh, say have meetings with, uh, with their customers and say, say it's like an agency and they're building things. So they might do things like have the meeting and then based on that, they need to create a requirements doc. And like maybe before they would've created that requirements doc themselves, but they just had a meeting where they talked about all the requirements. So they literally ask Fellow. Hey, based on this meeting, here's a template for, you know, how we usually create requirements docs. Based on this, can you just like make the requirements doc? And then it just does that. They send it to the customer, they verify, and then, you know, they send it to the engineering team. Or same thing for, you know, product research, right? So if you want to create clips or, you know, communicate things to your product team after an interview, you know, you can use the, the, the same sorts of workflows. The other thing that's kind of surprised me, I mean, I'm not necessarily surprised, but it's really validating to see these sorts of things. It used to be that when this idea of having an AI meeting assistant in your company, it really started with like the very basics of, oh, let's just record conversations, right? Like that's really where all this began. So then if you're recording things, you know, the immediate place that you start to think about is, well, I'm recording it just in case, the one person that missed it wants to watch this, then they can do that, right? And that's kind of where this stuff started. So you would never in that case record a one-on-one meeting, right? Because both parties are there, nobody's missing the meeting, so why would you ever have an AI meeting assistant there? But now what we're realizing is that it's not about that. Like that use case of, oh, if you miss the meeting, you can get the information is so like level zero. It's, it's not about the recording, it's about, you know, the organizational knowledge that is communicated that can be referenced. And so every time that you have an AI meeting assistant capturing that organizational knowledge, you're actually making your company better because you're adding to the information that is retained within the intelligence of your organization that can be looked back at, that can be manipulated, that, you know, You can create documents out of it. You can ask it questions. You can, you can do so many different things. And again, not having that is just like a huge opportunity lost. And so this is also the, the mind shift that we are also seeing there.

Michael Koenig: We'll be right back. Hey, it's Michael. If you've been on a

Aydin Mirzaee: Yeah, so great question. So I don't see the Google,

Michael Koenig: Well, let's talk about privacy and security. I know this is a huge focus for you all. Fellow is known to have that as priority number one in the market.

Aydin Mirzaee: How are you approaching it?

Michael Koenig: And how do you put people at ease?

Aydin Mirzaee: Yeah, so this is a great question, and I think that for internal meetings, this, this matters a lot. So one of the things that we're, we're hearing a lot about is, you know, I go into a meeting in my company, I, you know, I work at a 1,000-person company and there's, you know, 3 different meeting bots that joined and I don't know who invited them. And it turns out that one of them belongs to an employee that no longer even works here. So it's starting to get to be a problem. So amongst the things that we do is we've always had this organizational mindset. So we've always built for the enterprise in mind. So for example, customer number 1 for Fellow was Shopify. So when your first customer is Shopify, you build for the enterprise from the get-go. And so amongst the things that, for example, we put into place is various rules that your IT team can use to make sure that the right types of things are on the record and, and the things that you don't want on the record are not on the record. So for example, if my lawyer is in the meeting, then I would like there to be an organizational rule that prevents that meeting from being on the record. Like these are things that you can do with Fellow. You can have these, these rules. You can say, well, if it's a sales meeting with a customer, then I want that to be internally accessible to the whole company. But my implementation meetings, I only want those to be accessible to the entire customer success team. Or for my board meetings, I don't want there to be video or audio. I just want the board meeting minutes and I just want that to be produced and I want nothing else to be there. But for these other types of meetings, I want the everything to exist, the, the video, the audio, but I want it deleted after 1 month, right? So, These are the sorts of things that you can do with Fellow. Uh, the, the other thing I would say is that, you know, this concept of, uh, going on the record and off the record. So for example, with Fellow, you can do that. So if, if we're talking about something that we don't want on the record, we can put Fellow on pause during the meeting. Uh, but we also have redaction. So redaction is really powerful. So after the meeting, if there's something that you want to remove, it's very, very easy to do that. You go in, you select it from the transcript, delete it. And when you delete it, it doesn't just go from the transcript, it's gone from the video, it's gone from the AI notes and summary. So it's really enterprise-grade redaction. And so these are the sorts of things that we build towards because we have this focus on internal meetings. And of course, you know, people use Fellow for all their meetings, internal and external, but Internal meetings have all these extra requirements that I think most people don't pay attention to that we care a lot about.

Michael Koenig: Let's talk about the models. Obviously you guys aren't building your own. This is built on top of an LLM. Security, privacy, these things are always a concern, especially now, you know, I don't want to date this episode, but we saw just this very interesting innovation come out of, out of Deepseek, a Chinese company. Where they were able to highly optimize their hardware to be able to do a really inexpensive final training run, which has dropped the cost of using their product significantly lower than ChatGPT. How do you guys approach this? Which models are you using? How do you think about this? Because this not only has to do with security and privacy, but it also has to do with the bottom line and your unit economics.

Aydin Mirzaee: What, what's interesting is it's, it's really cool to see this all play out, right? All of the. Very large and very well-funded companies out there, all competing to build the best models out there. And I think this ultimately really benefits everybody, right? Because one of the things that you can assume is that the cost of using this intelligence will continuously lower over the course of time. And that's really important because there's a lot of things when you're building a product that you think about. So you think about, well, I could deliver that functionality, but it's going to require all of these API calls, which will cost a lot. So then the question is, is this worth it enough for the customer? And like, will they pay for it? And all those things slow down decision-making. But if you kind of assume that, oh, costs will actually come down over the course of time, it will allow you to build towards those use cases and, you know, build a better product and not worry as much about what things cost today, knowing that they will reduce over the course of time. So seeing something like Deepseek is very encouraging because it kind of accelerated this idea that, yeah, cost will come down and it might come down faster than we all think. And that's very encouraging. And I do think that as costs come down, all it will mean is that we will use more AI throughout our products. I don't think it's like all of a sudden your margins will go up so much because the costs are lower. I just think that it would be an opportunity to use a lot more of it. At the same sort of margin profiles that you always had, but just like deliver more value to the customers.

Michael Koenig: Great answer. And yes, absolutely. As these models continue to get commoditized, hopefully prices will continue to drop. One of the things that I love, you know, we'd look at so many different AI companies that are out there and you see just on a weekly basis, oh, company X, Y, and

Aydin Mirzaee: This is a really good question. I mean, so many things are, are changing in, even in terms of the way that We develop software. So for example, today at Fellow, all the teams are using Cursor. And so one of the fastest companies ever to get to $100 million in ARR. And so this kind of changes the way that even your software developers build products. It also changes the way that you get, you do your customer interviews and what's possible. So you might do all your customer interviews with AI. You might, you know, get all of the product management function to be like really powered by AI because it can, you know, decipher thousands of conversations and then come up with the best things that you should build or the most logical things that you should build next. On the sales team, again, this, you know, more and more people are starting to use agentic tools. As basically AI BDRs or AI SDRs so that they're not hiring for that role anymore. Even from a marketing perspective, there's this whole new world of go-to-market engineering that's starting to form. So really when I look at every part of the company, it seems that there's a new way to do all of those things. And I think companies that are starting from scratch, actually potentially have a leg up on everybody else because every other company's already operating in a certain way and they have to change to adopt the new way. But if you were able to start from scratch, you would say like, what is the cutting edge of this stuff? How does product management work now that AI exists? How does sales work that now that AI exists? Even how does HR work now that AI exists? And there's AI recruiters now. So any function that you look at, there are new startups, for example, that exist that are trying to build agentic versions of software that used to exist in those roles. So I think that's the first thing from a— if I were to start from scratch today, that's the way that I would build. And knowing that, my job is to make sure that we still do all those things. And we're still disrupting ourselves, not just at a company level, but also in the individual department level and the individual way that we look at things. You think about marketing, one of the most traditional things in the world of marketing was you would get your traffic through search engines as an example. But even the concept of a search engine is changing, right? Where people get their answers is changing. Less and less people are using traditional search and using AI chat interfaces. So it really feels like just the way to build companies is now different. And so when starting from scratch, I think you have the opportunity to adopt these new workflows from the get-go. So that's pretty exciting. But to some extent, a lot of the things don't change, right? You still need operational rigor, you still need a strong culture, you still need great communication. Like the fundamentals of building businesses don't change, but I do think that the tools allow you to pass on a lot of the busy work and really focus on the high-value strategy things more than ever before.

Michael Koenig: Yeah. So interesting. And you actually posted— I unearthed the post from LinkedIn a ways back, and you had written about how old companies are not going to be embracing AI and adopting it as quickly as new companies. And you quote, you know, I've realized that adopting AI isn't about doing the same thing the same way. And it means killing old habits and embracing new ones. I like the mindset of what would I do differently if I started this company today? And perhaps that's a way that older companies can actually go about trying to figure out what they would do differently.

Aydin Mirzaee: Yeah, it's so true because again, like, I think the misconception here is that people assume that I'm gonna use an AI tool and it's just gonna speed up my current process. Whereas using AI means coming up with completely different workflows. And this is a critical difference that I, I wanna say why it's very hard as a currently existing company to adopt AI and, you know, in, in the right way is because you have a deadline, right? Say, you know, by Friday I have to deliver this report. You know that if you do things the old way, that you're gonna be able to hit that deadline. But if you're gonna reinvent your workflow, then the reinventing the workflow means, okay, I'm gonna start by spinning up ChatGPT deep research. Then what I'm gonna do is I'm gonna talk to on voice mode and brainstorm for 2 hours, and then I'm gonna take that, you know, create a voice memo. You know, write something, get, you know, another AI agent to find the critical feedback there and then iterate. And it's just, you're doing things in different ways and getting that right is going to take some time. So the first time you do it, what you'll realize is, oh my God, this actually took longer and I'm not even sure that the output is better than what it would've been if I just did it myself. This applies to software, it applies to marketing work, everything else. So the first time you do it, Everybody's gonna say, well, it's quite not there because, you know, it would've been faster if I did it myself and the quality would've been higher. But the thing is that you, you, you're going to suck at it when you first start. And all these things require new workflows and iterations and learning how to change the way that you work. And once you do that for a little while, then you will get the efficiencies and then you'll be able to do things in half the time with twice the quality., but it's gonna take some time and effort. And the problem is we're all so busy doing that is very hard to carve out the time to reinvent your workflows and figure out how to use and, and work in the new way. So this is what makes it challenging. And Michael, what I will say is that some of the smartest people I know, some of the smartest leaders I know, are spending anywhere from half an hour to 2 hours a day just trying to keep up with AI, even if it's not like their day-to-day function. Like, this is how impactful it is. I, I just think we all need to carve out a whole bunch of time in our work days and work weeks, whenever it has to be, just so that we can, we can keep pace with, with everything that's going on.

Michael Koenig: Yeah, absolutely. And one of the things that I had asked you when you were first around was, hey, how are you encouraging your team to experiment with AI? And you had set aside a budget, I believe it was $500 or something like that, for each team member to go and experiment and then present what they learned. I think part of the success, and let me know if you agree or disagree, but perhaps part of the success with just how quickly you all have innovated not only your product but your internal operations has to have come from the tone you set from the top going back then. Do you think that is part of what has enabled you guys to really take that forward-looking approach throughout every fiber of this company?

Aydin Mirzaee: Yeah, so, so for sure, for sure about, about the leadership. I mean, again, we talked about how During town halls, we have this AI show and tell section, and every week there's someone who talks about something that they've done using AI workflows. And yeah, we encourage this, the leaders do it, but we also want to get people from all levels of the company to be able to do these kinds of presentations because when someone does that and everybody kind of acknowledges it and encourages it, that's when cultural change really starts to happen.

Michael Koenig: What's one piece of advice that you would give to COOs or other operations folks within, you know, not new companies, some of the older companies, on how to implement AI? Where should they start or how should they think about it?

Aydin Mirzaee: Yeah, so I, I think that again, every, every company's gonna be different, but there are now a slew of really awesome consultants out there. That are just like, their focus is, you know, they're AI consultants and, you know, they're like keeping up to date with everything that's going on. And sometimes again, depending on where your company's starting from, if, if you're at a place where, oh my God, we're doing absolutely nothing and we don't have the expertise in-house, it might make sense to, you know, hire someone like that to come in and say like, here are the basics. Like you guys need to be doing this and this and that. And then you can kind of start, but at the same time, try to change the culture too and encourage use, create experimental budgets and do those sorts of things as well. But if you feel like you're starting and you have a lot to go, you know, a lot, you're really behind, you know, start with a coach and a consultant to bring them in and they can really help accelerate things. That's something that I think makes sense, especially if you're a larger company. Part of the challenge is like there's so much cultural inertia in changing things. So getting an external party might be helpful. It's also cool, and I've seen a lot of companies designate people and teams within companies, like they've created like AI councils, whose job it is to look at new technologies, but with a goal and a focus. We're going to see how we're going to make our customer support better. Basically using AI. How can we do that? And they go off and they explore. And once they tackle one part of the company, then they can go and tackle another part of the company. So, that's a few things that I've seen, you know, some companies doing with good success.

Michael Koenig: Love it. And we talk about AI, not only new technology disrupting old jobs, but also creating new ones. Well, AI workflow consultant, that's certainly a job title that wasn't around most recently. So Aiden, this has been awesome. Thanks so much for coming back on. Where can people go to learn more about Fellow?

Aydin Mirzaee: Yeah, you could just visit us at fellow.ai.

Michael Koenig: Oh, you got the .ai now.

Aydin Mirzaee: We, we have the .ai.