Scaling and selling AI products for enterprise | May Habib (Co-founder and CEO of Writer)

Todd: Hey, so to start, May, I thought it'd be interesting if you could go into your background before starting Writer. What did your personal path to becoming a founder look like?

May: I grew up in a very entrepreneurial family, so it was just never in my conception of life that I would not be starting and running companies. people in my family didn't have technology startups, uh, they had, you know, small mom-and-pop shops, you know, it was just very clear that was in my DNA, really, and my makeup as a person, In high school I did a bunch of things, had a bookkeeping business, you know, I really started getting passionate about language technologies, it felt like. You know, it was an idea that could really be kind of the target of my obsessions for a long time. that was kind of the, the bar I had set for myself of, what it would need to look like and feel like to kind of leave the corporate life. 

Todd: So what was your first real startup and why did you start it?

May: So it was Qordoba it was a machine translation company. Uh, Waseem and I started it and moved to San Francisco, end of 14, beginning of 2015 to, to work on it. the problem that I was really, really obsessed with was, uh, how to make language a non-issue How to, create, you know, technologies that remove languages as a barrier. we built a really elegant solution, that made it much easier for companies to do software localization. So that was kind of how, the, the problem ended up manifesting and built machine translation models. that's how we discovered the transformer and encoder decoder technologies and how we became proficient with them. And so writer really came out of, you know, those years of experience.

Todd: Okay, when you transitioned from Qordoba to Writer was there a different problem you were setting out to solve with Writer? 

May: Yeah. And in a lot of ways it felt, honestly, I don't know if I've said this before, it felt like a betrayal of everything we had spent the last few years working on because, with machine translation, you're working always on target language and with writer, we now are pointing our sites to source language, English, and especially given that, you know, we had entered, the space for me, it really came from a place of wanting to address a lot of inequities, uh, that I thought were present as a result of, you know, folks being native in one language versus another. It felt like an abandonment of, that cause. But the reality is, being able to use AI or not is likely to create 100 X more in equity in society than all of us speak in different languages. and so, for lots of reasons, you know, built up the, the, the courage in the stomach to, work on a completely different problem.

Todd: And so in those days when you were just sort of, thinking about Writer, how did you go about validating the market and understanding kind of what the market wanted? 

May: Very deliberately actually. again, I don't think I've said this to anybody, but I literally read that Eric Ries book. I had done it in my first company. I'm like, Oh, let's see what this way more deliberate, startup thing is. I can't remember if I did a lot with it, honestly.

Um, but it really was like, I did feel like it was a very intentional process, but you know, it just went. Really came natural. Talk to people who were using competing solutions to do what we were trying to do Jamie Barnett, actually, uh, Jamie, if you're listening to this, hello, uh, was one of the people that I, approached with that problem.

She was the CMO of Netscope and we were, we had sold localization to her. You know, I thought chief digital officers, chief marketing officers, content people would be interested in the first iteration of the Writer product, which is an AI writing assistant. and she just loved it.

And you know, took us into like a whole, her and others, a whole category of problems around content and strategy and consistency. And it really felt like, Whoa, writing is super strategic, but is the last unstructured business process that a company does. You know, those were really the ideas that we started Writer on in, in 2020. 

Todd: And did you go to other companies too and kind of hear similar things? 

May: Oh, yeah. Dozens and dozens. And, I actually, did one of those surveys where you tell them who you want to talk to and we went and talked to everyone who was using Grammarly Enterprise. so, you know, it was really, it was a very deliberate market research effort. 

Todd: All right. So May, take us back because, you know, sort of like the very first product that Writer had, cause I know that you had been talking to these various business leaders that, you know, they weren't happy with the current tools. And so you've got these product ideas. And then I know you also, interestingly at Writer have built your own family of LLMs called Palmyra.

And so tell us those sort of early days of the first product and the first kind of technology.

May: When we launched Writer, the big vision slide in our sales material and our investor material, you know, showed how we're going to go from an AI writing assistant to AI writing. And there were customers where we took that out because it like was just too scary, you know, and it would like freak them out what we were going to do. For us, it was really just never a question that's what we were going to do. I mean, we, we, we started Writer because we knew Transformers. when GPT 3 was such a leap ahead. Right, we actually switched for a little bit for some of the microservice because it's like, Whoa, this is really good. But we caught up real fast. And now we've got 18 models. The biggest ones are GPT 4 level quality, but at a fraction of the cost of hosting. Super, super fast, for the enterprise use cases that, that our customers run them on. We now have a Palmyra, financial services model, you know, customers like, Vanguard and Northwestern Mutual and Goldman use, a medical model that UnitedHealthcare and other, um, uh, healthcare customers use. So for us, it was, you know, just always on, on the roadmap and, never felt contrarian as I think it felt to the rest of the market, where, you know, now I think increasingly, kind of having your apps built on a black box access to a large language model that you have no control or insight over, the weaknesses of that are starting to be really clear. 

Todd: So what was the initial product that sort of started catching on with some of these customers?

May: It was an AI writing assistant that helped folks across the company. So from support to sales, to marketing, follow, uh, and be compliant with, company rules around everything from brand voice to regulation around what you could and couldn't say. To kind of customer service best practice. So think like kind of canned emails on steroids for support agents, et cetera. When we launched generative features, we were really able to be head and shoulders above in quality, and customization right from a customer's perspective. And we've since really built out a whole platform. So we call ourselves a full stack generative AI platform. We combine large language models with all of the tools that companies need to build useful applications. Everything from RFP responder apps, to apps that help health care companies really modulate voice and tone to kind of get, high risk populations to respond to messages for care. The generative AI market has really evolved, obviously, since we started. Right now, enterprises are looking at a space that is mostly API access to raw large language models or productivity tools right? Here's a copilot license. It's a panel on the side of everything you use, go be productive. And the most impactful use cases really sit in between kind of those two poles, but it is really hard, expensive, and time consuming to connect large language models with customers data, with AI guardrails, with business users and business logic. And so we put that all in a single solution. It's highly composable. It can sit inside of a customer's own cloud. And that has been just an incredible way to, of course, you know, increase our footprint in a company and for, for the customer, really get high ROI out of generative AI. 

Todd: Okay. So you had marketing, you had brand, you had customer support, you had RFPs. Like did you go after all of those use cases kind of right from the beginning? 

May: There's like a pre Chat GPT world and a post Chat GPT world. So in a pre Chat GPT world, we landed in marketing and then really felt our way through the organization.

So sometimes went into support right away. Sometimes went into sales. Post Chat GPT, there's been so much interest, in kind of a horizontal generative AI product right? I don't then as a CIO and CISO have to go out and audit 800 models, right, to give my business users access to generative AI. You know, I am going to use GPT 4 and maybe one open source model and then I'm going to give them Writer, they can go build their own stuff on, on Writer and we can customize a bunch of our own applications.

And so, as a result, over the past year, we have really been selling to IT and the office of the head of AI. and then they really bring us into that first sort of big, impactful use case. And it really tends to be one of two flavors, either a customer experience use case that could be marketing or digital, or a, uh, we call them expert assist.

So, a knowledge management type of use case. you know, if you abstract away from kind of the function and the use case, what we're doing is combining structured and unstructured data, right, synthesizing it in some way, grabbing insights, doing generations, testing it against AI guardrails, which would be everything from brand to compliance, and putting that in front of the user, either in UI that they built on our product, or, you know, an interface that they've integrated Writer into. 

Todd: Okay. So in those first couple of years, you know, as you're, building out the LLMs yourself, you're building out the product. What were some of the like unexpected challenges that popped up for you? Either, either on the product side or the technology side. 

May: I don't think we really, had a sense for how the RAG market or RAG space would, evolve. Like, I think, even 18 months ago at Writer, we weren't talking about that, 

Todd: Can you just explain what RAG is to everyone May just so everyone understands.

May: it's a, it's a technique called retrieval augmented generation, and it basically, you know, will. augment the training data of a model with an external, you know, think of it as a database, that contains the structured unstructured data that a customer, really wants to append to the query. So let's say, Todd, you're like, Who have we talked to in the last two years at First Round that is, you know, founder, covering this type of space, et cetera. If every partner's emails are in a writer knowledge graph, you can actually answer that type of question, right? And that's different than the enterprise search market, right? Where you are really using enterprise search to kind of surface work product that has existed or has been built, right, or places where we do things or strategy docs, et cetera. Where we sit and where knowledge management use cases are very interesting, I think, is when you are really appending an LLM to structured and unstructured data to deliver work product. That really was all very new to us. And so there, yeah, there's just been so much, I don't, I don't know that another massive market, when this is all said and done, will have been as dynamic as this one in such a short period of time. When a market is really dynamic and technology is really emerging, the challenge is to take conversations you are having and get them to the product team as fast as possible. If you, when you're, if you're doing that, when you are scaling, that's like really hard. It used to be the case that I could go into Gong and like, Oh, what conversations did we have this week?

And I'm scrolling, you know, and I'm listening and clicking. Now I'm scared to go in there because I'm literally scrolling, scrolling, scrolling, scrolling. All right, that was Monday. Next, you know, and it just makes it impossible. And I'm excited for their AI stuff to get better so that you can actually, like, search across all of it.

We're not there quite yet, we, we just did things that didn't scale, right? I just heard this. Voice notes became my friend. So like I would come out of a conversation with somebody or see, you know, a technical leader use our product in some way or hear a great idea or hear that they're talking to somebody else about, you know, another company about something they're doing.

Hey, can we do X? And so that kind of like really closing the loop, you know, as we were developing our approach to RAG, it really did feel like zero to one Writer. And I do think that is the defining feature of a generative AI native company, is there a broad swath of your company that is in constant zero to one mode, right?

Do you really become a learning organization? And so we, we've now incorporated that into our operating principles of when we act with urgency and focus, we're doing that in a context where we expect change, where we expect flexibility, where the need for all of our learning curves to be steep. And I do think the capabilities that are possible are gonna outstrip most organizations' capacities to absorb them. For most of their employees to absorb them without a lot of effort. And so, as the company that is creating the tech, we can't let that happen to us. So, you know, the, the, the challenges in a lot of ways, I think we turned into strength.

Todd: So do you have any examples of that May? Because I think it's super interesting. You're in one of the most dynamic kind of markets ever. Part of you wants to sort of keep the team focused and like, we're doing what we said we were going to do, shipping the roadmap.

And then I assume part of you needs to sort of be paying attention to what's going on in the market and the technology is shifting. Are there any specific things you do in terms of how you lead the company or run the company to sort of allow both the execution to happen and sort of to keep your ears open and your mind open to new stuff?

May: We're not a culture that clicks on every AI headline. You know, there, there's a group of us, what's great about research, Twitter now is it's all of the AI research, in threads. It's never been easier to consume, you know, the latest, in LLM research specifically. Uh, and so there's, you know, a bunch of different DM groups on Twitter, Writer people, as we kind of dissect and debate and think about things. Then there's like some DM groups on LinkedIn, that's more kind of competitive Intel. And then, you know, our competitive channel and Slack, nobody shares shit that we know we all saw. there's just a really high bar for sharing what we think will have passed folks' radars when they are kind of skimming the headlines in the morning, etc. So overall, it's just a really high signal to noise, right? The stuff that sits there is, my friend was interviewing at X and heard Y, someone interviewed here and told us A, you know, I was literally like at an AI meetup or whatever, or someone's doing X. So like it is, Intel. And so, you know, when you add that to how tight the feedback loops are from our actual customers and our actual prospects, you know, I think we do a pretty good job 

just focused and and we're really up front what we don't do with the team. We don't do chatbots. We don't do ticket deflection. No one's going to build something that sits on a website. If it's zero shot, they don't need any data, right? And it's not medical and it's not financial services and it doesn't need to be on prem, they should use GPT-4 So, you know, we're really specific, about what we do and what we don't do. We don't work with agencies. We don't work with SMBs. We don't work with the corporate segment. there's just so much that we don't do. And we tell everybody what we don't do. We tell customers what we don't do. We tell customers what Writer is not good at today. 

Todd: I think this is fascinating where for companies have to sort of say, here is what we do. Here's what we're excellent at. And therefore, here are the million things that we're not going to do. And is that just sort of like how you are programmed to be very focused? Or, did you sort of like figure these things that we were not going to do? And the list kind of accumulated over time? Or how did that develop? How did that strength of opinion develop so strongly?

May: We're growing really fast and the reality is when you don't have the the benefit of, you know, instant distribution or, you know, universal name recognition, you got to be really focused and intentional to continue to grow that fast. And so, you know, as our team kind of dove in like, okay, we're trying to grow as fast as Datadog and Snowflake. There's a huge gulf between, you know, the trajectory that we're on kind of mimicking them so far, which is really great and literally thousands of other companies. What are we doing every week to make sure we're not thousands of other companies? Actually, what are we doing every day to make sure we're not thousands of other companies? And so it really sets the bar super high for how specific and focused you gotta be. We rolled out to 21 brands at L'Oreal, right? We know the 30 ish use case families that we did across that whole company, we built expertise around that. We built expertise around insurance. We built expertise around, managed care organizations. We're going to do more of that. And so really being able to like paint a picture to our employees around what business impact looks like, right? Showing people that we've got champions who are signed up with procurement to take a $100m of costs out of their organizations, completely built on Writer. We had a customer who posted their best quarter ever. And they attribute it directly to, they call it personal concierge. They rolled it out to 1500 people. They directly attribute it to raising the level of intelligence about their products and their positioning and their messaging across their company. And they built that on Writer. When you paint such a vivid picture of what customers are doing today, everybody is really clear when a new prospect comes through the door. Are they that company that's going to go out and look like a L'Oreal or an Accenture or a Commvault? Is that the champion? Or are they not the kind of person who sticks their neck out? who actually has the respect of their peers? Are they not the kind of person who is going to be able to have the positivity and the energy required to like, not move on to the next shiny toy, right?

And so our team is really good at politely declining to engage, right? not just because the use case doesn't fit or it's not ICP but we qualify the champion out, even if it's a CIO, 

Todd: You've got a really impressive list of enterprise customers. You mentioned L'Oreal, but you've got Accenture, you've got Vanguard, you've got Dropbox. I think it's often really tricky for startups to sell right into enterprise the way that you did. Was that, did that come naturally to you? Or how did you do that?

How did you figure that out?

May: We from our first company knew the long tail of enterprise features already. So your skin, your SOC 2 Type 2 your user groups and being able to tag on how to set up organizations and teams, how to be able to have settings that played well and could be inherited. All of the ability to roll out to thousands of users in a way that was collaborative, right?

And what I mean by that is folks could run reports around who was doing what and what kind of impact, the product was having and get insights just on their own utilization. That was just something that we invested product cycles from the get go, so that really helped. Understanding what enterprises needed from an AI guardrails perspective. The fact that we had such deep LLM and NLP expertise meant that, all of those hundreds of microservices that and some of them LLM based, some of them, you know, ML based that are applied post-processing on content that comes out of or answers that come out of Palmyra.

It was just really enterprise ready from the get-go and obviously, it didn't happen overnight. Some of our first customers were Enterprise. So Intuit and Accenture were 2020 customers of Writer. And so we learned on Enterprise use cases and what we would need to do pretty early. I think we had, like, 50 users at Accenture when we first started. Jill Kramer, the CMO, just gave everybody in marketing and comms licenses, thousands of people. and, you know, we had built our way up. It definitely is, not for everybody. You know, you're on site, you're doing workshops, you're doing office hours. But it is, it's so impactful. There are a lot of our enterprise customers who tell us that our account teams, you know, the solution architect and customer engineer and, the the CSM, the help manager know their organizations better than they do.

And, that is really, that's really cool. You know, I look at some of my friends and they're doing startups where they're excited if their software generates a hundred K for their customers and we literally get to take out a hundred million dollars of cost at, one of the customers you talked about.

So, you know, it's just a completely different scale of impact. We've had so many users, literally hundreds tell us they've been able to keep jobs longer as a result of like getting over, you know, ADD and frustration at work and being able to like access knowledge faster. It's just, you know, it's a next level of work transformation that generative AI brings and Yeah, we love working on this product.

Todd: Yeah. So I was going to ask that, you know, there's a lot of founders listening and pre founders, future founders. and if you were to sort of give, hey, here's a couple of pieces of advice about how to succeed in enterprise and how to succeed in making enterprise deals. What are some of those pieces of advice?

May: Succeed in enterprise, first, and then the deals will come second. I think, you know, the, the deals around POCs and innovation and, you know, sort of, I call it funny money. You don't really want that stuff, actually. I think that is a distraction because it's easy to cut. your team will have spent a lot of time and it'll be for naught And I think that's why a lot of founders ignore the enterprise, right? Or sequence their way through it years later,

Todd: How do you know if you're selling into the funny money or the innovation 

budget? 

May: The person has an innovation title and the money is innovation money, you know, versus the CIO is funding it or literally the head of North America. This is their ops budget. You know what I mean? and so in generative AI, you actually are in a constant state of reselling because it's such a noisy market, your buyer, your champions, there's so many shiny things come in their way, right?

So many people promising to help them sell more, save more, et cetera. I think in the enterprise, the prize is certainly, you know, bigger ACV. if you do it right, right, you are so sticky, you're going to, you're going to help. Everybody ignore all that shiny stuff. and I think it is just so much easier in mid-market and SMB from a functionality perspective to just switch right to something else, especially in generative AI, where, it'll be confusing for a while, it's going to be hard to scale a big business if your buyers are constantly right, shifting focus, shifting attention. And so there are a lot of reasons to go after it. But I do think really focusing on the actual urgency and pain and business need and not the generative AI. Let's try stuff fund, right? It's kind of like classic SaaS Don't let, the sexiness of the space really distract from the fact that sooner or later folks will be asking to see. ROI and, and actually for a lot of enterprise deals, you know, that happens very quickly, meaning that there's a lot of work about work that if you aren't productizing that, whether it's processes and approach or actually in the product, right? We do both. your teams could actually spend a lot of cycles on that and not actually deliver the value in the product and the rollout, et cetera. I think the deals part is, you know, after you have proven the business value, right? That's actually when the deal starts, right? There's got to be a really compelling thing. That has been proven already for you to get serious money out of the real budget, right? And serious money out of the real people.It's as simple as that because. We are not Microsoft. We're not Google. Right. They could be fired for choosing writer. and it is amazing to be able to see how, the trust with that team and the community that develops around us building, like, reference. Customers and just reference peers around that new champion, that new buyer, just how effective it is, how beautiful it is. we, we get our communities together a lot at every level, the AI program directors and managers, the practitioners, the executives. This is a really cool industry in that. You know, it's not like execs are like, Hey, person, three levels below me. Go try out this thing. They are hands-on we have had C-level people build apps and writer. It's like, so cool. I we're, we're doing a, another customer conference. This one's in New York in March of 24. And the two people who are giving the app studio workshop. Our customers are the champions of the customer is not our product team. 

if you are listening and you are a founder in AI, the enterprise is so great place to be because there are so many people, that are excited and the impact can be so large, but if they're sticking their neck out for you, you gotta be there for them. And that I think is the biggest thing.

Todd: When it comes to product market fit, was there like a moment where you were like, wow, Writer has product market fit. I can feel it. Or was it more sort of like a steady climb to getting to product market fit?

And how do you sort of think founders should think about that?

May: there certainly was, but, there's levels of stickiness once you are past a renewal threshold level of product market fit, We have 209 percent NRR. I like to strip that away from deals that are bigger than a certain amount. And then look at the NRR on just smaller accounts, right, who've got lots of optionality. That's still 160 something percent NRR, which is amazing, and the, the GRR, both revenue and logo, very healthy stuff that we look at all the time, you know, way, way higher than what a lot of folks in, in AI have and, you know, on par with kind of leading SaaS companies, folks that we have mentioned We look at, that amount ratios and I've spoken about this before, right? in the enterprise, the bar is really low, unfortunately. we, we don't like that. We much prefer consumer-grade actual. activation and, activity level and the way that we define being active is also not, you know, being logged in or doing one thing, but like doing the most important things, which are the reason, you know, somebody bought the product. for us, You know, we got to really great NRRs, uh, late 21. mean, they were, they were good when we raised our A, but they were really, really great, you know, starting a couple quarters after, right? This team, the, the product velocity on this team is, is ridiculous. and now, you know, it's just, it's, we're off every chart, you know, in terms of, what that looks like. I think this should be a 400 percent NRR company, the team knows that, and, you know, the land and expand works really well. This is a see it to believe it type of, thing, generative AI, right? You just really gotta, show people that it's intelligent. We have this acronym now internally, D A H, dumb as hell, you know, the, the digital assistants people build that are supposed to be really intelligent, but are D A H, and so, you know, you won't go back and use something if that's the case. you know, when we're talking to a customer, the first thing I pull up in Slack, we've got a command for it, is literally their, you know, their activity levels of, hundreds, thousands of users. 

Todd: Okay, May, I know you have some hot takes on AI, so let's, let's get into that. Palmyra, this, this family of LLMs that you have, I mean, is that something that you'd recommend that, you know, more companies invest in is, is sort of building and training their own smaller models how do you think about the strategy there?

May: Yeah, I would recommend it for sure. we are committed to state-of-the-art large language models for the use cases that we focus on. We don't do code, we don't do imagery, you know, we're really focused. It's what it says on the tin, right? Now we define multimodal pretty broadly, like as ingest data, but we are outputting words, right?

Whether that's an insight, whether that's a chart, whether that is a generation, whether that's an answer, it's, it's words. you know, a use case that is as narrow as that, you really could commit to state-of-the-art on that, for your underlying model. you know, for us, I'm really happy we're at GPT 4 level, but we don't have to be for, A lot of the use cases where, click below that in quality plus fine tuning and rag and AI guardrails is just a much better outcome, right?

Than the reasoning capabilities. Now, folks are always like, well, what about GPT 800, you know, and Um, I think folks get scared whenever, Sam says anything about the AGI capabilities of whatever is next, but no matter how smart the reasoning capabilities of our state of the art models, right, writers included, opening eyes included, there's just going to be a bunch of data that they're not going to have access to, right?

And these are transformers. It's not rocket science, right? They, there's got to be a way that you integrate your data with those models. And for us, the, Vast majority of enterprises we talk to that stuff can't leave their environment, and they're not excited about the kind of lock in that an Azure requires, right?

And the way that that scales the amount of engineering effort that's required, etc. So we built a whole business around, you know, that problem. that is scaling. generative AI startups targeting the enterprise face A uphill battle, but, you know, other spaces don't have and not the enterprise buyers are really uncertain about how the space shakes out.

Right? There are a lot of people on the sidelines. you have to figure out how to ignore those people and like, kind of weed through them. and weed through their unease about the speed. Of progress and just not knowing how to future-proof or not future-proof or, you know, what to do, to get value here. people are waiting for the dust to settle and there is no settling of this dust. 

Todd: Okay. So let's talk about, uh, Writer in 2024. what's next for Writer? What are you super excited about this year? 

May: I'm really excited about our next class of models. We're calling them large reasoning models. it's what we think is next after the transformer. Things that'll enable, agentic use, within our enterprises. you know, our kind of tests on inference, on quality, for agent use at work. It's just transformers.

Just it was current capabilities really not able to do what we wanted them to do. And so new architectures that we're working on are super, super promising. So I'm really excited about that. I am really excited for the world to continue to discover kind of the power of. AI graph-based approaches to rag, and we're seeing so much more inbound on people being like, okay, I am not getting the kind of answers and quality.

I need, like, it's fine when I am, I've indexed 50 pages, but when it goes to 50, 000, you know, I can't get this needle in a haystack out of my kind of vector DB approach. So I think 24, The folks who have spent the last two years, you know, to their elbows in this stuff, our customers will get to enjoy kind of the, the, the fruits of that first innovation that I talked about.

And, you know, the folks who are coming up to speed, I'm excited for them to, to really dig into our, our knowledge graph functionality. So, yeah, great year. And then, you know, internally scaling and bringing on The most fun, smart, exciting group of folks. Um, every time I meet a new writer class, it's like, whoa, pinch me. So yeah, I'm, I'm excited to work with all the new people that we have hired and we're recruiting. You can imagine every single role in engineering and go to market. So meeting all of our new folks, spending a lot more time with customers. we did our first customer conference in December. We've got one in Q1 in New York, one in Q2 in London. We're opening up, our London office and a Q1. I'm in London right now. we're hiring internationally as well. So if you are listening, New York, San Francisco, London, you can work from anywhere and those are our hubs. 

Todd: So May wrapping up now that you've done, you know, the zero to one journey twice. Uh, what advice would you give to early founders who are just starting out their own journey?

May: I think surrounding yourself with people who are going to lift you up when you are just absolutely utterly committed to being down in the dumps. Is probably the best thing that that you can do, you know, there are various parts of the journey both times where, you know, I have talked about not doing it out loud to, one or two very close confidants and definitely between Qordoba and Reiter Talked to a lot of people. the folks that told me to pack it in, I remember them. I will never forget that, you know, and you obviously don't listen. When you're as persistent as, you know, most founders are, and it's just really important to make sure that that last phone call that you have before you make the decision, you know, is going to be somebody that told you to just hang in there. 

Todd: But what was it that kept you going?

May: Oh, the team, we have such an amazing team and the idea of not working with them again. It's just, oh, I can't, I can't bear to not have that be the case to have to like, start from scratch, and obviously like you could all start something together again, et cetera, but it's not just you and your co-founder and a couple of people, right?

Like, when you really get into the groove, it's dozens and dozens of people now are in a rhythm making music, and you just want to, you want to get through the hard times together. And it was really like, it was really only 2020, right? we started right before COVID It was hard, to figure out product market fit that year and get to just where we wanted to be on the technology, right? Like this was also new to us. We had been in machine translation. and so that was, that was a hard year.

And now. You know, that team is just so, so tight, right? we didn't lose anybody. We had one person, move on, regrettably. But it was, yeah, it's, that whole team is, still here. 

Todd: And by the way, how did you get connected to the Writer co-founders And what sort of gave you conviction that you all would be a, such a strong co-founding team? 

May: Yeah. And it was a decade ago. Uh, and now I can't believe it. yeah, Waseem is my best friend now and, you know, we, we share so much and our families are so close and I think, you know, in the early months and, and really a couple of years of, you know, working together, before we moved to San Francisco to start Qordoba, just, side hustling type of things. The chemistry was always there, the creative chemistry. He'd push me, I'd push him. We thought so different and it was very yin and yang, but there was always so much, trust and respect for each other's ideas and like a real sort of headedness, honestly, of getting to the bottom of why do you disagree with me? I really don't get it. And so, you know, that has definitely, you know, developed into kind of a brother sister, super intentionally, diametrically opposed. And we actually, it's something that we, onboard people into. This trust battery is infinite. We will not shy away from disagreeing with each other in front of you. We don't care who you are. It is not anything to be scared of that your founders are disagreeing about something pretty critical in front of you. That is deliberate. And so the whole culture of connect, challenge, own at Writer I think a lot of that is derived from, you know, mine and Waseem's relationship, the relationship we have with Doris and Maiko and Mo'ayyad and Brock and a lot of the, the folks that we've been working with for a long time.

Todd: That's really interesting to be able to debate with each other so openly in front of, you know, others in the company. Are there things that you tend to gravitate towards or think about versus Waseem? Like, what are some of the debates consist of? 

May: You know, I'm, I'm always, holding the, the customer voice primally central to any conversation that we're having. Like, I am and lots of other folks, dozens of people are doing this every week at Writer, sharing clips and quotes and like literal words that somebody said, and we train people on it.

You're an SDR. I was just on with a CIO of a fortune 10. And we had an SDR has been on the team for 2 months during the call. It was very deliberate because He joined a CIO call earlier in the week where he hadn't done this, and it was a very specific, your job is to listen and the most important things, the most insightful insights that you hear this person say, that's what we want to go into this account channel. Really just helping people get inculcated into the mores of what it means to listen to the customer and have what they need, what they want, what their feedback is central to every conversation. So that's always kind of like what I am bringing to a founder, a friendly founder debate. Waseem's always thinking about and when 10, 000 people want this, or they're the only person who want this, right?

Kind of the two extremes, then what do we do? The framing he's bringing tends to be really around, it's a much more framework based you know, really thinking, around corners. And so that's why we're such a great pair really is so complimentary and that's why we also don't shy away from arguing in front of the team, and debating in front of the team. It also really reduces politics for folks to see that, right? We have an executive meeting every other Monday and then on the alternate weeks, it's a team leads meeting. So smaller meeting, a big meeting, but all leaders in the company and you know, we don't pre-bake stuff like we send pre-reads, but, you're not kind of being like, let's make sure May's on the same page on my idea, or, let me make sure Waseem's on the same page with me on on this proposal. Because me and him demonstrate that we don't do that. and we're really comfortable having very nuanced conversations openly in a group, and so we're able to have a really low meeting culture as a result, because, you know, we are able to do things as a group. That meeting is 90 minutes.

Sometimes it goes 2 hours a little bit longer. But, you know, as a result, I don't really have one-on-ones, we do them on an ad hoc basis and then folks meet, you know, as, as needed. But we really do try to keep to an asynchronous culture. It's a very fast moving company. And so just the level of trust has to be so high. The level of openness has to be so high. The, the level of comfort being uncomfortable, the ability to kind of be okay with eight people not liking your shit or disagreeing, you know, like we're pretty upfront about that. And so, yeah, a lot of that comes from our founding relationship. 

Todd: Well May congratulations on all your 

May: success with Writer Thanks, Todd. 

Todd: and really appreciate you being here. This was awesome. Thank you.

May: Thank you. Thanks so much.