The REAL Reason AI Gets the Attention But Trust Gets the Purchase Order

Tom Davis

AI gets the attention, but trust gets the PO. If an enterprise doesn't trust that you are safe, that you're going to be compliant, that your data retention policies are correct, you got the coolest demo in the world, but they're never going to buy you.

Samuel Boulanger

Tom Davis founded Solair in 2011, built it on Azure, and sold it to Microsoft in 2016. Today, he's a partner at Microsoft for startups, leading the Startup Success organization, and helping founders turn AI into real business outcomes.

Tom Davis

We've already seen a couple of companies become unicorns in just one or two people. I think that's going to be far more the norm. I speak to founders saying, "Well, this time I'm going to build a company that's no more than 20 people large because today you don't even need to go and get teams of engineers. You can actually get started just with an idea. You don't have to be technical. You can really just ask the system to build what you want. Get something in front of users very quickly and get that feedback."

Samuel Boulanger

Tom Davis founded Solair in 2011, built it on Azure, and sold it to Microsoft in 2016. Today, he's a partner at Microsoft for startups, leading the Startup Success organization, and helping founders turn AI into real business outcomes. In this episode, we get into why speed to a working product isn't the advantage it used to be, why plugging into something like Copilot can matter more than the product itself, when to build a tool yourself instead of buying off the shelf, and the one mistake that kills startups that looked great in the demo. Quick thanks to our sponsor OS4 Techno Services, a Quebec-based IT firm helping organizations navigate infrastructure, cyber security, cloud, and AI with strong Microsoft expertise. Before we jump in, I have a small favor to ask. If you're getting value from this show, hitting subscribe, dropping a comment, or leaving a like is one of the best ways you can support me. It helps more people find these conversations, and honestly, it means the world. Thank you. Now, let's jump in. Tom, thank you so much for joining me on the Frontier Playbook podcast. There aren't many people who've built built a company from zero and then watched thousands of others try to do the same from the inside. And I think that's exactly why I wanted you here. Tom, you founded Solair back in 2011, built the whole thing on Azure when I think at that time betting on the cloud was still kind of a real leap of faith. And then you sold the company to Microsoft in 2016 and today you're a partner at Microsoft for startups. You're leading the Startup Success organization and you're helping founders turn AI and cloud innovation as a whole into real business outcomes. And that's exactly where I like to start because you've lived this from both sides. So if I go back to that first chapter when it was just you and an idea if you were starting that same company today with everything that exists now what would feel compelling to the person you were then?

Tom Davis

Great question and thank you very much Samuel for having me. It's a pleasure to be here. Things have changed and we can't deny that fact. Things have moved very fast over the last couple of years and it seems like we've seen decades of change in just months at times and really back when I was building the company Solair as you said back in 2011 2012 things were very different and the cloud was just really coming about and our big challenge was really which cloud should we use and we did a lot of analysis and we landed on the Azure cloud and it was fantastic and really that was because we wanted to be very focused on the cloud that enterprises used and Azure is very diffused in the enterprise sector and so we thought we wanted to build there as that's where we would have the same customer base but what is different today I think is the speed at which you can do things. You have to really think through early on around what do you want to do and how do you achieve it you'll get there very very quickly whereas when I was building it out you had to really think about which architectures were you going to use you had to be thinking sort of several years ahead whereas today you can experiment get going see if it works if it doesn't, change it and spin on a dime it's a very different type of environment that founders find themselves in. And so that has some positives and some negatives. I mean the positives are that you really can get going very quickly with very little investment to be honest with you. I mean cloud was already a big thing. You didn't have to go and buy servers and things like that which was a very big capex investment. But today you don't even need to go and get teams of engineers. You can actually get started just with an idea. You don't have to be technical with the tooling, the GitHub Copilot and the Claude Code and things like that. You can really start to just ask the system to build what you want, get something in front of users very quickly and get that feedback. But at the end of the day, that's not going to help you sell necessarily. It's going to help you validate ideas quicker, but ultimately you need to be able to do the key things. You need to be able to address the pain that your customer is feeling. And just because you've got cool tech and do it in a very smart way, it doesn't mean that you're really actually addressing the pain. And so that is fundamental. And so it's actually more important than ever, I think to focus on the basics of building a company. You have to have have to address what your customer is looking for to solve that pain. You have to be secure in the way that you do it. And you really have to have those channels to market to help you do that. So the basics are still there. It's just changed the dynamics in which you address it.

Samuel Boulanger

I agree. So the time from idea or the time to an MVP is cut tremendously. I can't even express it. Anybody right now can put an idea into a reality. But again like you mentioned it doesn't mean that it will sell and doesn't mean that you have all the ecosystem around it to make it a success. Right.

Tom Davis

Exactly. Exactly.

Samuel Boulanger

Now from your seat at Microsoft you've been working with thousands of startups so you have this experience you've probably seen thousands or hundreds of them trying to use AI to come up with new ideas and putting it on the market. So what's the single biggest change AI has forced and how those companies are operating right now?

Tom Davis

That's a great question. I think that AI, we look at AI very often as it's going to help my product do better things, do it faster, address things better. That's only one aspect of it. I think the other aspect really is how can I build my company, the company structure, and how has that changed? I mean I look at it my son is now building his own startup, a company called Ether and it does auto-grading for teachers. It's an assistant for teachers. So going back to the thing it addresses something that teachers really need and really want. But what's fascinating is how he is approaching the way that he's going to go-to-market and how he's building his company. So they're just a couple of people. Admittedly, they are both sort of masters in computer science and things. So, they're very well-versed in the technology, but it's how they're thinking. How do they approach building the company out where are they going to hire people and where are they going to have AI agents and things working for them to do certain jobs to be done? So, for example, on the go-to-market side, how are they building out the ideal customer profile? Well, before when I was running my startup, I would be there. We would be sort of doing a lot of research, manual research that's sort of on the internet, speaking to a lot of people, bringing the information together. That is all available now. I can go to Copilot Researcher and just get everything I need, prompt it and say, build me out customer profiles for this type of teacher, for this type of professor or whatever. It would bring everything together for me. So it's thinking how you can leverage AI to do a lot of the tasks that typically would have been done by humans in an organization and then it can go out and build you your whole go-to-market plan. It can say these are the types of customers. This is where they hang out in their communities. This is the type of messaging that you should be using. And then you can start building things. You can have your personalized CRM system on the back end. You can build a CRM from scratch. You don't necessarily have to go and buy a large sort of enterprise CRM and try and tailor it for a startup. You can build exactly what you need based on that. AI can do that for you. And then it can auto-populate it with the leads that are coming in and all this. There's so many automations around the business that we don't even think about and we would have done manually beforehand. So I think it's an absolutely fascinating time not only to think about how you can leverage AI in your product but how you can have AI and really a team of agents working for you and then that sort of starts asking the questions where do I invest my money that I have whether that's my revenue or as my investment what are the real people I need and how do they manage teams of agents and so I think we're going to get to we've already seen a couple of companies sort of become unicorns and just one or two people. I think that's going to be far more the norm. I speak to founders that are sort of also repeat founders saying, "Well, this time I'm going to build a company that's no more than 20 people large because they don't need those other people because especially in certain markets, they can actually have the agents do a lot of the outreach and things. They can have and it's not just sort of writing emails and things like that for you and sort of spamming people. It's more getting how do I get precise in what I'm doing? Because if I can pull that knowledge that I that's readily available or I can get agents to find and get very precise in my messaging, it really helps me be successful in that outreach. And so AI is really going to help on that. So I think the smart founders are going to be getting very targeted in their approach. A lot of people think AI is just going to be a lot of AI slop and it's like throwing lots of emails out there and just really bugging if you like people to try try and sort of read them. But at the end of the day, that's just a sort of the it's very self-fulfilling in the way that it's just going to fill up people's emails boxes and they're never going to read what you are what you're sending them. So it's actually a very negative negative approach to it and a downward slope.

Samuel Boulanger

If I take back your example of building a CRM because I think it's a really good example. Let's say how do you identify where you should use AI or not? And I like the CRM example because not everyone knows what exactly a CRM is. So if I'm asking AI to build something but I just don't understand what it's tied to or what my own business processes are, the CRM might be useless in some cases because I don't understand the structure I don't understand my own processes now how do I identify where I should let AI do the work like in this example building me a system or should I take the off-the-shelf product and not spend my time trying to figure out that part or maybe getting the AI to build something that is not ideal. Because I'm seeing a lot of people getting lost with AI trying to use it for everything and first it can be very costly and second it's still your time. If I'm spending two weeks trying to build something I don't really understand I might have been better going with an off-the-shelf product. So what will be your recommendation on that?

Tom Davis

Great question. And I think it's very personal. It depends who you are and what your skill sets are and what is going to be valuable for you as a founder, for you as a startup. Where is your IP? And IP is not just what you've invented in your technology. It's what are your specific areas of specialty. If your market is very precise and you have very exact workflows and you can clearly understand them and can articulate them as a founder or as part of the founding team, you may want to build your own CRM. For example, I mean I have a great example. I can remember reading the LinkedIn post of one founder called Dave Clark who used to be the CEO of Amazon retail and he's actually one of our startups called Auger. So, a pretty big company but very serious senior executive and he was saying yeah we couldn't quite find the right CRM. They do supply chain and he said we couldn't find the right CRM for our customers. So he wrote one over the weekend and this is a senior exec because the tools that we have available because he understands exactly the processes and the requirements that he needs. He could build that on his own and there wasn't one off the shelf that he wanted to configure or things like that. So it depends where you are in your journey as well and you may start off building your own or you may start off say this isn't critical for me today. I'm just going to take something off the shelf and all I want to do is just manage, I don't know, the names of people, the companies, work out what my forecast is and things like that. There's nothing fancy about it. But then over time, as I learn and I develop and get more customers, I realize that there is this very particular selling process that I have to adhere to. And that may be the time to go, well, I'm going to build my own at this point and just import that stuff. And so it depends where you are in your flow and what your knowledge is. And you have to make those trade-offs because everything in a startup life is a trade-off because your time and resources are incredibly valuable and can be used anywhere. And you can you need to make those decisions of which way you want to go to invest it in doing this or to actually just get something off the shelf, maybe postpone doing something more custom and ad hoc till later on.

Samuel Boulanger

Quick break to thank our sponsor OS4 Techno Services. OS4 is a Quebec-based IT firm with 25 years of experience. They cover the full stack consulting, infrastructure management, cyber security, cloud and AI. Their hybrid outsourcing model gives you the technical depth of a large provider with the proximity and flexibility of a local partner. And they have a strong specialization in Microsoft technologies. If your organization is navigating IT complexity and wants a team that knows your environment, visit os4techno.com. Now back to the episode. You mentioned understanding your processes and understanding your requirements. If you don't understand your own needs, it's very hard for AI to build the right solution. Right. Something I've noticed I have a lot of well-being ads and content in my algorithm because I'm following that and I'm seeing a lot of people trying to do the same thing like life coaches or gym coaches all trying to build the same application because it's now feasible for them. They can just pay a monthly fee to any AI tool and get an app built in a couple of hours. So when a thousand teams can show the same idea just as fast like concretely what makes one of them break out because if I'm looking at my feed on Facebook for instance I have a hundred coaches that all build their apps and what will make one stand out more than another?

Tom Davis

Great point. I mean, with this we really have to define what type what is a startup and what their business is. I think it falls into two categories. If we're talking about life coaches, gym coaches etc. The actual business is teaching people, coaching people, improving people, helping them get fit. It's not the technology. Technology is an enabler for them to deliver the service that they want to provide. So for that group I think it's fine everybody can build what is comfortable for them and as you say the technology is open for them and that's where sort of great companies like Replit, Lovable, or Bolt are really helpful. You can build those technologies very quickly and it will help your business. The second group is companies whose business is technology and you are completely right it is very competitive very competitive it's never been easier to build a technology application and in theory sell it so there are three things around this which are fundamental I think the first which has always been true is you have to have very clear customer insight and really understand the problem that you're solving for the customer. If you have this clear and you build a technology that resolves that, you're in a great position because you're building something that people will want to buy. So that's the first point. The second point is distribution. Distribution is key. We've seen so many great products over time invented, but they just can't sell them because the startups haven't got the distribution. And this is where things like Microsoft Copilot come in as a key distribution channel. If you want to sell into enterprises, you really need to be able to plug into what enterprises are using. And so I go and speak to CEOs of large enterprises and the things that they say is we do not want another user interface for our teams. We don't want to train them up on something that they have to use and things like that. And so if my distribution is to try and sell this new application directly into an enterprise, I'm going to be combating that issue every time. And this is where Copilot comes in. What we're seeing is we're moving towards more agentic applications that Copilot becomes the UI for AI because it's in the workflow of all the users today. They know how to use it. They don't have to go an enterprise wouldn't have to go and retrain people on the UI. They wouldn't have to go and change the implementation structures and things like that and actually get it to because 50% is the value of the application. The other 50% in a buying decision is how do I actually get people to use it within my organization. So we're seeing a lot of startups now plugging in to Copilot and surfacing their application that way. So that's a key distribution channel. When you see people like the NHS in the UK that have just signed up for 500,000 seats of Copilot because everyone in their organization will be using it or KPMG 250,000 seats same thing that is definitely a signal for startups to say oh right if I want to plug in and I that's a great way for me to get that distribution so distribution is fundamental and the third is really to start thinking about how am I going to be able to tell enterprises that I'm trustworthy. Okay, it's like AI gets the attention, but trust gets the PO. If an if an enterprise doesn't trust that you are safe, that you're going to be compliant with something like agent 365, that they'll be using to manage agents and things like that, that your data retention policies are correct, that you're not looking at PII data and things like that, they're never going to buy you. You got the coolest demo in the world. But that is such a fundamental thing and very often startups miss that and they very they're so focused on what is the coolest thing I can use with these these models. They're not actually thinking about the bigger picture about how do I get it distributed and how do I actually make it trustworthy for enterprises just to say yes.

Samuel Boulanger

Now I found my idea. I found my customers. I'm solving a challenge. I've solved the discoverability problem. Now, how do I make sure that when the models evolve and the models or the providers of the models like Anthropic, like OpenAI, like Microsoft are slowly turning some products into features that are now part of the models out of the box. So, how is a founder able to build so the next model release makes them stronger instead of erasing them completely? Because we've seen thousands of startups at the beginning of this AI wave that have just appeared and disappeared because they were basically a wrapper around an LLM.

Tom Davis

Yeah, that's I mean this has been a concern since I think it was it December 2022 when chat GPT came out and has been evolving ever since. The wrapper concept I mean we have to define really what a wrapper is and what it's not. Everybody today leverages a model in one way, shape or form in their technology. I mean, it's just the way that it that it happens. It's very similar to in the sort of the dot-com era. Everybody started leveraging the internet. You're leveraging something and we're going to get to a point where we're not talking anymore about which model are you using or whatever. Doesn't really matter because the model we will just take it for granted. It's like when nobody says, "Oh, I'm an internet-based product anymore." I mean, it's just taken for granted. It's like, "Yeah, I use models and I will choose the most efficient model and I'll probably swap the models sort of when a better one comes along, a cheaper one comes along and I use different models for different tasks within my product and things like that." So, that's already becoming the norm in the way that people are approaching how they're building applications. What I think we need to go back to those fundamentals we talked about earlier is what are you doing which is to address the business problem of your end customer is that deep understanding. So how do I interact with that user? Am I solving the problem? Am I presenting the information in the right way at the right moment in time for that process that they're addressing. So a lot of the I mean we fall into a bit of a trap on in the west coast of the US of getting so fixated on sort of speeds and feeds of models and how they perform and things like that we actually forget that the people using the applications that we're building with the applications are not super technical and don't really care about the speeds and feeds. They're more interested in how fast can I grade my students' homework. If I go back to the case I was talking about earlier with Ether the how well can I manage the supply chain with how can I improve the procurement cycles and make sure I'm looking at the different contracts in enterprise procurement. We have a company called D-Silo that do exactly that and help people find the extra percentage points by getting more precise with contract management. These are the real problems. If we did it with paper and pen and we got like the savings that we get with AI, people wouldn't care. Doesn't matter if it's AI or done with paper and pen. They're looking to solve a problem. And that is the key thing. And if I'm architecting thinking how a model can improve over time, it will do it something more precisely and faster. I need to build that into my thinking. But it's what I do around it and how I address it and how I work with other systems around me. And it's just it works out of the box. And as a as a user, I just want something to work. I don't want to have to fiddle around with it or anything. The model may be super powerful, but I'd have to go and configure it and fiddle around with it to make it work. If I'm just a standard user, if I have a solution that just works for me and solves my problem, I'm going to be buying that every single time.

Samuel Boulanger

This is so true. We're so immersed in tech. Yeah, you and I and a lot a lot of people in technology that we just assume that everything can be done with a model while most people don't want to care about it at all. They just want something off the shelf they can use even if even if they could create their own, they don't want to go to the hassle of trying to build something.

Tom Davis

Yeah. Exactly.

Samuel Boulanger

And to that point, like going from a working demo to something an enterprise will buy is I think where startups stall because it's so easy to do this working demo nowadays. So what should a founder put in place early to clear that bar without slowing the build like making sure you can test your idea but also making sure it will be deliverable and it will really bring this value without stalling.

Tom Davis

Yeah, this is the classic challenge that very early stage founders face. It's I don't want to think too far down the line because I don't even know if people want to buy this product and so I don't want to sort of overinvest in it. Well, that's I think that's where AI comes into it. I can spin up a proof of concept. I'm not going to say an MVP. I say a proof of concept with a tool like Replit, Lovable, etc. Very quickly and very easily. And that I think allows me to get in front of users and say, "Is this something you would use? Is this the right workflow?" And test that out. It's never been faster to do that. You could literally do that over the weekend. They then need to think, okay, if they're getting positive feedback, again, going back and thinking, who is the ideal customer profile? And if you're selling into an enterprise, well, I mean, even an SMB company to be honest with you, you need to start thinking what the requirements are and it's going back to that trust. Is it going to work in a complex environment? Is it going to be secure? And getting ahead of that early on, you can do some very simple things. I mean just by in your early architectures I mean if you're going to be selling into a European financial institution and things they're going to be asking you for data residency sooner or later once you sign that contract. So you should be architecting your thought process around that or your solution around that and to make it a lot easier for yourself. And even if you're not selling into a sort of a complex environment like that, there's some basic fundamentals you need to be thinking about security from day one because there's no way on earth that anybody will actually purchase your solution if you haven't done it. You don't want to have to rearchitect your solution or build in these these fundamental enterprise security requirements. So you need to get ahead of that so you can move very fast with the solution going forward.

Samuel Boulanger

Yeah. Security, security and governance are the top of what you should worry about when building one of those solutions. Anything else you have in mind that a founder should absolutely think about even before starting to build a proof of concept?

Tom Davis

I mean, I would I would say get a proof of concept. You've got to get in front of customers. So I don't want to sort of I don't want people thinking too much about selling into an enterprise too early because building a startup means you're risking your personal relationships the stress you're putting a lot of money into it. It's a big step to take. So before you throw everything behind that, get proof of concept going quickly. And that again that's the beauty of AI. You can get a proof of concept up very cheaply, very quickly to test out the idea and take off the rose-colored glasses when you're listening to people and don't hear what you want to hear. You've got to sort of ask yourself the hard questions and then if you really are convinced after that, then yeah, you need to throw yourself into it sort of whole hog and go from there. And that's when you need to start thinking seriously, okay, what is the technology platform that I should be choosing that's going to help me in my long run? Because if you're chasing this dream, this vision that you have, you have to be convinced from day one that you will be selling into enterprises if that's your target market. And so you should be building for that and you need to be building early on. You need to be using the technology they want to use. I mean Azure fits very well at that phase as the enterprise choice. And so that's the moment where you really put everything behind it.

Samuel Boulanger

And what's the most common expensive mistake you watch AI founders make right now because they might have great idea but might not be tech people and may use AI not in the right way or for the right use cases. So what have you seen that was the biggest mistake you've seen right now?

Tom Davis

I actually think it's slightly the inverse. What I'm seeing is because AI is so fascinating, it is very attractive for technical founders because they can do great things with it, which is fantastic. But I often see with technical founders, they build something which is beautiful at a technical level. They can do incredible things, but they lose focus of really what the customer wants and they go off and build features and things that the customer doesn't really care about or it's too early and they need to build an MVP that starts selling and then they can add the sort of the more advanced features later on. Whereas they keep on holding back to release the product with the MVP and actually start getting real customers because they want to build these super advanced technical features because they're really interesting. Then you can add the features. Users are incredibly forgiving if you haven't got all the features if you can address some key things very early on.

Samuel Boulanger

Yeah, I can think of so many products or apps that I've tried that have just too many features and I finally get out of the application because it's too complicated. I don't want to spend time trying to understand an app. I will just move to the next one that is just more intuitive and easier to use. So For a founder listening, let's say I want to I want a framework that I can apply in the next 90 days. What are the three moves you would suggest they make in the next 90 days leveraging AI to accelerate their project.

Tom Davis

So I'd say first of all narrow the customer problem really understand what it is and what's going to provide a real quick dopamine hit of value for the user so they get very excited. It doesn't have to use everything, but it's going to be the one that really makes them go, "Wow, I want to use this. It really helps me out." Then you can, as we just talked about, you can build extra features later, but what's that real killer feature and who is that exact customer profile that you're going to sell to? So, that's number one. So, you can get very targeted and then build that proof loop. Get something in the hands of users to test out to see if it really does address what they want. Get them to use it. Is it something they go back to or is it something they're just going to have a look at and say, "Oh, that looks really nice." but then they're never really going to use it. Get something in the hands. Let them test it out. See how they use it over time to really understand if you've addressed that correctly. And then if you've hit those things, really start sort of forcing yourself to build for those enterprise customers earlier on. It's going to feel unnatural. You go, "Oh, no. I haven't got any customers or anything like that." No, no, no, no. Think about it now because I can promise you that very quickly you will get people that are interested and they're going to be asking you those questions. And if you're not prepared, then it's going to take a long time to get those sales. It's going to take a long time to sort of close that business. And unfortunately, you could actually go out of business in that period as well. How many times do I see people go, "Yeah, we've got an interest from this large customer and we've got to do this, but they can't actually sort of financially keep the company going for the nine months that it would take them to get enterprise ready." So, the earlier you can do that, the better it's going to be.

Samuel Boulanger

Tom, we're almost at the end of our time together. I have my last two signature questions. So if we forget about founders and companies as a whole and you tell me a bit more about you, what's one thing you do with AI that makes you more productive every single day? So, something a founder listening could set up this week.

Tom Davis

My favorite AI tool at the moment is Microsoft Scout. Yeah, it's built around OpenClaw. So sort of like the enterprise version of OpenClaw if you like with all the safety rails and guardrails and things like that and it's incredible. It enables me to do data analysis. I point it at data sets and just ask it questions and it tells me everything. I think probably the easiest timesaver for my team as the GM for Microsoft for startups. I send out each week a top of mind about what's going on to help keep everyone up to speed. And I started off by asking Scout to go through what I did the previous week and start giving me some ideas and bringing things together. And it's incredible these top-of-mind updates that would have taken me probably two or three hours to write. I'm writing in half an hour with the help of Scout. So it's just these sort of productivity gains that we can get from these types of solutions that are really helpful that I find incredibly valuable.

Samuel Boulanger

Yeah, this is great and yeah I got my hands on it and it's really powerful. Last question: you sit right at the front of where all of this is heading actually. So you see it before most of us do looking over the next 10 years. How do you think AI is going to reshape the way we build and run companies?

Tom Davis

I think we're already seeing the start of this and we touched on it earlier is how do you build a company in terms of the formation the people the jobs that they will be doing within that organization compared to what has been the norm until now. Do I have lawyers? Who do I have on my marketing team? What are they doing? The engineers, we've already seen with software development, the role of engineers has changed dramatically. It's going to change across every different function. Startups will be at the forefront of this because they are coming with a white piece of paper. They don't have any baggage to change or anything like that. They can actually just design the company as they want and leverage all this AI for the future. I think that is going to be the biggest change. We will still see startups doing things. We're not going to be run by AI. It will be how does AI work with humans to achieve the goals that we want.

Samuel Boulanger

I love it. Thank you so much, Tom, for joining us on the show. It was very insightful. I think we have a good framework here for founders that are already leveraging AI or are just starting to think of building their company and they haven't jumped in yet. So, this was a great conversation. Thank you so much.

Tom Davis

It's an absolute pleasure. Thank you so much.

Samuel Boulanger

Tom's point about his own son building a startup with just two people using agents to do work that used to take a team is the clearest sign of where things are headed. The barrier isn't building anymore. It's knowing what's actually worth building. Three things I want you to walk away with from this episode. First, only build something yourself when it's tied to what you actually understand deeply. If your workflow is precise and you can articulate it clearly, building your own tool can be the right move. If it's not core to what makes your business different, buy something off the shelf and revisit it later. Second, distribution matters as much as the product. Tom pointed to Copilot as a way to reach enterprise users without asking them to learn a new interface. If your product can plug into where people already work half the buying decision is already working for you. Third, bake in security, compliance, and things like data residency from day one, even before you have enterprise customers. Retrofitting those requirements after a big deal is on the table is exactly when startups run out of time and money. If you want more conversations like this one, subscribe to the AI Frontier Playbook whenever you listen and sign up for the AI Frontier Playbook newsletter to stay sharp between episodes. Thank you so much for listening. I'll see you in the next one. See you.