Latest / Exit Algorithms / Lessons from 3 Exits and How to Implement AI in Your Business with Andrew Brooks (#76)
Transcript
- Pete Vera, Exit Algorithms: Welcome to Exit Algorithms, the podcast where we decode what it really takes to unlock growth, streamline operations, and prepare your business for a high-value exit. I'm your host, Pete Vera, and today I'm joined by Andrew Brooks. He's a Princeton educated entrepreneur who's successfully built and exited multiple companies. He's now the CEO of Contextual.io, a firm delivering purpose-built AI solutions that drive immediate ROI. Andrew, so excited to have you here. Welcome to the podcast. Andrew Brooks: Thank you. I'm ⁓ I'm thrilled to be on. I think it's gonna be a great conversation and ⁓ excited to share. Pete Vera, Exit Algorithms: Yeah, definitely. yeah, do you mind starting off? Can you share a bit about your your business journey, your background and and you know, ⁓ maybe what led you to contextual? Andrew Brooks: Yeah, no, that's ⁓ sounds good. It's compressing, you know, twenty ⁓ twenty-nine years or so of of work ⁓ pretty quickly, but I think we can get through it. So I ⁓ like you said, I I graduated from Princeton. I actually had a chemistry degree. ⁓ the joke that I like to share is that my senior thesis involved sulfur compounds and so I smelled like rotten eggs and a skunk for most of my senior year, and that was enough to teach me that I didn't actually want to be a bench chemist. So ⁓ this is the late nineties. Obviously internet is just starting to bubble a little bit. left and joined ⁓ what was then Anderson Consulting, which has become Accenture now, just because I felt like getting some broad consulting experience was going to be valuable. jumped ship ⁓ like a lot of folks did into more of the startup land. that was a company called Seven Space. ⁓ and we built and sold that in two thousand five to ⁓ what was then Sun Microsystems, which is now part of Orac Oracle. ⁓ and that gave me my first taste of being in, you know, the startup world. Like what what it what did it take to build a company? What did it take to ⁓ to to win clients. I I I had had an engineering role there, I had a sales engineering role there, I had an account executive role there. So I kind spanned the the gamut ⁓ of experiences and that gave me the taste and so in two thousand five reached out to ⁓ a friend of mine who I knew was leaving his ⁓ position who who was an entrepreneur at heart ⁓ and that started me on my journey. We we built ⁓ multiple companies over the last 20 years, ⁓ largely with the similar group or a same group. And that's actually one of the things that, you know, I would share and we should we should dig into. ⁓ but built a built a consulting company that turned into a a call center, which we ran for 20 years. ⁓ we built a a company called SMB Live, which was a small business marketing company, ⁓ and sold that to Reach Local, which at the time was ⁓ Google's largest ad reseller, took that public. ⁓ Pete Vera, Exit Algorithms: ⁓ interesting. Andrew Brooks: Left there, started a company called Smart Things, which is probably ⁓ one of the the best known ⁓ products that I've worked on. We built that company for a couple of years, sold it to Samsung. If you have a Samsung device, a TV, a phone, you've got smart things in your world right now. So ⁓ the little icons on your TV, you can you can play around with it. ⁓ and stayed there for a number of years, ⁓ you know, growing that globally. ⁓ did a little more consulting and then started contextual a few years ago and I'd say there's a through line, you know, if you like, ⁓ small business marketing and consumer smart home and now AI for business solutions, it looks a little scattered, but there is a through line, which is there's some sort of technology shift happening. There's an underserved market ⁓ or or a fragmented market, and we believe in building platforms and services that that ⁓ bring that technology to that underserved market. So so for SB Live, that was hyper small businesses trying to get online. for smart things, that was, you know, how are all these connected devices gonna stitch together and create a truly smart home? ⁓ and now with contextual, it's how ⁓ lower mid market primarily businesses are gonna are gonna benefit transform using AI. Pete Vera, Exit Algorithms: Got it. Wow, love it. Yeah, there's a lot of tangents we could go on. ⁓ but thank you for yeah, you kinda connect ⁓ a lot of the things in your career ⁓ that you know may not on the surface seem related, but they they actually are. ⁓ awesome. Yeah, and you you mentioned you it's a common team for a lot of your your businesses. How did that form up? Yeah, how did that happen? Andrew Brooks: Sure. Exactly. Yeah. Yeah. Yeah. So ⁓ there were three of us ⁓ back in ⁓ you know two thousand five who started ⁓ SMB Live ⁓ and we happened to right before we sold that we ⁓ integrated a company based out of Minneapolis that had three co-founders as well. ⁓ and then also had pulled in a a a chief product officer who I had known for the last you know eight, nine years in my work. And it was the seven of us who then went on to found ⁓ Smart Things. So ⁓ we were all players in the first company, all seven of us co-founded Smart Things. ⁓ and now ⁓ at least, you know, four of those co-founders are with me ⁓ at contextual as well. And so, you know, there there's a magic I think to knowing the people that you're working with. Of course, there's trust, there's there's ⁓ you know, experience. You can do hard things, you can have hard conversations. You know, you know, when you're when you're friends at the core, you can you you know, you're not gonna offend people. ⁓ they know that everything's coming from the right heart. But I think, you know, magically it's everybody knows their role, everybody knows what they're responsible for. It's not, you know, there's not there's no ambiguity as to, ⁓ that's you, this is me, I do this, you do that. ⁓ and that's super helpful with the speed that is required from a startup standpoint. You can't be you can't be spinning your wheels. Pete Vera, Exit Algorithms: Yeah, definitely. ⁓ that's a that's a huge benefit. It probably makes it a more fun ⁓ work environment, I'd imagine too. Andrew Brooks: Yeah. I mean these are people that, you know, when you work with somebody for twenty years and you've seen their children grow up and they've gone through life milestones, you've gone through life milestones. ⁓ you you've you know, you know, you've done fun things together as well. A lot of these folks we we took a trip to Patagonia and did a bunch of hiking together. You know, when you break out of that work experience and and live life together, ⁓ I think it it tightens the bond that is that's pretty critical in a startup world. Pete Vera, Exit Algorithms: Yeah, definitely. That's ⁓ sounds like a great culture as well you you develop over time. Yeah, I'd love to learn, you know, y you've had two well basically now three ⁓ exits, right? Or partial exits at least. ⁓ yeah, what were some of those the lessons that you you've learned from those? Did did the did you do some things maybe the first exit that you course corrected on on the second one? I'd I'd love to hear ⁓ you know, some of your your stories around that. Andrew Brooks: Yeah, each each one has been a little bit unique. so SMB Live we we bootstrapped, we self-funded that. We actually were doing a bunch of consulting while we built the software platform. ⁓ and so that was you know, ⁓ s assemble your own desk with your own screwdriver, get on any plane you need to, you know, make it happen. And ⁓ and that worked with SMB Live and and and that meant we had a very clean experience in transacting that to to ⁓ reach local. We all took ⁓ you know, very specific ⁓ roles within the new company and understood our job. ⁓ but we were we were kind of dispersed within the company from that perspective. and so I I I liked that at some level. We weren't this kind of isolated thing. We were we were truly integrated into the company broadly. ⁓ with with smart things, it was a slightly different experience in that ⁓ first of all, it was a venture-backed. ⁓ Entity. ⁓ you know, we built hardware, we built software, we built cloud, we built mobile mobile apps, we had to do a lot. And so we had ⁓ you know, venture venture stakeholders. And that changes, you know, the the size of an outcome that you have to seek, you know, fundamentally. ⁓ you know, fortunately for us, ⁓ the timing was right, ⁓ and Samsung was a great strategic acquirer. You know, the difference there was we ended up as a as a standalone unit, a standalone entity. And there was there was great value to that in that we ⁓ Pete Vera, Exit Algorithms: Mm. Andrew Brooks: You know, we were able to preserve the brand. ⁓ that brand, you know, continues to exist to this day. It's it makes us, you know, all of us who are part of it very, very proud. But but there's a tension when you're a separate entity in an acquirer, right? ⁓ attention for resources, attention for ⁓ priorities. you know, the mobile division and the TV division within Samsung are enormous entities and they can push and pull and move you around. ⁓ and so I think one of the pieces that I would draw out. Pete Vera, Exit Algorithms: Yeah. Andrew Brooks: for ⁓ your listeners between those two is make sure you understand after the transaction is done where will you land and how long will that be? Are you integrated into the company or are you a standalone entity? What control or authority or permissions do you have if you know a standalone entity that those all really matter to kind of the success or challenges that you might ⁓ you might achieve. ⁓ This didn't affect us, but certainly I I know of colleagues who've gone through ⁓ transactions where i in in if you have specific earnout obligations or opportunities, if you don't fully control your ability to hit those ⁓ obligations, you you're putting yourself at a risky situation. And so as you think about how am I being integrated into this organization, how is my team, how is my product, how is my service, whatever it is being integrated, you have to look at, yeah, what what is that experience going to be like and do I have the control necessary to ⁓ ⁓ you know to to to achieve my destiny. ⁓ as you said, the the third transaction ⁓ we completed just last September. We sold a majority of contextual to ⁓ a mid-market private equity company called Southfield Capital. They're based in Greenwich, Connecticut. they're great partners, ⁓ but it's a different thing in that the transaction is not the it's not the end of the race, it's the beginning of the race, right? Our our job now is to is to bait to build a really ⁓ meaningful and important company. Pete Vera, Exit Algorithms: Yeah. Andrew Brooks: in this space with their support and backing. Private equity companies call themselves your sponsor. And that's true. They're sponsoring you towards your ⁓ towards your growth objectives. And so, you know, each one ⁓ is a is kind of a unique ⁓ a unique step in time. Pete Vera, Exit Algorithms: Definitely. Yeah. How how was it managing the you know the integration ⁓ post close? I I mean ⁓ now you've done it three three times. ⁓ was there any difference between, you know, a huge ⁓ fortune fifty company like Samsung or versus ⁓ re you know, reach locals not at n as big, right, at the time anyways? Andrew Brooks: Yeah. Yeah. Yeah. Yeah. No, absolutely. And and ⁓ you know, the Reach Local team was entrepreneurs at heart as well. The entire leadership team, ⁓ from the CEO down wa were were entrepreneurs. And so ⁓ there was a great ⁓ kind of cultural meld there. ⁓ and they were US based, right? Their headquarters was in LA, ⁓ their sales headquarters was in Dallas. ⁓ it was a lot of travel back and forth between, you know, Connecticut where I was living at the time and LA. But ⁓ But ⁓ but because of that entrepreneurial spirit, I think it the integration was quite easy. The support and vision, they were looking to grow. the storytelling required for their S1 filing for the for going public, we were a core part of. And so ⁓ that integration was easy. I think the integration with with Samsung was a little more difficult. First of all, flying to Seoul, Korea is not ⁓ South Korea is not the the easiest flight to ⁓ to take. They did require every the leadership team to relocate. to Palo Alto or the Bay Area in California, they wanted the center of gravity to be there. So it was a little bit life disruptive, honestly. You know, I had young kids at the time and you're you're moving them around. and ⁓ and then again that just the size of some of their entities, the mobile division, the consumer electronics division, ⁓ you know, they can they can have a they they have high expectations and you might not have the same resources that they do to achieve things. So I think the integration was ⁓ successful but definitely took a a a lot more ⁓ Pete Vera, Exit Algorithms: No. Andrew Brooks: energy and then in the in the private equity situation you you remain a standalone entity. The C Corp ex persists on the other side. ⁓ you're just there now to ⁓ you know to drive growth. There's resources, there's support, there's ⁓ access that otherwise you wouldn't have and so no real integration there other than maybe some maturation in terms of reporting and you know your financials and things like that that you know the LPs of a private equity firm are going to expect. And that's reasonable. Pete Vera, Exit Algorithms: Mm. Yeah. God yeah, that makes a lot of sense. Each one was a yeah, completely different scenario. Andrew Brooks: Very different. Yeah, very different for sure. Pete Vera, Exit Algorithms: Were there some things that were consistent as far as, you know, preparing your business for a for a sale, whether that's, you know, cl financials or yeah, what what do you what do you do to prepare for a successful exit? Andrew Brooks: Yeah. Yeah, this is this is gonna fall into the do as I say, not as I do, ⁓ because I haven't gotten it perfectly ⁓ any time. Every time I start a company, I promise myself, be much more ticked and tied with all of your paperwork, all of your documents, every vendor contract, every consultant contract. Have it in ⁓ ha have maturity around all of your corporate documents, all of your financials, like have maturity around that from day one. you will be going through a scramble during a ⁓ during a a a diligence period and during a transaction period to find all of those materials dating back years, right? and remember, you know, the the the lawyers on the other side of the transaction are are paid to reduce risk and risk creeps in when you don't have, you know, hey, you had a consultant working for you, but you don't you can't track down the signed version of the ⁓ of the ⁓ of the consulting agreement, you know, that's that's a risk. And so ⁓ you know, do yourself a favor from day one, just be buttoned up from a from a paperwork standpoint. ⁓ and again, that's you know, all of your vendor contracts, all of your supplier contracts, all of your consultants, all of your ⁓ you know, every every piece of paper that you could think. The second thing I would put in there is you know, get get on the right side of employment contracts, employment agreements, ⁓ you know, IP assignments, confidentiality ⁓ documentation. With modern tools, you know, we use ⁓ we use rippling ⁓ today with modern tools. ⁓ that shouldn't be difficult to do. But sometimes you start small, maybe you've got you know somebody just doing your books out of out of and and payroll out of QuickBooks, and you don't necessarily have all of that. It's just gonna add burden to an already stressful time. And so, you know, take that burden away ⁓ is one. ⁓ certainly ⁓ depending on the size of your transaction. having a an attorney on your side who is gonna see you through ⁓ that process is is worth is worth it, right? I mean if you if you're selling something for, you know, under a million dollars, maybe that starts to be concerning. But anything over that size, you know, having having an attorney on your side that ⁓ where where you can negotiate with some of these folks to be like, hey, it's a fixed fee to get us through this transaction, so you know what it's gonna be. ⁓ that just you just need that that partner. And that and that fear reduction, right? What what am I signing up for? What are what is this exposing me to? What ⁓ how do I understand this? How do I think about this for my family? You know, those are all important considerations. Pete Vera, Exit Algorithms: Yeah. Yeah. Love it. Well said. No, th thank for that's a really good really good insights there. ⁓ yeah, I'd love to talk a little bit more too about contextual dotio and maybe what is a you know, typical engagement look like? What are the first things you look at with a with a client that you on board? Andrew Brooks: Yeah. So we we focus on AI business systems and and what I mean w the reason I I draw that out is, you know, there are plenty of assistant tools, ⁓ chat tools, et cetera. Everybody should love co-work or or the equivalent on on open AI. ⁓ but those are assistants, they're kind of individual ⁓ enablers and and maybe you have an agent or two that does something in the background. So we focus on what we would consider a an AI business system that tends to cross people, it tends to cross underlying technologies, it it it ingests multiple data sources. And they're often back office operational, you know, what I would say, you know, unsexy in terms of ⁓ the AI pieces, but they're how businesses get work done. And especially in the lower mid-market, one of the things we flag is, you know, the these folks are going to go through a leapfrog moment, which is historically they they couldn't invest in purpose-built technology. They couldn't invest in software that they designed, built for their own their own selves. They were kind of limited to some ERP systems and a CRM and, you know, ⁓ your Microsoft Suite, et cetera. And with AI assisted development, because it's gotten faster and less expensive, as well as AI as a new tool, they can suddenly invest in in solutions that are bespoke effective effectively, what we would say purpose built to their ⁓ particular operation. And so we focus there, ⁓ largely midmark lower mid market companies, so you twenty five million, maybe up to Pete Vera, Exit Algorithms: Mm-hmm. Mm. Andrew Brooks: ⁓ a billion dollars in revenue, ⁓ who again historically have been ⁓ you know haven't had these technologies and now can leapfrog into ⁓ you know this new capability set. just to make that a little bit more practical, ⁓ you know, it's very common in the lower mid market that stuff's being done via email. Vendors are sending in invoices and contracts and things like that. ⁓ reading an inbox, attach you know, extracting attachments, deciphering that attachment. ⁓ extracting information from that attachment, enriching it, preparing it, sending it along. That's a great AI ⁓ tool and and solution. We we we see those sorts of automations all the time. And and oftentimes our clients think in automation as like, ⁓ AI is for automation. And that is very true. It's it's one of the major buckets that we work in. But increasingly a lot of our clients are thinking about, well, wait a minute, we need we can look at our data in a different way. We can look around the corner. Pete Vera, Exit Algorithms: Mm. Andrew Brooks: a little bit more proactively. We can enrich our data with outside information that otherwise we wouldn't have had. ⁓ and then all the way through, what if we brought a new product to market for our ⁓ for our customers, right? An entirely new revenue stream. All of that is kind of on the table right now. And you know, our argument would be, especially ⁓ in the businesses that we work with, you can only cut costs so much. You've got to find growth. And and AI can be part of both of those. Pete Vera, Exit Algorithms: Yeah, love it. Yeah, it ⁓ makes a lot of sense. What are some ⁓ you know, some of the most common mistakes you see business businesses make when they're implementing AI? Andrew Brooks: You know, i it's interesting. Twenty twenty five, twenty twenty four was very much you know, what's happening. The the models were still making f silly mistakes. It doesn't know how many R's are in strawberry, you know, all those sorts of things. ⁓ hallucinations were real. and so I think that created a little bit of a ⁓ a fear ⁓ in in in users. So twenty twenty five becomes a little bit of the kind of proof of concept world and time. Twenty twenty six is definitely the time to be implementing, seeking to deploy into production. And I think so one mistake would be waiting still. ⁓ and that that waiting could be caused again by a little bit of the hangover from you know ⁓ early stage AI that wasn't as dependable. ⁓ that waiting is also potentially because, gosh, it's it's insane in the market. There's a new an announcement every single day. It feels like Claude releases this, open AI does this, you know, Gemini does this. And so that causes a little bit of paralysis. It's too much information for most people to kind of really process. So number one would be waiting. You've got to you've got to dive in. And the way to dive in is to choose a very practical application to think about how change management's gonna work within your organization. Like how are people going to adopt this? What does this really mean for their jobs and communicate early, often and aggressively? Find ⁓ your cheerleaders in inside and don't do you know, in our recommendation, don't do blanket cutovers where you're like, okay, we're just gonna turn the system fully on. These systems can be staged, humans can be in the loop, humans can be in the lead, humans can be in command to help them get better, to ⁓ you know, observe you know unexpected patterns and ⁓ and ⁓ and you you then grow it ⁓ from there. So I would say waiting is a mistake. ⁓ choosing to go too big that you can find very small practical applications is a mistake. Pete Vera, Exit Algorithms: Mm. Andrew Brooks: Believing that your messy data is is gonna prevent you from succeeding. I'll give you a very practical example of that. ⁓ one of our clients falls into this bucket of of ⁓ receiving invoices from 2,500 different vendors, something like 14,000 invoices a month. These are all shapes and sizes and formats. ⁓ they wanna process those invoices and work orders, but more than just kind of line item extraction. They to categorize the work, what was done, why was it done, all that sort of stuff. Well. ⁓ the AI does a great job of extracting like, hey, this is the vendor name and address from this invoice. And then sometimes it can't find it in their core data set because somebody, you know, fat keyed it or entered it differently or something like that. So that's a great example of messy data. But actually what you get in that situation is if you've built the solution right, an override, an escalation where they where a human can fix the data. And so it can almost become a a data cleansing. Pete Vera, Exit Algorithms: Yeah. Andrew Brooks: mechanism for you. You you you know, you take the burden of the of the heavily manual work away. The humans can now be thoughtful around well, what what's happening here? How do I want to clean this data? How do I how do I want to modify this? So ⁓ you know we see that as a as as one of the best ways to ⁓ to address that messy data ⁓ situation. you know we have an importer where we'll get Pete Vera, Exit Algorithms: Yeah. Andrew Brooks: ⁓ you know, ⁓ records of of the manufacturer from China and it's it's chaos that data, right? Like it's it's it's there could be a gazillion different companies in the same street. ⁓ so how do you use AI to rationalize that? Well we can take a lot of their different records out of their core database, pump all of that into an L O ⁓ and say, What what is that? Which business is this? And it can look for pattern matching of course, it can look online and see if it can figure out from there. It can sort into other documents to try to figure it out. ⁓ Pete Vera, Exit Algorithms: Yeah. Andrew Brooks: You know, so so there's ways to use the AI to to get over that that data burden. Pete Vera, Exit Algorithms: Yeah. Wow, l love that answer. ⁓ yeah, and ⁓ my next question was gonna be around, you know, how do you how do you choose the the right tools? ⁓ when there's so much out there, you know, there there's a lot of yeah, ⁓ I I'm curious. You you but you touched on it a little bit, it sounds like starting small, ⁓ Andrew Brooks: Yeah. Yeah, I mean our so we have a a phrase that we use, ⁓ we've actually filed a trademark on it called own your AI. And what that means is not necessarily that we're saying, hey, you should own your own model or you know, build your own model. What what it means is don't ⁓ hit your wagon to a single hyperscaler or a single tool or a single provider. ⁓ we provide what would classically be an AI orchestration layer. What that means is we build our solutions on top of contextual our platform. And we have an entire infrastructure to determine which model to use, which which route, ⁓ which which which hyperscaler, which specific model. And that means it's very easy for us when, ⁓ gosh, ⁓ Gemini just released an update or Google just released an update that does that improved ⁓ you know PDF extraction by one percent or something like that. We just toggle a switch, right? It just we switch it. And so I think the fear of lock in. is real and should be acknowledged. There are there are per you know partners like Contextual who acknowledged that early on and said, let's let's make sure you own this. You can decide which models to use. You can switch models. You can stop using a model if you want to. ⁓ you know, we'll show show you very granularly within every single call what you're paying, what and what you might pay if you chose a different model. You know, there's ways to to get over that discomfort. but it's key that you own your business process, you own your data, you own your underlying integrations with your systems, you own potentially new interfaces that you're creating for humans to use. ⁓ you know, that that ownership helps ⁓ you know, lock you into the real value and the real the real moat that you're creating with AI versus ⁓ you know, what everybody else can get access to. Pete Vera, Exit Algorithms: Yeah. ⁓ love yeah, that's that's great advice. ⁓ yeah, and and you know, you've been in the tech space a long time, seen a lot of disruptions, a lot of tech shifts. How does AI, in your opinion, compare to other ones you've experienced, like, you know, the internet or or Andrew Brooks: I mean, certainly there's plenty of data out there about the, you know, pace of ⁓ you know, pace of adoption, right? We we love throwing a slide up like how long did it take people to adopt railroads and electricity and then you know computing and then mobile and now and the internet and now AI and of course the the pace of adoption is is insane, right? ⁓ and so so number one that. Number two, it's certainly one of the first technologies where I would say if if everyone in your organization is not fully embracing it, you are ⁓ regardless of role. you're making a mistake. ⁓ and I'm sure somebody could counter hey that does the machinist down on the shop floor ⁓ need to to use AI? I guarantee I can come up with a reason that they they should be embracing AI. so I I think that's one difference is ⁓ obviously pace of adoption, the need for everyone to be an adopter. And as a result, the kind of need for everyone to be a little bit of a system thinker ⁓ and a and a ⁓ a different because You know, when you're just using an interface that's a ⁓ you know, a a form against a a a a stack of software, ⁓ you just know, ⁓ I put ⁓ you know, data in A and and B spits out. with AI, there there might be a different interface. It might be a chat interface, might not be a chat interface, but you must understand kind of what is the system doing and understand what that system is doing. Cause you you now it this isn't fixed software logic on the backside of a bunch of if then statements. You have a thought partner that's engaging with you. ⁓ in the form of AI. And so think you know, understanding how it how it works at at the system level, I I think is is really really important. I also think we're about to go through ⁓ Microsoft just announced that they're pulling back some cloud cowork licenses, ⁓ or a code ⁓ cloud code licenses. I think we're about to go through a little bit of a an economic rationalization. ⁓ you know, I was using some of our own AI tools this weekend and just was really into it and go, go, go. And next thing I know I'd I'd racked up Pete Vera, Exit Algorithms: Mm-hmm. Andrew Brooks: you know, a few hundred dollars of ⁓ of overage charges, right? And so, oops, well, okay, that's, you know, fine. It was only a couple hundred dollars and, you know, we're a small company, but if you're a ten thousand person company and that sort of stuff is happening, it can it can be pretty meaningful pretty fast. So I think there's gonna be some cost rationalization and observance here happening soon. Pete Vera, Exit Algorithms: Credit. Yeah. And what do you what would that look like? I know it's hard to predict, but just using less or di or new new tools coming about as a result or Andrew Brooks: Yeah. I I think it it's gonna look like a few things. Number one is obviously people have to pay close attention to, you know, what ⁓ what is the relationship they have with those providers? Are they corporate accounts? What what tier are people on? People need to be a little bit more ⁓ you know, sensitive to the model they were using. I was I was, you know, at Opus four six and just blasted. I was going at it. And so, you know, that has a a specific token cost. ⁓ so you gotta be set so there's gonna be training around what does that mean? What's the right model for the right task? What's the right tool for the right task? I think you might see clients ⁓ or or users starting to put their own interface in front of these tools in a way that they can get better control and better observability ⁓ as to what's happening and why and why it's being used ⁓ you know in that way. so you know, data's your friend there and and sometimes the the hyperscalers don't wanna provide a lot of data because they're the they recognize that the data can help you manage that your ⁓ sp your spend. But we're we're coming to a rationalization there. You can't have just people blasting into, you know, thousands and thousands of dollars of unexpected charges every single month. That's there's no way to plan around that and that's not gonna be acceptable. Pete Vera, Exit Algorithms: Yeah. No, love it. Yeah. It it's fascinating where where it's going and the time we're we're at. ⁓ man, ⁓ well I I'd be remiss if I didn't ask about ⁓ time management. I wanted to bring that up. ⁓ how do you manage your your day? You know, you you get so much done, you've well s with business, I saw you you're an author, you're also athlete. ⁓ yeah, how do you structure your your day to maximize productivity? Andrew Brooks: For sure. Yeah, ⁓ well, ⁓ s certainly the the published book is not a great book, so I wouldn't encourage anybody to go out and ⁓ find it, but it was a a life milestone and ⁓ co-wrote that with my dad, which was fun. ⁓ if you if you want a book, go go get my daughter's ⁓ children's book. That's a good one. ⁓ no, so I you know, I have gone through peaks and valleys of starting companies, ⁓ doing ultra races. ⁓ you know, it's it's an energy allocation question more than anything. I I don't believe strongly in Pete Vera, Exit Algorithms: Mm. Andrew Brooks: ⁓ super fixed schedules because if you have a super fixed schedule and then you break it, ⁓ it can be very frustrating. So as an example, ⁓ you like I don't say, ⁓ I, you know, I I will always be training at nine AM or eight AM because sometimes you wake up and work needs you or an event needs you or your family needs you, right? So I think for me it's a little less about ⁓ super fixed time schedules. I know there's people who like swear by the up at four thirty, meditate for thirty, that's just not me. ⁓ I think it's more about goal management, right? And there's a you know, the classic view of ⁓ you know, you've probably seen this where there's a container and you got big rocks and little rocks and smaller rocks and sand and the only way to get it all in the container is you gotta start with the big rocks. And so, you know, if I have a a race that I'm training for, that's one of the big rocks on the day. And if that means I need forty five minutes to go out and run, you'd find that, you know, in the calendar. if I if I need if I'm Pete Vera, Exit Algorithms: Mm, yeah. Andrew Brooks: starting a business, maybe the racing goes to the side. I haven't actually done a a a race in a while. I'm just ⁓ starting to gear myself back up to do another one. ⁓ and that's okay, right? Like, you that's you you can do those trade-offs. I would say what is interesting, especially as it relates to technology and tools, is how do you manage the interrupt driven world of a cowork, Claude Cowork as an example. You set it off on a task and it might take a few minutes to do that task. And you get distracted doing something else and then you come back and you send it another one. There's a little bit of this Pete Vera, Exit Algorithms: Mm-hmm. Andrew Brooks: this hyper task switching. You have this amazing assistant sitting next to you, but now you have to like intellectually task switch back to it, you know, repeatedly. And I I don't have great advice there other than don't not use it because of it. But you you you think about don't think about this necessarily as like g identify the project and give yourself a timeline to get that project done even if you're stepping away and doing something else while it's doing its thinking and it's stewing and stuff like that. ⁓ or If it's not critical, if it doesn't need to be done in the next hour, set it up in the morning, fire it off, come back to it at the end of the day. you know, d use the use the tool wisely. I do get fearful of interrupt driven worlds. Of course, AI is not new there. ⁓ anybody with Slack, you know, or a f a cell phone understands that how damaging interrupt driven ⁓ worlds can be. Pete Vera, Exit Algorithms: Yeah, yeah, definitely. No, you said a lot of good things there. ⁓ I know, I love the focus on energy rather than necessarily the time structured time. ⁓ any calendar can work, you know, it's a matter of f choosing what you're focusing on, right? Andrew Brooks: That's right. That's right. Pete Vera, Exit Algorithms: And the yeah, the the attention residue concept too of switching tasks. There's a lot of studies on that. ⁓ I'm sure you've probably read Deep ⁓ Deep Work, Cal Newport's book. Andrew Brooks: Yeah. Yeah. And I think AI is an interesting deviation from that. I you know, I'll have three cowork sessions doing three different things. One's a financial analysis and one's a marketing analysis. And it feels very productive, but I you we have to pause and be like, Are ⁓ are you losing fidelity through, you know, that that task switching? And so I think that's a real thing to focus on for people. Pete Vera, Exit Algorithms: Yeah, definitely. Well great, Andrew. I have one last question for you. What is one practical tip for business owners who want to implement AI and actually see results now? Andrew Brooks: Yeah, I mean the the we obviously build you know more complicated systems. They take weeks to to develop and and deploy. So so I the the number one thing is has to be if you're not if you're not using long running assistance like cloud cowork or like GPT, where your documents are in there, ⁓ you've given it a lot of your thinking and your thoughts and how you approach it and you're dividing it based on the function that it is. This is my marketing assistant, this is my financial assistant, this is my sales assistant. If you're not doing that, you have to be doing that. Because it's like having an an additional employee, an addition, you know, a a business analyst ⁓ from McKinsey on your side for every single one of those, right? And you manage it though. Like don't let your assistants then write you 15 pages of of AI slop that you're not going to read anymore. Like focus on what the output is that you actually you actually really want. And so I think if if people aren't doing that today and I speak at events and I ask people to throw their hands in the air and they're not necessarily doing that. Y that's gonna be the biggest, most immediate impact. And for a twenty dollar license, you you you need that impact in your life. You have to be doing that. Pete Vera, Exit Algorithms: Well said. And it's been a a great conversation, Andrew. ⁓ loved hearing your your insights today. Where can listeners find you and learn more? Andrew Brooks: ⁓ certainly on on LinkedIn, obviously, Andrew Carol Brooks. And then ⁓ you know can you can always ⁓ check us out at contextual.io and reach out. I'm always happy to to share ⁓ with folks who really want to understand real life use cases, practical examples, ⁓ what we've learned doing this. We're we're pretty we're pretty open in in sharing our journey as well. Pete Vera, Exit Algorithms: Awesome. I'll leave that in the show notes. Th thanks for joining the show today, Andrew. It's great having you. Andrew Brooks: All right. Thank you, Pete. Appreciate it.