How to Align Marketing and Finance and Fix the Plumbing Problem with Alex Brower

Welcome to the Future Finance Show, where we talk about treasury management on bars. Forecasting isn't broken, but the plumbing is. What do you mean by that?

There is a tendency for sure to just get all the data into one place. A headline would be just because you've gotten all of your data into one place doesn't mean you've gotten teams on the same page. Today we are joined by Alex Brower, co-founder and CEO of Qflow.ai, also a limited partner at Fog Ventures. Previously led finance, operations, and marketing teams at high-growth tech firms, Alex brings a unique perspective on bridging strategy and data. If you're running agents or if you're relying on AI to fix a vibe-coded app and halfway through the month you're out of tokens, that's a problem. We built Qflow to change how teams approach fixing the data, planning the growth, running the business, and telling the story.

For a CFO or an FP&A leader that's trying right now to modernize their planning stack in the midst of all this noise, where do they start? What do they need to look at with their tech stack, and what should they stop doing first before they make any technology decision?

Good question.

I think that Future Finance is brought to you by Qflow.ai, the strategic finance platform solving the toughest part of planning and analysis. B2B revenue aligns sales, marketing, and finance seamlessly. Speed up decision-making and lock in accountability with Qflow.ai. Welcome to another episode of Future Finance. I'm Paul Barnhurst, your host, and today kind of a clown. I can't get things right, as Glen will confirm, but we're going to try to do a great interview today. So how you doing, Glen?

I'm doing great. And I was going to say, how is today any different than any other day?

Shots already, and we haven't even got 30 seconds in. Well played, Glen. You know, not all of us can be famous like you, big AI guy and all. Speaking of famous people, we got a great guest today, but before I give it away, I'm gonna let Glenn go ahead and introduce our illustrious guest.

Today we're joined by our good friend, Mr. Alex Brower, co-founder and CEO of Qflow.ai that listeners to this show will be familiar with. Qflow is a company redefining how finance and go-to-market teams plan, measure, and align. Backed by Precursor Ventures and K5 Tokyo Black, Qflow helps growth-focused companies connect their data, improve forecast accuracy, and increase analyst productivity by up to 5 times. Before founding Qflow, Alex led finance, operations, and marketing teams at high-growth tech firms including Apptelligent, which was acquired by VMware, and Cloud Academy, which was acquired by QA. Alex is known for bridging strategy, performance, and data, and for his belief that the real power of technology is in aligning people around shared definitions of success. How's that for my radio voice? That's pretty good. Alex, how you doing? I'm good. We're happy to have you on the show. We've been trying to plan this for a while, so we're glad to finally have you on.

I'm happy to be finally on. I appreciate you guys and for the FP&A and finance community more broadly, and it's an honor to be here.

We talk all the time and you and I have, we've done a little fire chat session before. So this is, I'm going to dredge up some of the stuff we've talked about. I'm going to lead off today with, it's probably this first question I asked you when, when we did our last chat. I love the shape of your journey. So if you could give us kind of a, a quick look at going back from leading finance to biz ops, to marketing, and then ultimately to founding Qflow. Was there a consistent thread or what problem were you trying to solve when you started Qflow?

Hopefully my answer doesn't vary too much from the last time I answered. I don't think it will. I came up with PE background operators at a startup in Chicago. I was on the finance team and when I left, I was running the finance team. And this was a startup that went from $200K EBITDA to $30 million EBITDA. In under a year. So operationally, it was challenging, but I was focused on building the models. We owned the forecast. We lived the pain of board prep. A couple years later, I moved out to San Francisco, you know, with the hopes of doing it again. And in this case, it was, you know, I joined a tech company. It was a self-service meets enterprise motion. I had RevOps reporting to me. We had to throw analysts at hygiene issues. Every cycle to make things like the, you know, MRR or ARR schedule tie out. It was a tax we paid each month. And after that company sold to VMware, an investor asked me to run marketing at one of his portcos. Part of what I had done previously was restructuring a marketing team and it was fun. So I said yes. And I then experienced, I wouldn't call it the forecast fight, the forecast tension from the other chair. Right? Same data, different numbers, hard work aligning people. That's really the arc. Every company I have been at, I was building the same Frankenstein, right? Across a data warehouse, Salesforce or HubSpot exports, ERP pull, billing reconciliation, and a lot of spreadsheets. So we built Qflow to change how teams approach fixing the data, planning the growth, running the business, and telling the story.

Isn't it amazing? It seems the more businesses you talk to, it's like everybody within a, you know, certain size of business, they're all dealing with the exact same issue. So I can see the desire to productize connecting all that and bringing and taking the time out of, well, it's what everybody's trying to solve with AI, with probabilistic results right now. And so that makes complete sense.

Yeah, Glenn, I was going to say, you said it much nicer than I did. I was going to say the same level of dysfunction.

In all the companies.

Dysfunction is opportunity, right?

It's all what side of the coin you decide to look at, right? Thank you for, uh, that kind of first answer in your journey. You know, that friction between marketing and finance is an interesting one. I've had some guests on and talking about that. And now having run my own business, and I imagine even Glenn some, I have a very different perspective. Like I would work differently with marketing. So I can see how having that help led to what you're doing, because being on both sides of that table gives you a very different perspective than just being on the finance side or just the marketing side.

You know, Paul, my first job out of business school was in marketing and not finance. So I've always had a soft spot for marketing as well.

Well, we have 3 people that know a little bit about marketing. We're almost enough to be dangerous.

Dangerous. Yeah.

3 finance people, I should say. That's a scary thought. All right. You've mentioned several times, you kind of have a saying where you said forecasting isn't broken, but the plumbing is. So I'd love to first kind of dive into that. What do you mean by that? And what does good data plumbing look like? You know, obviously you're trying to connect the CRM, the ERP, you got your billing system, right? There's never just one source of truth. There's multiple different tools out there. So maybe start by talking a little bit about that.

There is a tendency for sure to just get all the data into one place. And I have thoughts on that, but I think the headline would be just because you've gotten all of your data into one place doesn't mean you've gotten teams on the same page. Typically, we think about hygiene issues and plumbing issues in a couple different categories. There's the categories of things that you can address and fix And then there's the stuff that you have to live with and manage. And so examples of stuff that we typically see getting in the way between an organization and a top-notch, you know, quarterly or rolling 36-month forecast are, you know, things that obviously impact the top-line bookings forecast. Close dates in the past, garbage data entered by sales reps who are gaming overly. Aggressive CRM validation requirements. Then we see stuff that finance has contributed to, and this usually is because they can't trust the data out of the CRM, but they maybe lack the willpower to do something about it. You usually see this in the form of like an ARR waterfall that's like this massive file that gets updated every month. And that's because the customer list coming out of the CRM doesn't match the customer list in the billing system, which has drifted from the customer list in the ERP. And once we start addressing those, we see that there are missing transactions in the CRM or in the ERP, or hopefully not in the billing system, but it's happened, right? Probably biggest thing that companies can fix it, fix, and this is perhaps not new news, is around the data dictionary. So these are data dictionaries that may not exist. They may have been updated a long time ago. And when you have, when you lack a data dictionary across teams, every team's true number disagrees with every other team's true number because no one agreed on what a dollar of ARR means or what actual a qualified opportunity looks like. Or how we think about customer hierarchy. So those are like 3 categories of hygiene issues that I think are immediately addressable. The fourth is a category of issues that you have to live with. So many of our customers have grown through acquisition and they have no intention of consolidating their ERPs or their CRMs. Maybe they bought a B2B2C company. They have a couple B2B acquisitions that run vastly different go-to-market motions. So they're running in 4 different CRMs. And in that case, a data warehouse actually co-locates that complexity. It doesn't reconcile it. So that's what I mean by getting the data in one place. Doesn't mean you've gotten people on the same page. These, these companies live with sort of a bidirectional pull. They still need the fine-tuned models, the predictions, the business logic per entity, per segment within that entity. But they also owe the board one aggregate forecast. So it's a tension to manage and not a problem to solve.

You know, as I listen to you say all that, I was like, has he been sitting in, in the companies I worked for? Yeah, I can still remember the conversation with senior leadership. We were trying to define what a booking was as we moved toward a SaaS model. Yeah, we had the CEO, the CFO, the CRO all on the phone, and I was kind of leading that because my job was helping us make the transition. And so it was an interesting conversation. Or when you said, you know, missing billings, I remember I had to go through and dig through all our billing stuff. We had Someone we were giving a $10,000 discount for 2 years that was only supposed to be a 1-month discount and nobody turned it off. Like just all kinds of stuff where you're just like, really? There was, 'cause there was zero connection. I mean, it was all broken. So I can really relate to that. What do you say of this idea now of, well, as long as you can hook it up to an MCP connector, AI can help you figure it all out. What would you say to that?

Yeah, I think that, uh, What would I say? I would say MCPs are powerful. They're a great transport layer, but they're not a semantic layer. So we see, you know, MCPs being used sort of, again, like on two different ends of the spectrum. On one hand, like the one-offs. So I'll pick on like the HubSpot OpenAI plugin, you know. Great for querying the status of an open deal. Pretty good at, you know, updating, you know, updating a record, but it's not going to give you a nuanced view of your win rate by segment or tell you why your renewal motion is slipping or thoughtfully assemble, you know, a low, medium, and high growth plan for potential investors. So that's sort of like on the, on the one-off side. On the other hand, it's full-on vibe coding, right? A CFO asks marketing for conversion analysis. Someone's got, you know, Claude Desktop or Claude Code hooked up, uh, both powerful tools. Um, and someone vibe codes like a complex attribution conversion analysis in Node or Python. You know, it works for like the next QBR. But it breaks the next time a new definition or a new activity type is introduced. And it's difficult to maintain because that, you know, Mark, the operations person doesn't actually know Python or Node. So I'd say right now the MCP on its own coupled with a frontier model does not get you to the durable analytical layer that you need. Some of these things we're talking about, like segmented win rate, multi-entity waterfall, or even variance commentary that holds up at the board level. So I think that's the gap. MCP is the pipe. You still need the water treatment plant. You need the semantic alignment, the business model reconciliation, and that has to happen before any insight is generated.

Now, that's what we keep hearing, that semantic layer. I was going to say one thing and then I'll let you go, Glenn. I'd love to get both your thoughts on this. So as I listen, right, we hear a ton about vibe coding. And I've been thinking a lot about this and it feels a little bit to me like the Excel wizard or the VBA person, but it's available to everybody. What always happened when somebody did a bunch of stuff in VBA or they did complex Excel models and they turned it over to the next person? What happened 80% of the time? The next person scrapped it and started over. Are we just duplicating that now that it's much easier? There's lower friction with VIBE coding. Not to say there aren't some great solutions and there aren't some great Excel files to get transferred. And some good work that can be done, but it feels a little bit like we're repeating it just with new technology. Am I off on that? I'd love your guys' thoughts. Then I'll let you go, Glenn. I'm just kind of curious, Alex, your take on that, because that's kind of how I feel. And I like it. I think people will use it, but I worry we're going to see a similar type of kind of sprawl and problems that we've seen with the expert in Excel or the VBA, just in a different way.

Yeah, I think it's an interesting analogy. I think that it's a very powerful tool if you're looking at it, if you're actually reading what it's doing, if you're testing it. Basically, it's a very powerful tool for someone who probably already could write the code. And I think it's a cool tool for people who can, but I do think that You know, there's another aspect to this that we were chatting about before the show. Yeah, the vibe coding can get expensive fast. And if you're vibe coding basically a mini application for every aspect of financial analysis or go-to-market analysis, you're going to end up like some of the companies that I spoke to earlier this week. Um, one company we talked to now has a new policy where it's op— it's not, uh, opt out. You have to ask for permission to get access to the corporate Anthropic account because of pricing and cost concerns. Another company was literally— the whole company was out of OpenAI tokens for the month. So if you're running, you know, agents, or if you're relying on you know, AI to fix a vibe-coded app and halfway through the month you're out of tokens, that's a problem. So for sure, I think the team leaderboards on who's using the most tokens, you know, that's probably been over for a few months, but that's, I think, really over now. And people are starting to think about costs. And if you can do something with deterministic code, as opposed to reinventing it every time using an LLM, you're gonna be much better off. So, that are my thoughts. I think it's a good analogy.

No, I appreciate that. I was just kind of curious and sorry, Glenn, I know I took your question, so I'll let you have back control. But every so often I like to do what you do. Just one more. Yeah, right. Just talk over everyone.

Oh wait, we have a guest today. I think from my standpoint, what you said about, you know, if somebody's writing VBA, they at least know what they're doing. So even if they hand it off to someone else and they're not using it, it's like, well, this is traceable. If you could, you know, if you understood VBA, you could make it work. If you're ViBe coding an app and it's a magic black box that's writing code for you and you don't understand the first thing about coding and you're handing, and then you leave, or, I mean, I think, you know, the, what's the Mark Twain quote that, you know, the reports of my death have been greatly exaggerated. I think the whole SaaS-pocalypse is, yeah, I get it. It would be cool if we could just all vibe code our own ERPs, but come on. What, like, I mean, you know, we're, I'm a finance guy by day and I'm building my own SaaS platform at night. That's not to say there's not a place for it, but it, people, you know, if you don't know a semantic layer from a layer cake, you know, how are you gonna structure and architect all this? So to me, vibe coding certainly has its place, but I think when people, it, it's interesting, Alex, because in a lot of ways we're solving very similar problems. Except I'm doing it, every client is a snowflake and it's all bespoke. And you're doing it the smart way where you have the product that applies, you know, across the digital universe. But the whole idea that someone who is a CPA and has spent 25 years as a, you know, as a controller up to CFO, that they're suddenly gonna be able to build apps that aren't leaking their company data, that they can explain to an auditor. Sure, the code sits out there, but a lot of times the code doesn't even sit out there. If they're building artifacts in Claude and sharing artifacts, it's like, well, you just made that a public link so somebody could stumble across. I mean, there's just so many, there's so many ways that this could be a problem. So I think, you know, maybe one day down the road, but there's, it's a discredit to me. I mean, if you say, oh, we don't need developers anymore, we can, You know, we can all develop our own apps. That's a discredit to what true engineers are doing and what they're building. And you don't even know, it's the unknown unknowns. You don't even know the right questions to ask, the things to check, how to set up OAuth or whatever it is that you're doing. You can't, you know, you can have AI do it, but how are you going to check that to even make sure that it's doing the math right? Anyway, that's, that's a long, long road.

Yeah, we could spend the whole time on it. Sorry, I didn't mean to derail us. It's just. It feels like there's going to become a moment of truth where we're going to recognize there's a lot of stuff that we probably went too far on, but there'll be a lot of great stuff that comes out of it as well. And so it's just interesting. I think you have two camps. Don't do it at all. Almost like these people are, oh yeah, it should be developers. You're stupid. Nothing's going to work. And others of, I can vibe code everything, you know? So anyway, sorry, I'll let us get back on track here. Ever feel like your go-to-market teams and finance speak different languages? This misalignment is a breeding ground for failure. Impairing the predictive power of forecasts and delaying decisions that drive efficient growth. It's not for lack of trying, but getting all the data in one place doesn't mean you've gotten everyone on the same page. Meet Qflow.ai, the strategic finance platform purpose-built to solve the toughest part of planning and analysis: B2B QFO quickly integrates key data from your go-to-market stack and accounting platform, then handles all the data prep and normalization under the hood. It automatically assembles your go-to-market stacks, makes segmented scenario planning a breeze, and closes the planning loop. Create airtight alignment, improve decision latency, and ensure accountability across the team.

Well, and then I think the transition from that is, you know, we're talking, we started talking, I think, before we went off about that data layer and how important that is. And, you know, one of the things, I guess, the Qflow promises is that you're going to help companies prep the data, plan the growth, run the business, and tell the story. And that's a, a whole workflow that has to start with the data. But maybe if you could kind of tie those together of how those stages work together in practice, and, you know, if you've got a real example of, of how that would work, that, that would be great to sort of visualize it.

Prep the data is always part of it, and sometimes it's fix the data. Uh, everyone gets the data dictionary out of the box when they sign up for Qflow. There's a light configuration layer that is AI quarterbacked. And, you know, that will, that automatically builds several dozen datasets that are deterministic and ready for analysis. So everyone has that. Some people to fix the data. We talked about some of those examples earlier. You know, I can't, I can't give my board a good revenue number for, or forecast because You know, I've got the 4 different CRMs and the company names don't match and we've kind of hacked it together. So that falls in the fix the data. And there we have a series of tools that do entity matching and transaction matching and reconciliation on an ongoing basis. So sometimes there's, you know, a push to clean up the skeletons in the closet or clean out the skeletons in the closet. And then, and for those companies, that's like the hair on fire problem, right? We've talked about this earlier. The forecast isn't broken, but the plumbing is. They know roughly how to forecast their business. Maybe it's seasonal, maybe it's based on number of, you know, subscribers or number of cameras under management or something, but they got a data problem. And so often, you know, we'll, We'll, we'll work with them to get their data into a good place. And then we're off to the races in terms of planning the growth and running the business. We had taken, I think, a unique approach to growth planning. It's one workspace where both finance and go-to-market leaders share the same inputs. These are the finance-grade conversion stats, timing stats that you would need to properly forecast out revenue, but it is also the unstructured sales data, the call recordings that salespeople are having. So we integrate all of that in to get a true multifaceted look at, at pipeline and unknown pipeline. So that's where those marketing statistics come in and we assemble all those for you into scenarios. You can create as many as you'd like. It's a 36-month rolling view, you know, so that company, most, the company I was describing that, you know, has the, the, the 4, uh, 4 CRMs can manage the, all of these scenarios at a per-entity level, a per-segment level, and have them all roll up into one unified, uh, number that they call to the board. The run, run the business, uh, we've got dedicated agents that run daily doing things like evaluating why companies are winning and losing deals. That's a, uh, that's a surprisingly popular product. Uh, turns out that sales reps like to choose just the first loss reason in the dropdown. And when they win, it's always because of an excellent relationship that they've created with the client. When you peel back the layers of the onion and you actually look at the call recordings, And you map those to a canonical list of reasons that you wish you had set up and someone had filled out thoughtfully, you get very different answers, right? So we have agents that do things like that, deal evaluation, budget versus actual go-to-market stat monitoring. And it's built on a blend of deterministic BI and LLM reasoning, right? The why for the narrative. And the final piece is telling the story. This is where we do expose an MCP. It is a, through a Qflow desktop app, which allows you to run recipes, which are in our product. Recipe is a series of deterministic steps that are auditable, that have drilldowns, that have timestamps, that have, you know, history tracking in terms of who touched what, when, and why. And those serve up the biggest use case here is automating portions of the board deck. So you can drag your existing board deck into Qflow, it will analyze it, and it will create recipes to create essentially the deterministic basis for each one of those slides. And then you can take the results of that and do with it what you wish in Cloud Desktop or Codex in terms of formatting. And so that's really the telling the story piece. And it's all the same dataset. It's not 4 products.

Yeah. It's running on Ada, which makes sense.

Interesting to hear QFlow has clearly evolved as AI capabilities have evolved too, and you're bringing them in. And I think that that's maybe another way. If you ask a CFO how many people in their, you know, if they're using AI, they're, you know, gonna say, heck yeah, we're using AI. You know, we're doing, but for a lot of people using AI means just dragging stuff into a chatbot, doing an analysis kind of in a one-off situation. And everybody's kind of, it's kind of like the vibe coding thing. If everybody's building their own apps or having their own analysis or whatever, But when you can bring it closer to the engine and you have it all in one system where everybody's using the same AI and it's in, to your point, which a lot of people may not be thinking about, when you build that recipe, there's deterministic steps. So that's repeatable, it's explainable, auditors get it, everybody gets it. And that's, I think, and obviously everybody in, in SaaS is, is trying to do this right now. But solving for that is what is gonna keep people on your platform and make them trust AI. Even if you're using the same model underneath, you've got the guardrails in because engineers are actually thinking about the layout there. And I think that's a significant thing to watch and it's been pretty fascinating to me. Whereas, you know, it's kind of like the same, maybe it's in the same vein, but where if you're worried about AI taking your job, I, I focus more on how am I gonna be better at my job with AI? And then with SaaS, it's, well, could AI do this? I don't know. Maybe if you had the right, you know, guardrails in place. But what if the people that made your favorite SaaS product actually just incorporated it in and you didn't have to worry about it and you got those better tools out of it? And it sounds like you're really, that's the path you guys are on too, evolving with, um, the technology as it's coming out there.

Yeah, for sure. I mean, AI was in our domain name because we were AI and ML first from a statistics perspective and a model perspective. The large language stuff, we're clearly evolving as everyone is, as that gets better. But I think the two advantages are the reliability and repeatability of using something like our recipes feature where you can see what it's doing and what's being, what's resulting from that run, number one. And number two, back to the comment earlier about expenses for companies that, that are focused on budgeting for AI and specifically who are averse to tokens. Having— you can bring your own token to Qflow. That's great. It's great for our margins. Go for it. But you can also just use one of the models that we have built in and that's reliable. You know, that reliable pricing is attractive to certain CFOs right now.

Totally makes sense. I think everybody's trying to get a handle on the pricing. It's, it's kind of like we're dealing with AI sprawl, just like we all dealt with SaaS sprawl. Right, something new great comes out, you all run to use it, and gets adopted widely. And all of a sudden you're like, wait, we're spending how much? What's the benefit we're getting? We really did— that was a million dollars last month, are we crazy? And you start to put discipline in place. It seems to be a recurring theme. But I want to shift gears a little bit. You know, something you've often said that I agree with is, right, the real purpose of technology is around aligning people. So how does that shape, you know, kind of Qflow and how you do things on a day-to-day basis, especially as we all know, you know, sales and finance, there's often a lot of friction, disagreement about forecasts. They kind of come at the data from different perspectives. So talk a little bit about that.

You were talking earlier about getting everyone into a room and getting everyone aligned on, you know, from the CRO to the CFO on what a booking was. And I think a great FP&A professional is the one orchestrating meetings like that and pulling people into the room. Probably the, there's a class of FP&A professionals who would just say, nope, this is the definition of bookings and I'm going to filter the dataset down this way and I'm going to continue, you know, with my my predictions off the side of my desk or in the closet, right? Predictions in the closet are worthless really. And forecasts matter when they justify operational change or future operational changes. And operational change requires having a team that agrees on what is true. So I think alignment is the prerequisite, not necessarily the byproduct. And most leaders get this intellectually, But day to day, they're stuck hunting the facts because of bad data and they're not strategizing. And that's the tax. So what we're building flips the ratio, right? With one source of truth, with evidence attached to analyses, even if they're AI-powered, right? The same number in the QBR, the board deck, and the variance analysis. So I think It's about fact hunting time collapsing and strategy time expanding. And in our product, where the agents stop, humans start, right? Not humans on the Keyflow side, but the workflows are built in there. So agents handle, you know, the reconciliation, the hygiene, daily variants, the deal risk surfacing, scenario assembly, and then they're surfaced to humans so that they can handle judgment calls, you know, evaluating genuine risk, the strategic bets, and ultimately the numbers that they share with their stakeholders. So the fact hunting theme is one that whether, you know, I'm talking to a chief revenue officer or a chief financial officer is a common thing. And whether it's fact hunting on a forecast call with the sales reps about, what's actually happening with the deal. Or we're fact hunting with, you know, between CFO and FP&A who's trying to get a handle on, you know, pipeline conversion and, you know, 365-day sales cycles and what's real and what's not. And so I think it's about getting to a place where we can make decisions, not arguing over who's right or who's cut of the data is right.

Right. As I always like to say, if you're in a meeting and I don't recognize that number, how did you calculate it? You're pretty much done the meeting. You're not going to get anything good done the rest of the meeting if that's, that's brought up early. That's what I always find because it just, right, delves into a conversation of data instead of what we're trying to accomplish.

That's right.

God, you're talking about aligning people and I just, my, I think my old military roots came out and I was like, the way you tell, the way you align people is tell them, this is the way that I'm doing it. I'm the captain. And we're just, but so I, it's funny, the soft skills, all this, the soft skills and all this are tough. And I've really just been sort of fixated on, I'm not going to go too much further down the road of, you know, what we started with generative AI and how that's changing SaaS today. But I do think Right now there's so much noise between you need to do AI and this is the tech stack you need to have and this is the data warehouse model. Here's your semantic layer. There's so much going on. And I think again, going back to my domain expertise comments earlier, it's hard if, you know, you didn't come up as a technologist. I think we all have to be to whatever extent in our roles now, but for a CFO or an FP&A leader that's trying right now to modernize their planning stack in the midst of all this noise, like where, I mean, from your point of view, Alex, where do they start? What do they, you know, what do they need to look at with their tech stack and what should they, you know, kind of stop doing first before they, make any technology decisions?

I think that I would focus more on the questions that your team can't answer in numbers as opposed to evaluating just like the tech stack. Clearly, I have opinions about different vendors and models and, and, but I would focus on like the one or two questions that your team can't answer in numbers. And to solve those first. So some recent examples that come to mind, I mentioned deeply understanding of what's going into long sales cycles and partnering to fix it, or understanding why, you know, 3% of marketing leads convert when you think that number should be closer to 30. So I would separate these from longer plays. And I would stop the scavenger hunts for answers, right? People should know, should have one place to go to ask the question and get the answer and not piece it together from CRM, BI, finance files, Slack threads, and, you know, MCPs that hook into each one of those things. I would also probably stop adding dashboards. You know, that's, you know, every new Looker dashboard is yet one more thing that someone can change a filter on and disagree with. You know, one more place to where people have to check before they trust the number. Yeah, and I would also stop treating manual reconciliation as routine. I mean, the number of companies that are well endowed with funding and, and lots of revenue who are still, who are still doing a bunch of manual reconciliation as opposed to having humans investigate flagged issues is, continues to surprise me. Did I, I, I answer your question sort of from a different angle there, Glenn? Not really the, like, the tech stack angle. Is that—

Yeah. And we're aligned on that. And I, for me, it's, you could always tell when the right salesperson got to a CFO and sold them on an ERP or, or whatever before even knowing the requirements. So you, you, I think the mistake is there's some shiny new fancy whether it's a SaaS product or AI itself, something out there that you think, oh, we're just going to go sprinkle some AI on this, or we're just going to go plug in this one magic piece of software and it's going to solve all of our problems. I think you nailed it in the first part of that answer of, how about let's come up with requirements first? What are we, what are we solving for? And then search for software around that rather than trying to, uh, you know, jam the square peg into a round hole.

Yeah. And I think that then it's also about setting timelines, honestly. There are certain things which are absolutely quick wins that AI can help with. And then there are other things where, you know, like replacing that 29-megabyte ARR waterfall spreadsheet where someone from finance has literally typed out, you know, the contract details for the last 4 years into columns that drive the waterfall, like that might not be— today's Tuesday, we're current, you know, I'm not going to promise that's going to be fixed by Thursday, right? We can use AI to help us in that project, but there's some underlying data and data structures that need to be fixed and need to be transitioned somewhere else. So requirements and, and, and realistic timeline depending on depending on what those requirements are.

We're going to move into the fun where we ask a couple personal questions developed by AI. So I thought it was fun that we got to use Fable 5 before, you know, it got taken away from us. We all got banned, so to speak. So I'm going to read what it said because I thought it was kind of funny at the beginning. And then I'm going to let you pick a number. So Glenn and I both take a different approach. We both get one question. You get two options with me, Alex. You can pick a number yourself between 1 and 25, or you can take a human out of the loop and I let the random number generator pick a number between 1 and 25. So I'm going to let you choose that. And then I'm going to read this first little part because I think you'll find it interesting. So which one do you want? Random number generator? Are you going to pick yourself, Alex?

You know, there's probably, without knowing the, uh, the random number generator, there's probably an argument that it's not truly random. So I'm just going to say 25.

All right. So we'll get there in a second. Before I do that, here's what it said. So it said I gave it the prompt and it goes, excavated candidate background to enrich interview questions, found some good material to work with. Alex has a BA in international relations from the College of Wooster, studied international relations and history in Paris. Is based in San Francisco, spent some time at UC Berkeley. And here's the part that I thought was interesting. I think you'll laugh at. And his Twitter bio includes, fear is the mind killer, a Dune reference, which is a gift for this kind of question list. And then it goes on to say, here are 25 fun, personal, and slightly quirky questions. All right, here we go. Number 25, it says lightning round. So I'm going to ask you 4 questions and you have to pick one or the other. Excel or Google Sheets?

Excel.

In person or remote?

It's hybrid, but in person, I guess, then.

All right. Tabs or spaces?

Tabs or spaces?

Your guess is as good as mine.

I guess I'll say tabs.

Pineapple on pizza, yes or no?

No.

All right. You passed. That was number 25.

With tabs or spaces, is that something like in a new line, like, I don't know, I'm trying to think of when you would tab to a spot. I don't know.

That's kind of what I'm thinking is that with typing for cat, do you space all the way through or you just hit cat? I was struggling a little with that one. That's why that one was actually listed first, but I decided to put it toward the end because I'm like, uh, the other ones make sense. We'll, uh, switch the order a little. So I did do a little human in the loop there.

That's good.

I just realized what I did. So my approach, I just copy and paste whatever model comes up with and ask it what's the best question, but ended up deleting question at my point. But I think I'm going to stick by my rule. So I always figure AI wrote the questions. I ask it which one's the best one to ask, but this one, based on what Paul read at the beginning, and then at the end, a few notes, question 1 is probably the strongest. So with all that as prompting, of course it picks question number 1. So I'll go with that and, uh, say question number 1, your Twitter bio says fear is the mind killer. Are you a hardcore Dune fan? As a hardcore Dune fan, I'm asking this. And do you have a hill you'll die on about the books versus the movie?

I don't have a hill I'll die on about the books and movies. I actually enjoyed the movies, but obviously read the books long before. In my wedding vows with my wife, there are quotes from Dune.

Nice. We got some good material outta that one. Well done, Glen. You wanna share any of those lines that you used?

I could probably search for it in Google Docs.

We won't make you do that.

I think my comment would be Denis Villeneuve killed it. I'm super excited about the third installment coming. David Lynch, interesting guy. Boy, that was a botched movie. That was his version. I mean, I've still seen it a dozen times, I'm sure, but it's nice to see that, you know, because for years it just seemed like Dune was an unfilmable book. And I think breaking it into 3 movies—

they had the version with Sting. Like a, a—

yeah, that was the one that David Lynch did that was just weird. Yeah.

In all seriousness, thank you so much for joining us. Alex is our guest. Recognize Alex does run Qflow. He's our sponsor, so we're incredibly grateful for his support on the show. We think he has a great product and really enjoyed the interview with you. So thank you so much for joining us, Alex.

Thanks for having me.

I appreciate you guys.

Thanks, Alex. Always fun.

Thanks for listening to the Future Finance Show, and thanks to our sponsor qflow.ai. If you enjoyed this episode, please leave a rating and review on your podcast platform of choice. And may your robot overlords be with you.