Many companies are rushing to adopt AI tools hoping to unlock new levels of productivity and innovation — and failing. Why? According to Sarah Jeannault, former FinTech founder and now VP Marketing at ProcedureFlow, it’s because they skipped a crucial step: building a strong operational and knowledge foundation. In this episode, Sarah dives into why AI initiatives fall flat, how to fix broken knowledge systems, and what a true AI-ready organization actually looks like. We also talk about marketing’s evolving tech stack, change management, and why AI must empower frontline teams — not just leadership.
Auto-generated transcript. Speaker names, spelling, and punctuation may be slightly off.
Mark Evans: It's Mark Evans, and you're listening to Marketing Spark, the podcast dedicated to helping b to b SaaS CEOs, entrepreneurs, and marketing leaders unlock growth, scale their teams, and navigate the ever evolving world of go to market strategy. Today, we've got a practical and insightful conversation with Sarah Janeaux, the VP of marketing at Procedure Flow. She's a fintech founder with two successful exits and an expert in AI adoption and knowledge governance. On today's episode, we're diving into a topic many SaaS teams are grappling with right now. What happens when companies rush into AI adoption without first fixing that foundation of their knowledge systems and internal workflows? Sarah, welcome to Marketing Spark. Let's start with your journey from fintech founder to VP of marketing and now a leader in AI and knowledge governance. What experience has shaped how you think about scaling teams and systems?
Guest: I have this experience firsthand by somebody who thought, this is brilliant. I'm going to build and sell a business. And of all of the hard work that requires, and we all know that as people who are entrepreneurs and founders who have built and scaled, That process of which you work through as the business grows is something that we all have to adapt to. Right? And we're only working with limited budgets with huge expectations, and we need huge wins. And one thing I learned on my exit of my second business was that there was a lot of process internally within the organization that was all sitting in my head. What's interesting and why I joined Procedure Flow was for that very experienced because sometimes when we're entrepreneurs and we're trying to build so fast, we're not often thinking about what are we doing with our teams internally to make sure that we're really operationalizing for efficiencies while we're trying to hit all of those targets.
Mark Evans: When you talk about documentation and getting workflows out of your head and into some form that can be shareable, is it playbooks? Is it operate operating manuals? What does that look like in the real world so entrepreneurs can understand where they're at right now? Many of them are operating right in their heads right now and where they need to be.
Guest: That's a tool that ProcedureFlow actually uses. And, again, why I joined them because I found that level of efficiency was so important. Sometimes those documentation pieces are all over the place in a company. Let's be real. We don't always have the time to sit down and write all the steps. So what our tool does is take many of those SOPs and put those in a nice workflow so that it is efficient for other people to be able follow those processes. So they're not always texting you and calling you and saying, hey, how do we do this? I have this problem with this client. I'm not really sure what to do. And really because that tool is so essential because at the end of the day, if you're at the top of a small organization pushing it hard or you're at the top of a really organization, large organization with complex processes, you need the ability for your team to be able to efficiently go to the steps, figure out how you can solve the problems for your customer, and then make sure that everybody is happy without picking up the phone every day and calling somebody to be like, actually, I don't know how to do this.
Mark Evans: Let's bridge this operational foundation to a subject that a lot of companies are dealing with right now, which is AI adoption. As much as people are excited about AI and the tsunami of tools that are coming our way, Some of them are eye candy. Some of them actually have some substance. But one of the common refrains these days is that many AI initiatives within companies, big or small, fail. You hear a lot about the fact that there's very little ROI that's happening or the productivity gains that a lot of companies anticipated aren't there. Your contention is that failure is often a result of the knowledge layer isn't there. I'm interested in, obviously, digging deep into that, but let's start with what exactly do you mean by that, and why is it such a fundamental issue, especially when it comes to a new and emerging technology like AI?
Guest: And be remiss if anybody was not want going on here with what's AI. It's supposed to be the answer for everything. And maybe it will be. We'll see how it goes down the road. But in essence, I think we all need to stop for a moment and really think, okay. What am I asking AI to do in my company? And what do I need it where is it pulling that information from to actually give me the answer? Because AI is very good at lying to you. It's very good at convincing you one point of view, and that is important, especially for regulated industry businesses. You don't want to have AI telling you something where you're gonna stand up there confidently and say, this is how it goes. And that is so critical because we don't want anyone getting in trouble with following the wrong steps, is we need to say, okay, where is AI pulling their information from internally to make sure that it's giving our employees or me the answers that I need that I can then go out and represent or action on? If it's we're talking about, let's say, a context The utilization of AI, especially in this lens of knowledge layer, is really utilized across all industry and it'll be that output will be different. But you really need to stop and say, where is the information coming from? Where what is it? Is that information up to date and accurate? Is it even correct still? Just so we can't move forward sometimes with AI. We can't move forward with its ability to innovate if we don't stop and just make sure where is it pulling that information from to ensure that it's actually the right one, and it's giving people the actual efficient decisions that you need for your business.
Mark Evans: Great point. And I think it's the question that I have going forward is how do you get that consistency? How do you determine that that you're getting the information within the enterprise context that you need? So there are a couple approaches here. One is you can use a Claude or a ChatGPT, and you can feed information through custom GBTs into the system. And you can basically provide it with a library of documentation that you hope it will use, but it'll also go to the external rules and start pulling information externally. Then the opposite is that you use something like a Glean, which essentially just provides us almost private AI environment where it's only pulling information that you fed it. So you can theoretically guarantee that it's gonna give you more accurate information, more relevant information. So how do companies move forward and leverage AI and, to your point, make sure that the information is accurate? What are the different approaches that they should be looking at?
Guest: One, I think it's stopping to think, who owns this? Honestly, we've moved ahead with AI as, like, our solution without really stopping for a moment and saying, within our internal organization, who owns the information that's being fed into it? Who and also, who are the people going and collecting the information that they're looking for in terms of the answers? We need to define operationally within the organization who owns that piece, who owns the that responsibility to have the single source of truth so that we can then go in and layer AI into our processes to make sure that answers. And whether you are using various private or public AI tools, it is still essential that the information is a single source of truth that's going into to educate that as we are looking for outputs. And honestly, I know that's not the fancy answer because people are always looking for what's faster, what's more efficient, and how do we go. But I do think that there is a necessity for some governance here about really being reflective of what exists within my organization to make sure that output is the piece that I'm really looking for.
Mark Evans: The question would be, is this a new role within in many organizations? Is it the chief information officer, the chief data officer, the chief AI officer, or is it a role that somebody in the existing infrastructure has to step up and take? Because it sounds like you need a point person to make sure that at the end of the day, someone is ultimately accountable, someone is leading the way strategically and tactically so that there is consistency. As to your point, there's proper governance in terms of the information available. How do you appoint that person or create an infrastructure where that person exists?
Guest: I'd leave it to the structure of the organization to figure out who they wanna appoint. But I would say the thing you need to be most mindful of is how am I connecting to my frontline teams, and how am I getting information back from the frontline team back up to the people who are making the leaders who are making the decisions. The efficiencies can be found within reducing headcount if you wanna go there or increasing the time that clients are getting information, whichever variable you wanna be tracking. I do think that you can find efficiencies in headcounts over there and that you would you'd need to utilize somebody that is now responsible for this. So I do think it's an iteration of finding someone within an organization for sure to make sure they're the single source of truth. But I think that the operational piece of efficiencies grows exponentially from that one person when we're making sure that AI actually is trained correctly. Because then that it's almost like then going out within the universe of your organization, which now has a single source of point of truth. While you could think about it as something adding on to someone who's higher up in the organization, my only caution there would be to be reflective of how much of a feedback loop do you have with your frontline employees to make sure that we're really hearing where those pain points are and making sure that some of those pieces are correct. Some organizations might have them, especially regulated spaces. You already have those governance principles already in play, and I think it's perhaps shifting and how we're looking at that information and how we're reviewing those various departments. But I don't want this to sound too controlling either. And again, because there's such scope here of whether you're an incredibly large organization or small and depending on the industry, there's absolutely a role to be played in AI in any iteration you want to use it in. There are absolutely efficiencies there. But moving ahead sometimes with so much speed without being reflective of what are we asking the tool to do and why, and how am I ensuring that the answers that are coming out actually are accurate to help really improve my bottom line. Those questions just should be asked before we jump five steps ahead into it all.
Mark Evans: One of the biggest challenges for a lot of companies that they don't know what they don't know. It's almost like they operate in a bliss blissful state of ignorance. Everything is great. The status quo is amazing. There's no reason to change. When you're talking about a broken knowledge layer, A lot of companies may not recognize that even exists, that they have a problem, that their knowledge layer isn't rock solid or as or as rock solid as it needs to be. So what are the signs that your knowledge layer is broken? What are the symptoms that you could look at and go, oh, jeez. There's something not right here. There's something that needs to be fixed. Because otherwise, they just operate on a continual basis, and the problem never gets solved.
Guest: Yeah. I think it's also sometimes, like, how much is in your desk drawer kind of scenario. I don't know about you, but I have kids. And sometimes I look at all of the things that they've got stored in their lockers and that they don't really deal with, but it's okay because we'll just slam it shut and we'll just keep going forward. I think we need to be thinking about that same way within an organization. And, I think that's that connection with that frontline team. How often is an entire team going to one person to ask for the real answer? How often are we hearing back from customers that they were hearing things that were just a little bit off? Why are we looking at our satisfaction rate from customers to see how they're feeling about some of the services that they're receiving? Those feedback loops, I think the answers are there. Now do we have the time to look for those pieces? I don't know because we're always, again, pointed towards that revenue trajectory. So tools like ProcedureFlow can actually help prevent that in the first place. I do wanna throw that out there by making sure that there is a tool layered in with your AI, if you have AI, or layered in with your humans depending on how your organization is working just to make sure that's not even a problem in the first place is actually quite helpful. But you're right. Sometimes we don't know what we don't know. But stopping just for a moment and asking, I think the biggest question is, who answers all my questions when I need the answers? Who is everybody going to? And what are people actually really struggling with from a customer feedback point of view? Your answers are right there.
Mark Evans: Curious about your take on the adoption of AI tools within organizations. So there's a couple ways that it happens. One is is is bottom up is you get people within the organization starting to use ChatGPT or Claude or one of the very myriad of tools out there, and then it's top down. Where the organization says, we're gonna use Glean. We're gonna use these other AI tools, whether it's a clay or some other AI powered tool. And then what inevitably happens is you've got a mishmash of tools. Some have been have the stamp of approval. Some don't. So when you're looking at companies that are looking to adopt AI tools, like, obviously, they have to have the right knowledge layer in place. But what other advice would you give companies as they look to leverage productivity and improve efficiency? Where do they start with AI tools? What's the best approach forward so that everybody is on the same page?
Guest: I think what's happened is everybody has said, oh, AI. Okay. And then they're asking, oh, what can AI generate? And I think we need to stop for a minute and say, what do we want AI to know? So that is a fundamental shift. Right? This isn't just a feature tool, I think, that's just here and will leave tomorrow. This is a fundamental shift in how organizational operations happen moving forward. I think when we think about that, longevity within a business, we need to have the operational rigor to decide how this is gonna be implemented. And I would hope that those at the top are really having listening conversations with teams on the front lines to ensure that we're putting in AI in areas that not just support them, but that actually builds real efficiencies for them. That's an operational shift within an organization. It's so important that just because we're going to put AI in here to answer some questions for customers, that we're still looking and listening to our internal employees to make sure that the process is still in place within the organization, so we're setting everyone up for success. And like you said, that is exactly what's happening. We're seeing top down decisions happening and then bottom up people saying, this stuff's working. This isn't working. And we need the middle. There there needs to be some mixture here of communication for them to figure out within their organization on how they're gonna make that most efficient because it's not going away. And the problem will just keep growing there in the middle unless they figure out operationally in internally how they're gonna solve that.
Mark Evans: I think one of the biggest challenges is that top down senior management is looking at a world of productivity and efficiencies and, in many cases, reducing headcount. And bottom up, they're looking at these tools, and there's concerns about is AI going to replace me or make me do some kind of different job? How do organizations manage that conflict? Obviously, two different parties with two different mandates. What's your advice in terms of allowing an organization to go forward leveraging AI in a way that's a win win for management, win win for employees.
Guest: There are organizations who are doing this really well, and there are organizations that are still working towards that success goal. It's in the messy middle. I think what AI is doing for us, and I know everyone's talking about efficiencies, but it's also really being reflective for us to say, how should an organization be set up today? What is a new reality of customer experience? How do we define efficiency today? It's a different answer than it was ten years ago. And how do we layer that in? I what I fear most is that people are afraid of the unknown and afraid of the not knowing what you don't know. I think we need to be approaching this with a growth mindset. With that ability to say, this how can I make this as great as I can make it by realizing that I need to continue to grow? AI is not a stamp on a piece of paper or a check mark on a budget line to say, I did this. It's an ongoing process of looking for efficiencies within your organization. And I think that is the critical piece that top level leaders need to recognize. It's not a one and done scenario. It's an ongoing conversation. And what and for companies who do embrace that opportunity, what a beautiful thing that is, that we finally have actually an opportunity for frontline teams to be able to have a voice that's loud enough that is going to the top tables, and that there is this beautiful communication back and forth. Sure. There'll be tension. There's budget lines. There's deliverables. All those things will still exist, but it is providing opportunity for communication. And those who are most afraid of AI is generally because they're afraid of what it's taking away. But I would actually suggest we should be looking at this as what it's bringing to an organization, how this is actually empowering you as an employee within an organization to be able to have a voice to look for efficiencies, to find ways to be more operationally efficient and to be more relevant to an organization where they can't let you go because you are so good at making sure that feedback loop is happening. And at the end of the day, we're building revenue, and you're helping customers feel satisfied.
Mark Evans: Let's stick with AI, but let's turn it from a philosophical or operational to the practical. A lot of marketing leaders, including yourself, most likely, are being asked to figure out a AI. What should they be thinking about in terms of their go to market and enablement strategies in the context of AI adoption? And give me some personal insight into how you're embracing AI strategically and tactically with your new job at ProcedureFlow.
Guest: So this is fun because I think it's the same thing. I definitely have approached with my teams with that same growth mindset that we are also going to walk the walk here, and we are also gonna integrate some of our own products internally within our organization, which does provide some opportunities for sure to be like, yep. These are the steps. But also for us to say, why are these steps this way? We have opened up opportunities to have a little bit more of a feedback loop internally, not just within our marketing department, but a real go to market strategy. Where can we find most ways to provide accountability within the organization that doesn't just sit on one or two people, that is something that we are empowering different people to make different decisions now because they can rely on a knowledge tool like ProcedureFlow. And that's important. And you to be honest, the tension there is always that go faster, and we will go fast, and we will be efficient. But we do have marks of opportunities where we're stopping weekly to make sure that the entire team is accountable to doing what they need to do and that we are understanding that we're breaking down some of these silos within an organization because now everybody can see what those steps are. It's not hidden anymore, and we can keep each other accountable as well as a group to make sure that we're delivering the things we said we're going to deliver and make sure that we're listening to other teams. So it's a real efficiency space, but the root of it really starts with being open and okay to having conversations, listening to other people, and figuring out perhaps a new way next week that might need some tweaks based on what we're talking about this week.
Mark Evans: I think what's what's most interesting to me from a marketing perspective is that over the years, the tech stack has gotten bigger and bigger, and we all have dozens of tools at our disposable. As a marketing leader, I'm interested in how you assess and audit your current tool set, and many of them are traditional SaaS tools, and how you determine where there are opportunities to leverage AI and AI powered products and how you measure the success of the adoption of these tools. Because it's one thing to talk about AI. It's another thing to embrace it and actually get the outcomes that you want to make marketing more efficient and more successful. So walk me through that process where you've got all these tools. There's all these AI tools out there that promise to make you basically, make you the smartest person in the room, what is that process where you say, yes. This is these are the things we're gonna focus on. These are the tools we're gonna embrace, and these are the results we're looking for.
Guest: I'm all about empowerment with my teams and the my own personal I'll give you, I guess, my internal monologue on this one. When I have a team meeting, a cross functional team meeting, and I hear people working through how we're gonna solve something, and they have just casually brought up something like Descript, something where we're adding in AI and they're just naturally doing it in terms of the steps they're gonna follow to achieve what they're looking for. To me, that's real success. Because what's happened there is that someone else has identified a way they're going to layer in one of the tools we're using without be prompting it, and then actually making a real connection to their day to day what they're going to do, which then will help with that output piece on the other end. That is my true success. Now, getting there, for example, is not as, oh, this works every day because it doesn't. It still takes some queuing questions as we're working through what are we looking for, what campaign are we working on, what are we actually trying to achieve, and then just keeping being mindful of, hey, we have these tools. Is there any way that this tool I'll just use Descript again as an example, is that this tool can be utilized at all through our process here. I'm not putting pressure on and saying, I need this to be two days faster because you said that you were gonna use this tool. That's not the conversation. It's a coaching conversation of how do we embrace this in an inclusive manner to make sure that they're feeling that there's some operational efficiencies in what they do. I'm trying not to be top down, but to be able to provide tools. Now, I'm also looking at, and certainly a finance department would definitely be looking at, how often are these people actually utilizing this tool? Does this really make sense to put spend and then really linking back to efficiencies? So those conversations also need to occur, but this isn't I just feel strongly that this is a changed management time within organizations as a whole. We there's different times in the world where you have pressures in terms of economic pressures or pulling back and scaling back on what we're spending, and everybody's gone to their board with whatever plan that they have at the time and they need to get something across. I do think a principle on all sheets of in a c suite should be saying, how are we using AI tools efficient efficiently, and where can I find evidence of them being used efficiently within the organization? That key critical question is something that should be happening within just top layer, middle layer, and frontline teams to ensure that we're not just using this for us, that stamp to say, hey, I use this, but to actually be listening for how people are using that in an organization.
Mark Evans: Let's drill down into AI tool selection. How do you identify the problem that potentially could be fixed or made more efficient with an AI tool? What does that approval process look like? And we can even look at a tool like Descript, which I'm a big fan of in terms of the ability to edit video and podcasts, add captions. It makes an individual into an amazing video editor as opposed to using an external person or legacy tool. So how do you identify there is an opportunity here for an AI tool to make something better? And then what does the rest of that journey look like before you would actually adopt it?
Guest: So it's in that planning phase within our organization, at least for us. We've come up with a campaign we're gonna do. We've decided we've linked it to, like, why are we doing this? What does this mean for the organization? What does this mean for our clients? So we're going through that structure of planning out a campaign. And it's in that process when we're starting to identify what are the pieces that we have, which and and if you had procedure flow in your organization, it would be popping up as prompted questions. And then it's again saying, okay. Of the tools we have, are any of these tools being utilized? And then the answer can be yes or no, and this is where it's at. So in the planning phase, one. And then two, also in those coaching conversation, in your one on ones, when you're listening to your teams and hearing what are they working through, where are the problems, one, prompt back to that knowledge flow within the organization to say, oh, are there anything here? We identify that we were going to be thinking through where our AI tools and does anything need to be used here. So that prompting piece there is also critically important. And also, it's just that open mind. Like, one thing I caution is I don't want everybody taking one tool and saying, I'm always gonna use this one tool. In every single iteration, I'm always always there are times where Descript is great, and then there are times when, no, I can't use that here, and I need to have something else. So you need to have flexibility. So thinking about as a leader ownership, who owns it? Where does it fit in the process? What is our something I'm looking for as the end result? Is there any efficiencies here in adding AI? And then also just from because we're in marketing, we get to do the creative stuff, the fun stuff, the push the envelope stuff. Do you in an organization have a sandbox place? Do you have a place whether it's a Slack channel, whether it's just a little you're training a little AI bot somewhere else where your testing ideas, it there is no it's not a wrong answer. It's just a try. Do you have that going in your departments and teams? That is so critical because those might not be the things you're actually gonna take to market, but that's where you're pushing the envelope that you can then decide, okay. Now I'm gonna bring that back. There is an element here that I'm gonna play with that works. And just personally, internally, thinking about something like Descript, where what we originally got it for, which was thinking, oh, this is gonna do all this video editing. But then from playing with it, realized it can actually it can fix the color. It can do a couple extra pieces that I never even realized. And then just like that, I can do it in instant. I can solve my own problem now. That came from the sandbox that I didn't think was the original need for the tool, but now I use it all the time. That creative aspect in pushing how you might use a tool in a place where it's safe for your teams to experiment somewhere else is also really important.
Mark Evans: Final question. Given the, for lack of a better word, pressure on marketing teams to embrace AI and the number of AI tools coming our way that are irresistible and intimidating at the same time, What's one piece of practical advice that you'd give about aligning an AI strategy with going back to where we started with scalable knowledge systems?
Guest: I think it does go back to scalable knowledge systems. It goes back to saying, what am I doing? Why am I doing this? Who is going to do it? And what's the output I'm looking for? And I know those sound simple, but sometimes we try to overcomplicate things, and it's really in the basics where the biggest improvements and gains are. So if I'm going back and thinking through those questions and saying, okay, who owns some of this? And then from that ownership, do they actually have the answer that we need? And then do we have that efficiency down in terms of that process is clearly outlined? And then is there any areas that I can layer in AI here to look for some efficiencies? But really, just back to that core piece, who is the authoritative person, the responsible truth source of truth here to make sure that before we layer in AI, it actually is getting the answers we're looking for, and it's giving the efficiencies we're actually wanting as an output.
Mark Evans: So this has been a very practical and thoughtful conversation. Where can people learn more about you and ProcedureFlow?
Guest: Hope everyone will come check out procedureflow.com, and you can always reach out to me over on LinkedIn.
Mark Evans: Thanks for listening to Marketing Spark. If you enjoy this conversation, share it via social media, subscribe via your favorite podcast app. And if you're so inclined, leave a flagstar review. Thanks for tuning in, and we'll talk to you soon.