[
  {
    "speaker": "Speaker 1",
    "text": "Hi everyone, Charlie and Chrissy here with the cooking up go to market podcast. So we have a very special guest today, Allison from our team who leads up our solutions architecture and"
  },
  {
    "speaker": "Speaker 1",
    "text": "analytics practice. I think every time we've had Alison Allison on has been one of our best performing podcasts, right, with a lot of good stuff around analytics."
  },
  {
    "speaker": "Speaker 1",
    "text": "Probably the one person that knows the most about attribution, Salesforce, so many different things. Um, and we're going to have a big part of the topic today talking about revenue planning and"
  },
  {
    "speaker": "Speaker 1",
    "text": "AI and how we've been using AI to prepare for revenue planning, get the most out of revenue planning, and help our clients plan. Um, but before we get into that, we've got a few other topics."
  },
  {
    "speaker": "Speaker 1",
    "text": "So, the first one we're going to start with, so CLA is going to throw this up on the screen when when we get there."
  },
  {
    "speaker": "Speaker 1",
    "text": "Um, but this is a this is a a tweet by Aaron Levy from from Box. So, I'm going to read it out for anyone that isn't um watching and then I want to get the team's reaction here. So, CEOs are"
  },
  {
    "speaker": "Speaker 1",
    "text": "uniquely prone to AI psychosis because they're sufficiently distant from the last mile of work that still has to happen to generate most value with AI."
  },
  {
    "speaker": "Speaker 1",
    "text": "So, when they play with AI, they see the happy path results, often not considering the next 10 or 20 things that have have to happen to get sustainable results from agents. Look, I've made this awesome product"
  },
  {
    "speaker": "Speaker 1",
    "text": "prototype. Yes, but you didn't have to review the code before it went into production and fixed a bunch of issues."
  },
  {
    "speaker": "Speaker 1",
    "text": "Look, I generated a contract. Yes, but you didn't verify all the terms before it goes out to the counterparty and didn't have to wire up all the past contracts to work with. The best thing"
  },
  {
    "speaker": "Speaker 1",
    "text": "you can do as a CEO is to use AI a ton to figure out the real implications for agents in the enterprise and come out the other side with an appreciation for both the upside and the real work that"
  },
  {
    "speaker": "Speaker 1",
    "text": "goes into them. So my question for Allison, do Chrissy and I suffer from AI psychosis. [laughter] Tell us the truth. But you can you can avoid that comment or that question if you want. And what do you think? Do you"
  },
  {
    "speaker": "Speaker 1",
    "text": "agree with this? Where where do you see value in this statement? And and do you agree? What do you think?"
  },
  {
    "speaker": "Speaker 2",
    "text": "do, but I feel like um my opinion on on that is similar to opinions I've had in the past, especially when it comes to operations in general, that often leadership doesn't kind of understand"
  },
  {
    "speaker": "Speaker 2",
    "text": "all the work that goes into what we do in operations."
  },
  {
    "speaker": "Speaker 2",
    "text": "um especially when it comes to like as we'll get into later like things like revenue planning where your data has to be in such a good place to really be able to effectively build a model. Um"
  },
  {
    "speaker": "Speaker 2",
    "text": "but also um your data has to be in such a good place in order for AI to work well with it. So I think there's like a probably a misconception in a lot in a lot of leaders eyes that their data is"
  },
  {
    "speaker": "Speaker 2",
    "text": "clean enough to to be able to leverage AI the way they want it to and that there's not a lot of work that's still going to have to go into that first."
  },
  {
    "speaker": "Speaker 1",
    "text": ">> That's a good point actually. I hadn't even thought about how it is a continuation of a an old an old problem that we've always had to deal with in office. Do you do you agree Chrissy?"
  },
  {
    "speaker": "Speaker 1",
    "text": "Because it kind of sounds like is it just like an expectations has gone up problem but actually the reality on the ground floor is always been difficult or is it both reality is even harder to"
  },
  {
    "speaker": "Speaker 1",
    "text": "achieve good results now with AI makes it even more complicated and expectations have gone up like what do you think has changed the most or if anything in this dynamic that's changed"
  },
  {
    "speaker": "Speaker 3",
    "text": "the most though is kind of what he's saying in terms of like people like now leaders are are using the technology and they're getting this kind of false perception on like what's possible. It's like I talked about it"
  },
  {
    "speaker": "Speaker 3",
    "text": "multiple times on this podcast. I've said it multiple times. It's like a little bit of information is a dangerous thing. And I think like that's what they have in their purview is like on"
  },
  {
    "speaker": "Speaker 3",
    "text": "operations now. It's it's a little bit like before they didn't actually know anything about operations. They just knew it was hard and they knew it was important."
  },
  {
    "speaker": "Speaker 1",
    "text": ">> Yeah. weren't logging into the marketing automation platform and building stuff, were they? But now they are."
  },
  {
    "speaker": "Speaker 3",
    "text": ">> And now, oh well, I could do this or I got cloud to, you know, and that's even like worse because yes, they're using it for certain applications that are quite simple. They have a very happy path, a"
  },
  {
    "speaker": "Speaker 3",
    "text": "short goal, um like a simple goal. And as we know, if you're taking any type of technology, like it's kind of like when you see a demo of something, you're seeing it like happy path. You're seeing"
  },
  {
    "speaker": "Speaker 3",
    "text": "it work in its best environment, but then it's not you hooking that up to your production systems, all the tech debt that's already in place and so forth. And so I think what's even worse"
  },
  {
    "speaker": "Speaker 3",
    "text": "here is now that like they have that like perceived a and then on like you know that perception but on top of it they're hearing their peers saying how much they're doing with AI but in a lot of"
  },
  {
    "speaker": "Speaker 3",
    "text": "ways it's their peers having to say they're doing a lot with AI because and also build be AI savvy because they also need to impress their investors get more money from investors and so forth and"
  },
  {
    "speaker": "Speaker 3",
    "text": "and stay relevant in their job. And if there's like one thing a seale person is really good at is like, you know, painting that persona and picture and being, you know, that's a big part of"
  },
  {
    "speaker": "Speaker 3",
    "text": "their job is being that kind of leader and and so forth. Um, and yes, some people are and some some CEOs I will say are pretty tech-savvy, like especially engineering CEOs and stuff like that."
  },
  {
    "speaker": "Speaker 3",
    "text": "And so in some ways they do have that but a lot of these a lot of them don't."
  },
  {
    "speaker": "Speaker 3",
    "text": "And so I think the difference here is like it's a little bit of a dangerous thing like I said where they see things happening for themselves with AI from a a personal use case and but that doesn't"
  },
  {
    "speaker": "Speaker 3",
    "text": "translate back into like what's actually it takes to add that into their company's like systems and um it's the hardest part about technology. We know that it's not using the technology. It's"
  },
  {
    "speaker": "Speaker 3",
    "text": "it's not adopting. It's not figure out how to push that technology to its limits. It's more how do we integrate this or use it on top of like what we already have and the people how do we"
  },
  {
    "speaker": "Speaker 3",
    "text": "align the people around that tool and how do we you know and there's a lot of tech debt obviously that you would have to clear up and no one has time for that in [laughter] a lot of ways too. So uh"
  },
  {
    "speaker": "Speaker 1",
    "text": "yeah. So I agree."
  },
  {
    "speaker": "Speaker 1",
    "text": ">> Yeah. Yeah. So, I know I started off this asking if Chrissy and I are delusional, and I don't know if anyone saw the newsletter that's listening that we put out last week, but one of the"
  },
  {
    "speaker": "Speaker 1",
    "text": "things that >> we've personally been focusing on to make sure we're not delusional is, you know, we don't have a massive team anyway, so we are very pretty close to the ground level of how things are"
  },
  {
    "speaker": "Speaker 1",
    "text": "working and what we're doing. But we also did also take on a client each so we could be like personally involved in how to deploy AI for clients and leveraging AI in the work that we do at"
  },
  {
    "speaker": "Speaker 1",
    "text": "CS2. So we don't don't have that disconnect. But for people that are working at companies where that's like less realistic, right? Or you wouldn't want your CEO to like just start getting"
  },
  {
    "speaker": "Speaker 1",
    "text": "too in the mix. How do you think that you I'll pass this question to Allison first, but Chrissy I'll get your take too. like, well, how do you how do you balance this? Because I think we see it"
  },
  {
    "speaker": "Speaker 1",
    "text": "a lot and maybe you could come up with some examples cuz I know we've spoken to both of you about this. There's a lot of CEOs that are kind of like like asking to hook up to claw code to Salesforce,"
  },
  {
    "speaker": "Speaker 1",
    "text": "like asking for kind of these dangerous permissions and things like that. And then maybe um maybe that's a good thing because they might understand how hard it is or break something and realize, oh"
  },
  {
    "speaker": "Speaker 1",
    "text": "[ __ ] I should hand this back to the experts. But also it's a bad thing because you know we need to keep control and governance of these systems but also they they might have unrealistic"
  },
  {
    "speaker": "Speaker 1",
    "text": "expectations on what is possible. So like give some people some advice or anything you can think of on how to how to approach this if you if you've got any ideas or anything that we've worked"
  },
  {
    "speaker": "Speaker 1",
    "text": "just managing expectations with execs in the past."
  },
  {
    "speaker": "Speaker 2",
    "text": "Um, I think well I think one thing that's really important is just to remind them of the risks involved. Um, and when when rolling anything out, making sure you're scoping permissions"
  },
  {
    "speaker": "Speaker 2",
    "text": "accordingly. Um, you know, I I know a lot of uh a lot of our clients are hooking into their Salesforce orgs uh via like MCP or via the API and having their agents look at look at the data in Salesforce. Um,"
  },
  {
    "speaker": "Speaker 2",
    "text": "making sure that that's like read only, that you're protecting PII wherever possible. But I think yeah, explaining the real risks involved is is key. Um cuz I mean there's a lot to be excited"
  },
  {
    "speaker": "Speaker 2",
    "text": "about, but there's a lot at at risk as well. So just kind of tempering that a bit."
  },
  {
    "speaker": "Speaker 3",
    "text": "Yeah, I would say from from my experience, so I had a very like just real practical use case where the one client like Charlie said I'm on like the CEO is like oh I want to connect to our"
  },
  {
    "speaker": "Speaker 3",
    "text": "sales force and um they were just like do whatever you need to do to make this happen because he's our CEO. So [laughter] um and and I kind of took a step back and was like okay but what you know and"
  },
  {
    "speaker": "Speaker 3",
    "text": "Allison kind of pushed on this too. It's like well I need to understand like what is what are they trying to do like what you know what objects are they trying to access like what are they are they"
  },
  {
    "speaker": "Speaker 3",
    "text": "trying to just pull data or they want to write back to um the system and so forth and so I think if you kind of have the like questioning parameters I think it's good at the same time also um we had to"
  },
  {
    "speaker": "Speaker 3",
    "text": "uh so I'll just use Salesforce example but try and do as much as you can to kind of control and get insight into like the monitoring after that. So we had created a um connected app for for"
  },
  {
    "speaker": "Speaker 3",
    "text": "that but also an integration user just for that connection so that we could see like okay if we see a spike in API calls that just tied to that user and we didn't give it any access to actually"
  },
  {
    "speaker": "Speaker 3",
    "text": "write data um which is we figured out that they didn't need in the end but um we could at least start to monitor and also let them know like hey you're actually like already running up against"
  },
  {
    "speaker": "Speaker 3",
    "text": "your API calls anyway, like it probably be careful around what you're going to do. Um, or else you're going to kind of take and they were like, \"But it doesn't matter.\" Like they were like, \"Oh, it"
  },
  {
    "speaker": "Speaker 3",
    "text": "doesn't matter.\" Like, no, it 100% matters. Like, do you want to take down your HubSpot? Do you want to follow up on leads that like you're sending over to sales? If you do, like you cannot hit"
  },
  {
    "speaker": "Speaker 3",
    "text": "your API limit. So, um, I think it was a definitely an education point. And since then also we've kind of um had they've had to actually talk about okay do we increase our APIs uh limits um internally as well and so"
  },
  {
    "speaker": "Speaker 3",
    "text": "forth because we actually do see that they're doing that not because of that user actually a different user which is also interesting. We haven't seen any activity from the one really tied to the"
  },
  {
    "speaker": "Speaker 3",
    "text": "CEO. So just another thing where you you maybe see like oh this is going to be a big thing and then maybe it's not. So >> it's a nothing burger. [laughter] Um, but yeah, I I I think there's a bit"
  },
  {
    "speaker": "Speaker 3",
    "text": "of that like use it as a chance to educate and put in like some governance, but also depending on the size of the company, like do allow experimentation, do allow them to do some of that work"
  },
  {
    "speaker": "Speaker 3",
    "text": "because I I I think you don't want to stifle uh that totally just out of the sake of governance, especially if it's a small company in my opinion."
  },
  {
    "speaker": "Speaker 1",
    "text": ">> Yeah, totally agree. um my post last week actually had a post on LinkedIn RIP Salesforce API limits that >> I think right now it's coming up to like 500 reactions and 70 odd comments. So I"
  },
  {
    "speaker": "Speaker 1",
    "text": "think that definitely resonated with people but um I would say the other thing that I I would just add on to this is that I think you just have to understand where the pressure is coming"
  },
  {
    "speaker": "Speaker 1",
    "text": "from maybe thinking about more often the expectation side. So the CEO isn't just making this up. Like they're getting pressure from other people like the board, the market, their peers."
  },
  {
    "speaker": "Speaker 1",
    "text": ">> So one of the best things I think you can do for leadership is to just to make them feel confident that you understand kind of the priorities and and feel that pressure too. And because they often"
  },
  {
    "speaker": "Speaker 1",
    "text": "people in that position they if they don't see something kind of obviously happening then they kind of like try to zoom in and try and figure it out and that's where maybe they get annoying."
  },
  {
    "speaker": "Speaker 1",
    "text": "But if you're proactively showing them, okay, I'm in marketing ops or I'm in go to market ops and this is every time we do something with AI, maybe we kind of announce it or we have like a portal"
  },
  {
    "speaker": "Speaker 1",
    "text": "where we're sharing content or we put it into Slack or in the QBRS we talk about it. I think just like the more vocal you can be about like where you see your le leverage of AI going and what you're"
  },
  {
    "speaker": "Speaker 1",
    "text": "doing practically on the ground floor I think then it just builds the confidence in your leadership team to like know that you are you are doing it and then they're not going to come maybe have ex"
  },
  {
    "speaker": "Speaker 1",
    "text": "extra expectations and then get all delusional about kind of what you can do and what my team isn't doing and kind of that kind of stuff. But then at the same time, like you've already mentioned,"
  },
  {
    "speaker": "Speaker 1",
    "text": "having some kind of education in there, right? Like, okay, we deployed this AI thing, but then maybe talk a little bit about what it took. Oh, we had to fix like three layers of data. We had to"
  },
  {
    "speaker": "Speaker 1",
    "text": "build all of these other processes. We had to implement a new tool. We had to get hook up, you know, integrations and fix integrations, but then we got this to work. So like try and bake in the"
  },
  {
    "speaker": "Speaker 1",
    "text": "exciting stuff around AI and like what you did and then like in that explain what it took to get there and I think just over time it's an educational thing. That's always been true to"
  },
  {
    "speaker": "Speaker 1",
    "text": "Allison's point. This is what we've always faced in ops and it's no different really now. It's just the expectations have maybe gone a bit further. Um so let's get into the meat of this and and why we wanted to bring"
  },
  {
    "speaker": "Speaker 1",
    "text": "Allison on to talk about analytics and revenue planning. So, um, just to kind of set the stage, Allison, um, because there's a lot of ways that we can leverage AI in analytics, but let's just kind of like"
  },
  {
    "speaker": "Speaker 1",
    "text": "zero in on this particular use case and just at a high level explain to anyone listening kind of what kind of the bounds are of this type of leverage of AI in this use case and what is kind of"
  },
  {
    "speaker": "Speaker 1",
    "text": "the ultimate goal and outcome that we're trying to get to with clients. just kind of explain it at a high level and then we'll kind of dig keep digging deeper into it."
  },
  {
    "speaker": "Speaker 2",
    "text": ">> Sure. Yeah. So, I mean I guess one thing that we've always helped clients out with is in preparing for you know an upcoming fiscal year um building building a revenue model or revenue plan"
  },
  {
    "speaker": "Speaker 2",
    "text": "based on historical data. So looking back at your funnel metrics, volume, velocity, conversion rates, and you know what your revenue goals are and kind of deriving well what what do we need to"
  },
  {
    "speaker": "Speaker 2",
    "text": "produce as a marketing team, as a sales team to reach our revenue goals. Um and then yeah, in the past oftentimes that would be done kind of quite manually, usually in spreadsheets. You'd build a"
  },
  {
    "speaker": "Speaker 2",
    "text": "bunch of formulas. Um, often times you would need to segment that model based on your key business segments cuz maybe selling to enterprise is a lot different than selling to SMBs and so you'd need"
  },
  {
    "speaker": "Speaker 2",
    "text": "to segment that way. um potentially segmenting by region. And so there'd be a lot of formulas and tabs and it would take, you know, weeks and weeks potentially, especially if there were"
  },
  {
    "speaker": "Speaker 2",
    "text": "changes that after you met with other folks. Um whereas um using AI to help with something like that. Um, for instance, I just did this with one of our clients where we went through seven full rebuilds of the model"
  },
  {
    "speaker": "Speaker 2",
    "text": "and we had to pull fresh data uh and have three different methodology changes throughout all of this. And um every one of those if we had done that by hand would be days of re redoing the formulas"
  },
  {
    "speaker": "Speaker 2",
    "text": "making hoping that you updated each uh each one correctly and you didn't miss something. Um but now it's just you know a a you you you pull the fresh data in you tell you know in this case we're"
  },
  {
    "speaker": "Speaker 2",
    "text": "using cloud code um tell tell cloud code exactly what what changes you wanted to make and then um it would re it would it would reiterate the Python script that created the model. um very quickly and"
  },
  {
    "speaker": "Speaker 2",
    "text": "easily. Um not to mention like we're able to pivot more easily on a bunch of different segments. It's able to find gaps that we have in the data. I think one thing I mentioned earlier in"
  },
  {
    "speaker": "Speaker 2",
    "text": "the podcast is how your data needs to be in a good state in order to do something like revenue modeling. So it would find things like, hey, you have a lot of deals that close in zero days and that's"
  },
  {
    "speaker": "Speaker 2",
    "text": "really skewing your velocity down. Um, maybe we should eliminate those from our, you know, average velocity or median velocity. So because our but the the the main deals that we're we're"
  },
  {
    "speaker": "Speaker 2",
    "text": "we're working on in this segment do take longer and so things like that that it can easily find more than a human."
  },
  {
    "speaker": "Speaker 1",
    "text": "Um so yeah >> so let's let's kind of walk through this cuz I think one of the things that a lot of people are really hungry for now is kind of a a playbyplay step by step like how exactly did you go about this? Like"
  },
  {
    "speaker": "Speaker 1",
    "text": "you mentioned claude code, you mentioned Excel, you mentioned formulas, you mentioned kind of the output like walk through. Okay, client comes to us says AllA Allison I need a revenue model now"
  },
  {
    "speaker": "Speaker 1",
    "text": "what what do you do next? Like and yeah, we'll keep digging into different parts of that, but like kind of walk us through your workflow both from like leveraging AI but also just like"
  },
  {
    "speaker": "Speaker 2",
    "text": "gathering the inputs that you need for this model too."
  },
  {
    "speaker": "Speaker 2",
    "text": "Yeah, I mean well the first thing is knowing what the goals are. Um so getting with them and understanding well what are the different segments of your business and what are what are the"
  },
  {
    "speaker": "Speaker 2",
    "text": "revenue goals and one thing to be aware of too when you're getting those goals is often times you actually you actually need to look farther in the future depending on how long your sales cycle"
  },
  {
    "speaker": "Speaker 2",
    "text": "is for goals. So like one of the clients that we built this for has very long sales cycles. So I actually needed to get >> what are you talking about very long?"
  },
  {
    "speaker": "Speaker 2",
    "text": "Like give us a sense for how you see like what's a very long sales cycle in this in this context. Do you think >> um in their case in many deals would take longer than a year. So their like median velocity was,"
  },
  {
    "speaker": "Speaker 2",
    "text": "you know, 9 to 12 months type of sales cycle. Um, and so when you have that and you're building a revenue model, especially based on velocity, well, you need to make sure that you're"
  },
  {
    "speaker": "Speaker 2",
    "text": "producing the deals in 2026 for 2027."
  },
  {
    "speaker": "Speaker 2",
    "text": ">> And if your goal is increasing in 2027, then you need to start producing those in 2026. So making sure that you know what what the the the current the next year goal is but probably the following"
  },
  {
    "speaker": "Speaker 2",
    "text": "year after that and if you don't know that at least kind of getting an idea of the percentage growth that the company is going to expect in 2027."
  },
  {
    "speaker": "Speaker 1",
    "text": ">> How are you getting velocity? Are you looking at opportunity create to close?"
  },
  {
    "speaker": "Speaker 1",
    "text": "Are you looking at sales ready to close?"
  },
  {
    "speaker": "Speaker 1",
    "text": "Are you factoring like STR time in that?"
  },
  {
    "speaker": "Speaker 1",
    "text": "Give us a sense for if someone wanted to get their velocity. How do we think about true sales cycle velocity?"
  },
  {
    "speaker": "Speaker 2",
    "text": ">> Yeah, so we're we're looking at um median or average velocity between different funnel stages. Um for this client in particular, they're like MQL to opportunity creation is actually fairly short where where the"
  },
  {
    "speaker": "Speaker 2",
    "text": "deals obviously take the longest."
  },
  {
    "speaker": "Speaker 2",
    "text": "There's once the deals created. So there's date timestamps that we have when opportunities reach each stage. Um so a lot of that velocity was driven off of those historical timestamps."
  },
  {
    "speaker": "Speaker 2",
    "text": "Um one thing that actually AI helped identify was what what I'll call like survivorship bias in some of the historical data. meaning that there were opportunities from like 3 4 years ago"
  },
  {
    "speaker": "Speaker 2",
    "text": "that are still open. Um that made the velocity look longer than it typically is. Um but yeah, so when looking at more recent deals, they are closing faster. But yeah, it's what it also highlights I guess too is"
  },
  {
    "speaker": "Speaker 2",
    "text": "hygiene as well. Like is it is the opportunity you created in 2021 that's still open really going to close this year or or not? So another thing we had to build into the model especially for"
  },
  {
    "speaker": "Speaker 2",
    "text": "them was what we're calling like a slippage rate where okay we say the close date is this this quarter but how often in the past so looking at stage history so how often in the past or"
  },
  {
    "speaker": "Speaker 2",
    "text": "sorry close date history how often in the past did the close date slip um how many opportunities on average that you said we're going to close this quarter actually closed in a subsequent quarter."
  },
  {
    "speaker": "Speaker 2",
    "text": "And it was actually a lot for them um because they have these long sales cycles and maybe they're more ambitious initially on the close dates."
  },
  {
    "speaker": "Speaker 1",
    "text": ">> So you're gathering all of all of this data, you're getting all of all of these inputs, then what what's the what's the next step to and and how do you use AI in these in these next steps to build"
  },
  {
    "speaker": "Speaker 2",
    "text": "out the model?"
  },
  {
    "speaker": "Speaker 2",
    "text": ">> Yeah. So um luckily I had done this by hand for other clients in the past. So I was able to give uh claude code an example um and explain how how to build a reverse waterfall model. Um so that was the kind of the"
  },
  {
    "speaker": "Speaker 2",
    "text": "starting point. Um and yeah, so I would feed it that I would also then pull the the raw data."
  },
  {
    "speaker": "Speaker 2",
    "text": "So um with sock like pull out like opportunity history, the funnel stage, the funnel stage attainment. So our custom funnel object um yeah opportunity history funnel object opportunity object pulling all the"
  },
  {
    "speaker": "Speaker 2",
    "text": "historical values and feeding that into the um feeding that into claude code along with the example and um yeah then it would derive uh the revenue model."
  },
  {
    "speaker": "Speaker 2",
    "text": "But there was lots of, again, like I said, there was lots of back and forth with like, oh, hey, there's these these deals that close in zero days that are like service renewals. Let's exclude those. Um"
  },
  {
    "speaker": "Speaker 2",
    "text": "um or, you know, there's these deals that have been that have been open for four years that is skewing the velocity high. How do we want to think about those? But then the the nice thing about"
  },
  {
    "speaker": "Speaker 2",
    "text": "that is in the past I would have to do that analysis manually and put it put in the exclusions manually where um with this you know it was not able not only was it able to surface those issues that"
  },
  {
    "speaker": "Speaker 2",
    "text": "I didn't identify out of the gate but then it was able to quickly you know remove them from the model as needed."
  },
  {
    "speaker": "Speaker 1",
    "text": "So you're going back and forth with Claude and Claude code kind of refining things. It's giving ideas, okay, these opportunities are skewing some of the averages and data. So we need to remove"
  },
  {
    "speaker": "Speaker 1",
    "text": "these. You're then saying, well, maybe we should look into this, look into that, and you're kind of going going back and forth versus you having to like run the report, kind of identify, filter"
  },
  {
    "speaker": "Speaker 1",
    "text": "out, cut them out the Excel, remove them from with the Salesforce report, kind of run it again. And that's kind of when you and then when you say derive the revenue model, kind of outline what it"
  },
  {
    "speaker": "Speaker 1",
    "text": "looks like to people and maybe we can get um a screenshot that we can throw up on the um on the screen here in production, but kind of for those who aren't watching kind of what does the"
  },
  {
    "speaker": "Speaker 1",
    "text": "the result look like at the end kind of what's the artifact that you're trying to build?"
  },
  {
    "speaker": "Speaker 2",
    "text": ">> Yeah. So, um, based on historical data, there's an essentially an inputs tab."
  },
  {
    "speaker": "Speaker 2",
    "text": "That's going to be all the the numbers that the model uses to then calculate what you need to reach your goal. So, those would be things like by segment, what's the average average or median"
  },
  {
    "speaker": "Speaker 2",
    "text": "deal size? What's the conversion rate between each of your funnel stages?"
  },
  {
    "speaker": "Speaker 2",
    "text": "what's the um velocity between each of your funnel stages by segment? Um and then things like the slippage rate."
  },
  {
    "speaker": "Speaker 2",
    "text": "Um so that's on your your your inputs tab. And then you have kind of your plan tab that then based on the historical data and the goals which the goal is also an input. then um calculates what"
  },
  {
    "speaker": "Speaker 2",
    "text": "you're going to need in Q1 of FY26 to not only re reach your goal for Q1 of FY26 but to plant the seeds for future quarters."
  },
  {
    "speaker": "Speaker 1",
    "text": ">> Um another >> get your pipeline goal in Q1 but then depending on your sales cycle it might be your revenue goal in Q1 of the next year is what you're saying right."
  },
  {
    "speaker": "Speaker 2",
    "text": ">> Right. Yeah. So like for instance this client you know they need to they need to create stage two opportunities at least two at least two quarters in advance of when those opportunities are"
  },
  {
    "speaker": "Speaker 2",
    "text": "going to close. Um another thing that we did as well is included current pipeline in the model as well. So we looked at well what what is already what already has a close date for the next two fiscal years and based"
  },
  {
    "speaker": "Speaker 2",
    "text": "on history what's the likelihood of those not only closing in in the quarter they that were saying they're going to close but um also what percentage of them are going to close based on"
  },
  {
    "speaker": "Speaker 2",
    "text": "conversion rates and baking that into the model because you know we're obviously not going to need to produce as much um if we were starting from zero because we're not starting from zero. We"
  },
  {
    "speaker": "Speaker 2",
    "text": "have some in pipeline already."
  },
  {
    "speaker": "Speaker 1",
    "text": ">> And then so you're having Claude Code create the Excel file too with those tabs with all of the formulas connected to then produce kind of like when you when you update an input if you want to"
  },
  {
    "speaker": "Speaker 1",
    "text": "update it live kind of all the formulas will then show the outputs. Claude code did that. give me a sense for so as you as you were going through it was it just like one shot create the Excel boom it did it"
  },
  {
    "speaker": "Speaker 1",
    "text": "perfectly or were there some little tweaks you had to make um and obviously over time we've done this with more clients so maybe we're getting better at it and or like instructing Claude and we"
  },
  {
    "speaker": "Speaker 1",
    "text": "got a more of a playbook around it but give me a sense for how well it did kind of first time with all that formula generation Excel tab generation etc."
  },
  {
    "speaker": "Speaker 2",
    "text": "I'd say it did a pretty good job. Um, well, one mostly because it did have an example of one that we've built before and >> that's probably a big deal. Worked well."
  },
  {
    "speaker": "Speaker 2",
    "text": "Yeah."
  },
  {
    "speaker": "Speaker 2",
    "text": ">> Um, >> cuz Yeah, I'm c I don't know if I was building it from scratch how well it would would have done. Um, but it did a pretty good job. But yeah, there were things that even I was like questioning"
  },
  {
    "speaker": "Speaker 2",
    "text": "like this the the velocity is zero days."
  },
  {
    "speaker": "Speaker 2",
    "text": "That's I know [snorts] that's not right."
  },
  {
    "speaker": "Speaker 2",
    "text": "Um and yeah, then it would surface things like oh well you've got all these service renewals that are zero days. Um so going back and forth there. Um there was also a lot of like client feedback"
  },
  {
    "speaker": "Speaker 2",
    "text": "too. That's the reason why we went through a lot of iterations because obviously they know their business a more closely than we do. So for example, we found that even though it wasn't tagged in"
  },
  {
    "speaker": "Speaker 2",
    "text": "Salesforce, there's a subset of opportunities that are larger deals that tend to close more quickly than some of their smaller deals. And so they wanted to segment out basically have a"
  },
  {
    "speaker": "Speaker 2",
    "text": "different model for those types of opportunities. So we had to like tag all of those opportunities in Salesforce, make it something that does get tagged going forward so we can measure that."
  },
  {
    "speaker": "Speaker 2",
    "text": "Um, and then we had to recalculate everything. And that's another thing where like if I had to do that by hand, that would have taken forever for it to then like be like, okay, we're"
  },
  {
    "speaker": "Speaker 2",
    "text": "rebuilding this and we're going to build, you know, half the model is going to going to be based off of this set of historical data and then half of it's going to be based off of this other set."
  },
  {
    "speaker": "Speaker 1",
    "text": "It would have Yeah. Just >> Yeah, that's where I want to go next."
  },
  {
    "speaker": "Speaker 1",
    "text": "Next with this. Where do you see cuz with AI there's always like AI can help save you time in certain tasks."
  },
  {
    "speaker": "Speaker 1",
    "text": "Also AI can help you do something better as well as save you time or maybe it's the same amount of time but you did it better. walk through that workflow and give me a sense for and give the"
  },
  {
    "speaker": "Speaker 1",
    "text": "audience a sense for where AI was most impactful maybe top three things and was it like an efficiency gain or was it also did something better that maybe Allison plus AI was like a superpower"
  },
  {
    "speaker": "Speaker 1",
    "text": "like give me a sense for where you see if people are going to do this obviously you've got a good framework here but like where do they lean into the most what what what can they expect from"
  },
  {
    "speaker": "Speaker 2",
    "text": "using AI >> yeah I think where AI really helps the most is so one is finding the the insights among all the all the data much more quickly. Um so yeah like the zero day service renewals surfacing surfacing"
  },
  {
    "speaker": "Speaker 2",
    "text": "those things. um you know it's something that I probably would have found but I it would have taken some pivoting and filtering and and things like that. So it's able >> yeah it just can do infinite slices of"
  },
  {
    "speaker": "Speaker 1",
    "text": "data so quickly and then just find the outliers right whereas you're in Salesforce doing like okay and let's drag this column here and group it by this and then and then look at it like"
  },
  {
    "speaker": "Speaker 1",
    "text": "this and then do another one and like you're just grinding through all of that. So do do you think that that that could be potentially both an efficiency and a and a performance game because do"
  },
  {
    "speaker": "Speaker 1",
    "text": "you think that because people are limited by time um maybe you don't find all of the things like AI right or do you think you would always have found everything that it found or do you think there's some"
  },
  {
    "speaker": "Speaker 2",
    "text": "certain things you would just maybe never have had time to find because you just couldn't do a hundred slices of data with the time available? Yeah, ex exactly. There's I know that there's"
  },
  {
    "speaker": "Speaker 2",
    "text": "things in the past that I have missed because I just Yeah. simply don't don't have the time and it's Yeah."
  },
  {
    "speaker": "Speaker 1",
    "text": ">> And that's not on you, is it? That's that's just like a reality of just like Yeah."
  },
  {
    "speaker": "Speaker 1",
    "text": ">> Do we really want to spend like a 100 hours doing this or like should we spend 50 or whatever it is? Like there's just a there's a trade-off there, right?"
  },
  {
    "speaker": "Speaker 2",
    "text": ">> Yeah. So like having it having it be able to do that. Yes. It's an efficiency game but also a performance game because you get a much more accurate model out of it."
  },
  {
    "speaker": "Speaker 2",
    "text": ">> Um I think yeah the other thing that it is a major efficiency game is just the speed at which like it can create 21 tabs with like thousands of formulas in you know 10 minutes where I'm like"
  },
  {
    "speaker": "Speaker 2",
    "text": "fiddling with the spreadsheet. Oh, like I know I'm I'm pretty good with Excel formulas, but I'm not like, you know, I don't know every single possible formula. So, having it be able to"
  },
  {
    "speaker": "Speaker 2",
    "text": "generate the the right formula. I mean, I will say there's bit there was a lot of back and forth in the beginning cuz I use like Excel for Mac and some of the formulas it created were not compatible."
  },
  {
    "speaker": "Speaker 2",
    "text": "So, there's some some things that we had to work through there, but um much still much quicker than um than me trying to figure out the best formula. Um yeah, and I would say the so that's"
  },
  {
    "speaker": "Speaker 2",
    "text": "mostly yeah, just an efficiency gain."
  },
  {
    "speaker": "Speaker 2",
    "text": "Mhm."
  },
  {
    "speaker": "Speaker 2",
    "text": ">> I think the the other thing is just the ability to change things very easily that normally would be like well now I have to like rewrite you know 100 formulas. Um >> yeah you're like sorry your client oh"
  },
  {
    "speaker": "Speaker 1",
    "text": "you want to change this okay give me a couple of weeks and I'll be able to change it. Now it's like okay tomorrow I'll have this done."
  },
  {
    "speaker": "Speaker 2",
    "text": ">> Yeah. and being able to slice things up because like this client in particular like >> we're like okay well why is the conversion rate this way? Well, can we look at it by industry? Can we look at"
  },
  {
    "speaker": "Speaker 2",
    "text": "it by sales rep? Can we look at it by >> this? Like one thing that came up cuz oftentimes we use what we call like um all conversion rates versus progressed conversion rates. So all meaning where"
  },
  {
    "speaker": "Speaker 2",
    "text": "the records that are still in the preceding stage haven't been >> have haven't had an outcome yet. So um we exclude them from the denominator in the progress conversion rate. Whereas in"
  },
  {
    "speaker": "Speaker 2",
    "text": "an all conversion rate we're just putting those in the denominator."
  },
  {
    "speaker": "Speaker 2",
    "text": "Um and so one thing that we did with this client they they still wanted to use the all conversion rates but what we decided to do was like different cut off based on the velocity. So what was the"
  },
  {
    "speaker": "Speaker 2",
    "text": "conversion rate you know 90 days after the quarter end versus 120 days after the quarter end and creating the formulas for that just like would have taken forever. But it was we were able"
  },
  {
    "speaker": "Speaker 2",
    "text": "to and so we now have this like conversion is the conversion rate mature for the stage two opportunities that reach stage two in Q4 or do we still need to wait based on velocity to know"
  },
  {
    "speaker": "Speaker 2",
    "text": "whether the conversion rate is mature."
  },
  {
    "speaker": "Speaker 2",
    "text": "Um, yeah."
  },
  {
    "speaker": "Speaker 2",
    "text": ">> So, that was one thing that really improved the model because we were able to >> to see like, well, what what is the most likely conversion rate based on velocity?"
  },
  {
    "speaker": "Speaker 1",
    "text": ">> Mhm."
  },
  {
    "speaker": "Speaker 1",
    "text": ">> So, so I have a couple of couple of questions I'm going to pose to both of you here. So, number one, um, so first off, thanks for walking through that."
  },
  {
    "speaker": "Speaker 1",
    "text": "Um, we've got one more topic we're going to go through and I think we might need to do a part two on this. If there's any questions, please send them to Russ and we'll bring Allison back on. But I think"
  },
  {
    "speaker": "Speaker 1",
    "text": "it one thing it it makes me think is um so often AI is in this category of where people are like well if we use AI we're going to cut down the workload and then people aren't going to have"
  },
  {
    "speaker": "Speaker 1",
    "text": "anything to do and people are going to get know laid off or whatever. And I think there's some companies using excuses to do that. But I think this is a great example of like you used AI, it"
  },
  {
    "speaker": "Speaker 1",
    "text": "made things a lot quicker, but you didn't just not you ended up just doing more on top of that. you you can do more slices with the data. Yeah."
  },
  {
    "speaker": "Speaker 1",
    "text": ">> You can do you can go deeper with it."
  },
  {
    "speaker": "Speaker 1",
    "text": "You can add on the um the >> what you just said there around like >> the the conversion rate type, right? And it's like I think the needs for businesses around like analytics and plenty of other use cases are are just"
  },
  {
    "speaker": "Speaker 1",
    "text": "infinite that I mean this is my maybe optimistic take. is infinite that even if you bring the time down to do so many of these things, it just gets replenished by just like doing even more"
  },
  {
    "speaker": "Speaker 1",
    "text": "stuff and being even better and having an even better model. Do you both agree with that or am I am I just delusional >> delusionally? Actually, the whole time that Allison was talking, I was thinking"
  },
  {
    "speaker": "Speaker 3",
    "text": "about like the reason why this is so powerful and and just knowing having done this before and you kind of and or see clients doing it is just your ability to just um like like we talked"
  },
  {
    "speaker": "Speaker 3",
    "text": "about like do the different slices but also inform it to do like a lot of scenario planning cuz like planning is great and having one overall plan is great but like also having like your"
  },
  {
    "speaker": "Speaker 3",
    "text": "different scenario plans but also informing that plan too and not 100% tied to your historical data is also where I see like really big efficiency gains because like [snorts] for for me"
  },
  {
    "speaker": "Speaker 3",
    "text": "when I think about like working especially with startups who are you know changing their ICP constantly uh changing like their value prop like proposition and and like even their key"
  },
  {
    "speaker": "Speaker 3",
    "text": "products like constantly you almost also need to get it to analyze your data first, your data set, your historicals, then inform it to say this is actually where we're going now. So, can you also"
  },
  {
    "speaker": "Speaker 3",
    "text": "build in all this new context to uh to to do that? And also, there's a lot of outside factors that then could like a a a head of sales or even a CMO might also then say, let's inform the model with"
  },
  {
    "speaker": "Speaker 3",
    "text": "this. So, you're talking about industry."
  },
  {
    "speaker": "Speaker 3",
    "text": "Maybe some industries weren't um you know a target for them in the past but now they are. or um sometimes in a way it's even tied to internal employee performance like hey can you look at"
  },
  {
    "speaker": "Speaker 3",
    "text": "like the tenure of a certain like AE and you could maybe look at like when they first joined you know your CRM or whatnot for that or maybe HR has that data but you could see okay for certain"
  },
  {
    "speaker": "Speaker 3",
    "text": "deals their velocity is going to be a lot you know shorter just depending on the tenure of the the AE and if you know you're going to grow your sales team out by you know x amount of people you can"
  },
  {
    "speaker": "Speaker 3",
    "text": "then have like that is just my like you know all our brains are exploding like how can I just then do get to inform the model and all of that well now with AI it's like you have this like amazing"
  },
  {
    "speaker": "Speaker 3",
    "text": "statistician at your like disposal but like doesn't have any of that context or doesn't really know so you still need to inform it but like the possibilities now are just like greater so to your point"
  },
  {
    "speaker": "Speaker 3",
    "text": "like yes we would have done a very basic model in the past everyone's like, you know, crossing their fingers, hoping and then they're, you know, sticking to these goals. Now it's like planning"
  },
  {
    "speaker": "Speaker 3",
    "text": "becomes a bit like it's it's it's smarter dynamic and you can actually have scenario planning that you then go back to as a business to say like hey you know we're adding on actually we're"
  },
  {
    "speaker": "Speaker 3",
    "text": "decided we're going to add on two more new reps instead and they're you know and so you can then like inform the model that way and I think that's the best part of it."
  },
  {
    "speaker": "Speaker 1",
    "text": ">> Totally. Yeah. And another question a question for you Allison. So another part of optimistic me today is that one you had to give an example two there was some work with AI you know and I'll"
  },
  {
    "speaker": "Speaker 1",
    "text": "break this up in kind of two questions there's one which is >> um you still needed expertise right this this was a what we would classify like a level two AI project which is human and"
  },
  {
    "speaker": "Speaker 1",
    "text": "machine working together to get like a really good output and it sped you up it helped you But like you still needed to figure it out. And the other thing and I would love to get your take on if you"
  },
  {
    "speaker": "Speaker 1",
    "text": "agree with that and you see that kind of longevity in there. And then the other question is around this isn't just like throw your data in and just say build me a revenue plan. It you're what you're"
  },
  {
    "speaker": "Speaker 1",
    "text": "doing is you're getting and then like every time you want to build a revenue plan you just like start again from scratch, right? You're you're building a tool, right? You're using AI to build"
  },
  {
    "speaker": "Speaker 1",
    "text": "like a deterministic system with inputs that will produce outputs, right? You're not, and this is where a lot of people get tripped up, like using AI, the probabilistic machine of AI, you can't"
  },
  {
    "speaker": "Speaker 1",
    "text": "always rely on that to just like generate a revenue plan every day. But what you can do is build this deterministic system for the revenue plan that you can then tweak and update and then you can iterate on that. It's"
  },
  {
    "speaker": "Speaker 1",
    "text": "kind of that again another blend of like yes human and machine working together but also deterministic versus non-deterministic kind of process and code and and what you're building"
  },
  {
    "speaker": "Speaker 1",
    "text": "working together as well. Do you want to take either one of those or both? But like anything you would add on to those agree with, disagree with or any thoughts you would want to expand upon"
  },
  {
    "speaker": "Speaker 2",
    "text": "there as you worked on this project?"
  },
  {
    "speaker": "Speaker 2",
    "text": ">> Yeah. Um yeah, I mean I definitely agree it's like in the the level two area. I do think having experience having done it manually greatly helped build it because well yeah one I was able to give it an"
  },
  {
    "speaker": "Speaker 2",
    "text": "example um which I think I could have done maybe without the example but would have had to had a more detailed prompt of what exactly I was looking for. Um, the the nice thing about what"
  },
  {
    "speaker": "Speaker 2",
    "text": "was built as well is that I mean essentially created I think about 2,000 lines of Python code that um yeah, basically I could pull the data again and have it regenerate the whole thing"
  },
  {
    "speaker": "Speaker 2",
    "text": "using using the the refresh data, which did actually happen a couple of times where um you know the data I was using was from 2025, but now we wanted to bake into more recent deals now that we're,"
  },
  {
    "speaker": "Speaker 2",
    "text": "you know, reaching the middle of 2026."
  },
  {
    "speaker": "Speaker 2",
    "text": "Um, and things potentially have improved. So, was able to just rebuild it using more recent data really quickly cuz the the Python script can handle that. And then, um, but yeah, some of"
  },
  {
    "speaker": "Speaker 2",
    "text": "some of the changes are more permanent where I I still had to tell it like, okay, from now on we, you know, we need to split split the model. um into those two different segments and so the Python"
  },
  {
    "speaker": "Speaker 2",
    "text": "script needed to be edited to handle that. Um yeah, so I mean I think it in the future it there's still going to need to be a human component when it comes to this because a lot of the data itself too is"
  },
  {
    "speaker": "Speaker 2",
    "text": "not always going to be in a perfect state. So, um, especially when you get especially when you get like a low volume. Um, I mean that's one thing that I've had to kind of coach Claude code to"
  },
  {
    "speaker": "Speaker 2",
    "text": "understand that like if something's if there's not enough volume of something, then we we don't want to we want to flag that as like maybe directional but not not going to be useful in the model or"
  },
  {
    "speaker": "Speaker 2",
    "text": "at least highlight that in some way that um we may want to take a closer look at this or have sort of a human gut feel override for this metric. because there's just not enough volume to to"
  },
  {
    "speaker": "Speaker 2",
    "text": "really say that this is going to be the conversion rate."
  },
  {
    "speaker": "Speaker 1",
    "text": ">> Yeah. Well, thank you so much for walking through that. Like I said, if anyone's listening to this and they want us to go deeper, let us know where where the questions are and we'll bring"
  },
  {
    "speaker": "Speaker 1",
    "text": "Allison back on because this is an an awesome use case um and a lot to unpack and continue to unpack as we do more and more. Um we're running out of time a little bit here, but let's let's do the"
  },
  {
    "speaker": "Speaker 1",
    "text": "final section. um speed round. So, uh a post by Harry Stebings, who he actually runs a podcast called 20BC that I listen to quite a lot, which I really like. Um this LinkedIn post though, not"
  },
  {
    "speaker": "Speaker 1",
    "text": "the you know, we'll see if we're a fan of it. I'll read it out for people not not watching. CLA will throw this up on on the YouTube video, but so I just interviewed a CEO. This is assuming on"
  },
  {
    "speaker": "Speaker 1",
    "text": "his podcast. I haven't seen it come out yet, but we'll see who it was. He said three things that blew my mind. Number one, we replaced our 600K Salesforce contract, $600,000 that is, um, with a"
  },
  {
    "speaker": "Speaker 1",
    "text": "vibe coded CRM built within three weeks."
  },
  {
    "speaker": "Speaker 1",
    "text": "We will get rid of 80% of the SAS we use internally. If Anthropic doubled pricing, we would not change usage in any way. So, let's talk about just the first comment, this first comment here."
  },
  {
    "speaker": "Speaker 1",
    "text": "So, Chrissy, I I'll pass it over to you first. What do you think? Should everyone be vibe coding their Salesforce away in three weeks?"
  },
  {
    "speaker": "Speaker 3",
    "text": ">> I think the question is, can anyone do it? Not should, cuz I still am not seeing any like receipts here. But when that podcast comes out, definitely share it with me cuz I want to listen. But um"
  },
  {
    "speaker": "Speaker 3",
    "text": "the the interesting thing is I I could potentially see that happening for a very small business that isn't so relying on a CRM that has like various integrations, a lot of systems and"
  },
  {
    "speaker": "Speaker 3",
    "text": "process setup. If you are basically just housing con contacts and the 600k part was just like housing a bunch of people that you're no longer talking to or whatnot, then I could maybe be like,"
  },
  {
    "speaker": "Speaker 3",
    "text": "\"Oh, okay. I could see that.\" There's probably a reason to be like, \"We don't need this tool anymore.\" And someone was like very much okay [laughter] with that. And so I could see that happening."
  },
  {
    "speaker": "Speaker 3",
    "text": "But I think for a majority of people, especially right now, I think that it it's we're just maybe in the future, but I just don't think right now is the case to just say you could just vibe code a"
  },
  {
    "speaker": "Speaker 3",
    "text": "lot of your tech stack. I think it's we were kind of we've kind of joked about this before but um that we even started filming but like you know there's like recently Anthropic was at an event and"
  },
  {
    "speaker": "Speaker 3",
    "text": "they were showing even the go to market tech stack and they they showed like lean data and Salesforce um being like [clears throat] a part of that and I think even a market automation platform"
  },
  {
    "speaker": "Speaker 3",
    "text": "and um cloud was kind of sitting in between some of that which is that is doable I understand um but that just shows that even a company who is at the cusp of AI literally building the models"
  },
  {
    "speaker": "Speaker 3",
    "text": "have like consciously and purposely decided to still use a platform to power their most important kind of processes um for the go to market team and there was a post that actually I think I"
  },
  {
    "speaker": "Speaker 3",
    "text": "shared and talked about in the past too for someone in Gong was saying like like there are things that you need for that it's like reliability and I actually talked about this on on a walk with you"
  },
  {
    "speaker": "Speaker 3",
    "text": "this morning, Charlie, where I was like, I definitely see the value of AI. I know it's going to change our our what we do massively, and we're talking about efficiency gains, and I I 100% see a lot"
  },
  {
    "speaker": "Speaker 3",
    "text": "of the efficiency gains, but when it comes to actually having AI power like all of our like systems and and go to market processes, I think of one thing that we will always need and and whoever"
  },
  {
    "speaker": "Speaker 3",
    "text": "is in Obsson has gotten a message from someone to say, hey, why did this work this way? what happened to this like you know that reliability is 100% necessary and we see AI hallucinating we see that"
  },
  {
    "speaker": "Speaker 3",
    "text": "if you're hooking up things in different with different platforms you can hit issues in in terms of token consumption um just API calls like just not you know running out of API calls uh sometimes"
  },
  {
    "speaker": "Speaker 3",
    "text": "even those like some failures just just purely out of like performance issues Um and so like and then if you're building a tool that uh we know that tools need a man need a product manager, you need to"
  },
  {
    "speaker": "Speaker 3",
    "text": "innovate them, you evolve them, you need even almost customer support. You're basically signing up your team to do all of those things for that tool that you're vibe coding for the go to market"
  },
  {
    "speaker": "Speaker 3",
    "text": "team. So I just I I I get it. I'm a bit concerned because I think someone made a comment on that post that I really like."
  },
  {
    "speaker": "Speaker 3",
    "text": "is like, oh, the things like CEOs will say to just get m money from investors."
  },
  {
    "speaker": "Speaker 3",
    "text": "And it's true. It's what we talked about. CEOs want to be looked at as AI savvy and and and building for the future, but I think we also need to be a bit more practical in terms of like what"
  },
  {
    "speaker": "Speaker 3",
    "text": "actually is doable from most companies."
  },
  {
    "speaker": "Speaker 1",
    "text": "So, >> Allison, you lead our solutions architecture practice that does all of our Salesforce work. So what do you think?"
  },
  {
    "speaker": "Speaker 2",
    "text": ">> Yeah, I think yeah, I think I think it's it would be challenging to build and support like a vibecoded CRM. Um for like many of the same reasons why like oftenimes we'll get the question of like our Salesforce"
  },
  {
    "speaker": "Speaker 2",
    "text": "or is a mess. Should we just start over?"
  },
  {
    "speaker": "Speaker 2",
    "text": "Um yeah, and often cases when a client asks that, you know, we're thinking about, well, if you're going to start over, you're gonna have to start over everywhere else as well. All the"
  },
  {
    "speaker": "Speaker 2",
    "text": "everything that integrates with Salesforce, you're going to have to start over. So, same thing like are you going to vibe code everything you're integrating with Salesforce? Um, there's"
  },
  {
    "speaker": "Speaker 2",
    "text": "also like the enablement piece in, you know, you've you've spent however much time enabling your team on using this tool. Are you going to be able to quickly ramp the team on your vibe coded"
  },
  {
    "speaker": "Speaker 2",
    "text": "CRM?"
  },
  {
    "speaker": "Speaker 2",
    "text": ">> Yeah."
  },
  {
    "speaker": "Speaker 2",
    "text": ">> Um, there's also security concerns. So, >> um, when you're buying a platform like Salesforce that's like multi-tenant architecture, like your data is secure."
  },
  {
    "speaker": "Speaker 2",
    "text": "Um, is it going to be secure with something that you've vibe coded? Um, essentially also like kind of what Chrissy said is when you're buying a product, you also get like documentation. There's feature upgrades"
  },
  {
    "speaker": "Speaker 2",
    "text": "that just come automatically. You get a support person. You can log tickets."
  },
  {
    "speaker": "Speaker 2",
    "text": "like um and often times like I I think especially when your company is already trying to sell its own product um you're also limited in time. So are you going to focus on building a CRM or focus on"
  },
  {
    "speaker": "Speaker 2",
    "text": "building your own product and selling that product? Like there there's only so much time. So >> Okay. Yeah, I completely agree with you both. I mean, I actually think this I need to listen to the podcast when it"
  },
  {
    "speaker": "Speaker 1",
    "text": "comes out, but I'm calling [ __ ] on this one for one cuz it's very easy to say they created a vibe coded CRM, but what actually did they build, right? Is it the equivalent of Salesforce? Because"
  },
  {
    "speaker": "Speaker 1",
    "text": "I would say even in 3 weeks, right? Like I don't I think that would be quite difficult to get something with like everything you talked about with like the level of security permissioning all"
  },
  {
    "speaker": "Speaker 1",
    "text": "your users in there all your data in there all of your integrations built out or have they just basically replaced Salesforce with just a data like they just threw all their data into a data"
  },
  {
    "speaker": "Speaker 1",
    "text": "lake and they called that a CRM and they have like a tiny little front end on it or something. So, like first off, like what actually did they build? And I think it the comment is designed for um"
  },
  {
    "speaker": "Speaker 1",
    "text": "like getting likes more than uh business reasons. Um and then like Allison, you talked about, I mean, it's the opportunity cost, right? If you're using >> first off, like it's not free to build"
  },
  {
    "speaker": "Speaker 1",
    "text": "this. You've probably got engineers doing it or someone doing it, right?"
  },
  {
    "speaker": "Speaker 1",
    "text": "You're using token costs. You're basically using this the most transformative technology of our lifetime to just build something that already exists. No, use that to go build your product to make it 100 times"
  },
  {
    "speaker": "Speaker 1",
    "text": "better. Why do you have engineers working on a CRM? Get them to be working on the product. I bet if you added in all of the cost of the people working on this, their salaries and how much"
  },
  {
    "speaker": "Speaker 1",
    "text": "they're going to put in how much they put into it in three weeks plus the percentage of their salary and allin cost for the people that are going to maintain improve support this assuming"
  },
  {
    "speaker": "Speaker 1",
    "text": "that it is actually a CRM and not just they're not just lying and it's just a database like literally with no front end or anything. um then it I I I I actually don't think there's much"
  },
  {
    "speaker": "Speaker 1",
    "text": "savings there. And if anything, if you then you add the opportunity cost of what these people could have done with that time and budget to then build your product, build your go to market,"
  },
  {
    "speaker": "Speaker 1",
    "text": "whatever, I think you just it's just a ridiculous thing to do. Um, and I'm willing to be proven wrong one day when Anthropic releases their new CRM killer model or something, but until Anthropic"
  },
  {
    "speaker": "Speaker 1",
    "text": "builds their own CRM, I mean, [laughter] with with their own tools, I think we're probably going to need like the off-the-shelf CRM. And I think um, there isn't much value in trying to vibe code"
  },
  {
    "speaker": "Speaker 1",
    "text": "them away at this point. And I think most people doing it are either setting themselves up for failure or they're just lying. And it and it's not they need to show more receipts to get my uh"
  },
  {
    "speaker": "Speaker 3",
    "text": ">> to get me bought in."
  },
  {
    "speaker": "Speaker 1",
    "text": ">> I know. I have so many questions. The more and more you think about it. I have so many questions. Like >> I'm going to listen to this podcast."
  },
  {
    "speaker": "Speaker 1",
    "text": "We'll listen to it and then we'll come when they release it and then we will chat about it on our next episode. I'm excited to dig in and and uh try and sniff out the [ __ ] in that"
  },
  {
    "speaker": "Speaker 1",
    "text": "conversation. Um but Allison, thank you so much for coming on. Um this was awesome. So, if people have questions for Allison, hit us up and we will respond and and bring you back on. And"
  },
  {
    "speaker": "Speaker 1",
    "text": "really excited to see everything you're doing with AI uh for our clients. So, and everyone listening, we'll see you on the next uh next podcast. Thanks for tuning in."
  },
  {
    "speaker": "Speaker 3",
    "text": ">> Bye."
  }
]