SaaStr 825: How the AI Era Has Directly Impacted Marketing and Sales with Snowflake's CMO and Founding CRO

0:01Welcome to the official Sastr podcast where you can hear some of the best Sastr speakers. This is where the cloud meets. Up today on the Sastra podcast. Let's talk a little bit about the use cases that we use AI for on the marketing organization. And on a daily basis now, 90 % of our marketing organization are using AI on a daily basis. We have about 450 marketers on the Snowflake team around the world. And we have seen in the range of 90 % time savings for a lot of different tasks that we've been able to use AI for instead.

1:04Hey everybody, it's Saster. FIN is the number one AI agent for resolving complex queries like refunds, transaction disputes, and technical troubleshooting, all with speed and reliability. See how FIN can deliver the highest resolution rates and highest quality customer experience at fin.ai slash Saster. That's fin.ai slash Saster. Hey everybody. Get excited. Saster AI London is this December, December 1st and 2nd, and we're on track to completely sell out with the playbooks for AI and B2B. Join 2 ,000 plus BI and AI leaders for two days, two days of practical advice on scaling in the new year.

1:46We'll have speakers flying in from OpenAI, Wiz, Clay, Intercom, Fin, and all your favorite B2B companies, including yours truly with Harry Stebbings and more doing our live 20VC podcast. It'll be fun all in the heart of London. don't miss out. Get your tickets and you can still go. Go to podcast.sasterlondon.com. That's podcast.sasterlondon.com for a special discount just for you.

2:12We are live. Hello, everyone. We're very excited about this next session. It is how the AI era has directly impacted marketing and sales with Snowflake. It's been a great day for Snowflake so far. This will be a very timely session. So, so excited. to have Chris and Denise here with us today. For those who maybe don't know our presenters, Chris was employee number 13 and the very first sales hire at Snowflake. He recently retired as CRO. He now advises Snowflake and is on the board for about eight or so companies, many of them AI startups, which he can talk about too. And after taking three other startups public, wow, Denise Denise joined Snowflake as employee number 120 back in 2016, and she is still actively the CMO.

3:04She's also spoken at Sastra before, so we're so great to have her back in the mix and bring Chris so they could do a bit of sales and marketing. Together, they have solidified the data cloud market category, which has resulted in many great things for Snowflake, including the largest software IPO in the history of Wall Street in 2020. Denise and Chris are going to cover a lot. They've got a lot of great learnings they're going to share today about getting into the IPO with sales and marketing and also how AI has impacted all that. So a lot of great learnings they're going to cover today. But for things they don't cover, for things you want to know more on, they just released Make It Snow.

3:41So this is the book to read, recommended to all of you here today. It captures all the lessons, the near misses, and stories in the book. It's available everywhere. But again, we're super grateful to have them here today to do a deep dive on how this has impacted in the AI area and the go-to-market lessons you can apply. With that, Chris and Denise, thanks for joining us. Super excited to be here with the Sester community today. It's always fun to be on the event. And likewise, thanks for having us.

4:15Let's jump into it. You've got over 10 ,000 customers, so let's kick off with setting the tone for the day. How do you succeed with so many customers and with AI in the enterprise? No, I think we wanted to start just by talking what are the things we are seeing out there in terms of which customers are succeeding with AI? How are they succeeding? And a lot of these learnings are applicable to us here at Snowflake internally as well. And the first thing is around company culture. Culture matters in a big way. It's really a make it or break it factor for AI success. You really need to have a culture of curiosity and an environment where people are really encouraged to experiment.

4:59And it's really been true here at Snowflake as well. And on the marketing team, we have an AI council with representation from every marketing function. And their job is really to go out and learn from others and test new use cases for their function. And on a quarterly basis, we host an AI day for marketing. And then the council share what they have learned with the use case that we plan to implement internally more broadly. And also all the tips and tricks for how to use, you know, Gemini or, you know, chat GVT on a daily basis. So we have that quarterly in a forum for the entire organization, market organization, to learn from the council.

5:46What we haven't seen working well is really to go out and ask everyone on your team. To tell everyone to go out and test new things, it really creates a lot of unnecessary deprecation of efforts and also chaos as well. So we asked all leaders to really identify those who are super interested and really curious about testing new use cases and new tools across all functions. Again, we're definitely seeing kind of the best ideas coming from, again, those who are the closest to the problem. But it's also important to have an executive mandate at the same time. If we look at Snowflake's customer base across the board, those companies that are the most successful are those who are really combining that top-down leadership with the bottom-up innovation.

6:38And if the CEO doesn't put AI as a top strategic initiative for the company, it's not going to be seen as a priority for employees either. So it's really important that AI needs to be one of the top priorities really coming from the CEO. If it is perceived as something optional, you're basically sending the signal that this isn't strategic enough for our organization. But the real transformative impact is really coming from when you embed AI into your core business and including it in the development of new products and customer experiences. You know, what we found, you know, AI hit, came super fast, like a fast-fitting train out of nowhere.

7:22And so at Snowflake, we had a lot of customers asking a ton of questions about how they were going to deal with this AI train. And so as we started to partner with our customers to help them deploy AI within their Snowflake environment, The thing that Snowflake was super hyper-focused on since our earliest days was making sure that we stored the data and protected the data that the customers want to do analytics in a centralized location in Snowflake. And then making sure that it was trusted and governed. And I think, you know, that's key to, you know, to this day of, you know, customers are, especially large enterprises are super nervous about sending their data out to random AI tools out there because they don't want personal identifiable information, PCI data to make it out to these AI tools that, that then can get published to who knows who.

8:25And so a lot of what we had the conversations from the earliest days is making sure that you had a data foundation that was ready for AI. And so really to this day, Selflake is partnering with all of their customers, 10 ,000 plus customers, on making sure that there is this solid foundation of a secure, organized data set that is really ready for the enterprise to then take advantage of AI. You know, AI is only as good as the data that it gets. And a lot of times customers in a large enterprise, Snowflake enables any large language model that the customer wants to use to get access to the data, but the customer has to approve those large language models.

9:11And so, you know, if an end user, you know, within a large enterprise decides to bring in a non-compliant LLM, Snowflake will lock that down and not let them do that. And that's really by design. And so what's super important is that the data that the customers are getting is something that is allowed to be accessed by the AI tools. And the quality of that data is good. Because also, using these large language models can get expensive. And so making sure that the tools are being used in an effective manner against the appropriate data. And so I think, you know, what we focused on at Snowflake is initially building a data warehouse.

9:56And now Snowflake has built this all-encompassing data platform that can do structured, unstructured, semi-structured data and allow customers to do, you know, not just analytics, but AI, apply AI across all of the different data sets that sit and reside within Snowflake. All right, so let's talk a little bit about the use cases that we use AI for on the marketing organization. And on a daily basis now, 90 % of our marketing organization are using AI on a daily basis. We have about 450 marketers on the Snowflake team around the world. And we have seen in the range of 90 % time savings for a lot of different tasks that we've been able to use AI for instead.

10:46Two of our bigger projects that we've implemented include two agentic models that are specifically built for marketing. And one is a campaign agent that is helping us run all our campaigns. It doesn't automate in every step of our campaign process yet, but it really provides real-time ROI data on every single campaign we have running. And it's helping us in real time to optimize all our channels, especially our digital ad spend that we can really optimize in real time, which has been pretty game-changing for us. It saves us a lot of money and have increased the ROI of our ad spend significantly.

11:28significantly. It also helps us with all our customer and prospect segmentation as well. In addition, also we have built a compete agent and it's really been developed both for us on the marketing side and the sales team. It really gives us real-time answers on how to position Snowflake in the most effective way in every compete situation. And it gives all the talking points back to the sales team if they're saying okay i'm competing against this company for this use case in this company the agent will give them the whole talking points in a back and that's pretty game-changing and i think many of you here in b2b can relate that it's really hard to enable both the market organization and the sales organization to compete against every single competitor at every single use case and at every industry level, for instance.

12:29That agent has been pretty game-changing for us. We also use AI for use cases like pipeline forecasting. That's a use case that we have been running for years. And also in B2B, the ability to be able to forecast exactly where your pipeline is going to be six months from now, that has been game-changing for us also many of you in b2b can relate right that you're suddenly in the quarter and you realize okay we don't have enough pipeline in let's say california this quarter and from marketing perspective it's pretty much you know too late at that point right to do something about that every initiative every program we're implementing now is really all about in a building pipeline for the quarters coming ahead that ability to really see where pipeline is going to be in the next kind of six months that has really allowed us to then completely reallocate and resources investments to make sure that every territory is in good shape also for lead scoring we're generating millions of leads now on an annual basis so the ability to score those in much more more granular and effective way that has been really game-changing in terms of just optimize our whole journey as well.

13:47Of course, like many of you, we're using AI for assisting us to create copy. We also use it for localization. Snowflake Marketing owns localization for the entire company that includes documents and for product and everything. So localization has been a big use case for us, both from cost efficiency and also from the speed of execution as well. Also, drafting interview scripts, video scripts. We do a lot of customer interviews. We have our own TV channel, Data Cloud Now. For them, they have seen that they're saving 90 % in terms of time savings for creating their scripts and preparing for interviews and everything.

14:37And also mentioned that everything around digital ad optimization, that has been really a critical use case where we now really in real time can see how channels are performing and we'll be able to shift to different channels in real time. And I also started out talking about our AI Marketing Council, what has really been instrumental to the success here, we started out by really identifying those folks that really leaned in a big way and that were really curious in terms of experimenting with AI. And that's about, we have about 30 people on that AI Council representing every function. And that also alleviates a lot of stress for the rest of the team but they know okay we have a team who's in charge of this on the marketing side i can go on with my day-to-day work and then again on an on a quarterly basis we have that ai day where we're kind of rolling out new tools and use cases to the team jesus i want to ask you a few questions that have come in from the chat and the live stream just on a few things you said here that's Okay, before we move on to sales.

15:49The AI council that you've mentioned, like who started it originally? And then how did you determine, you just mentioned it's cross-functional now. Like how did you determine the folks that should be, who started it? And then how did you determine the folks that should be on that council now? Yeah, it was something that was initiated by myself from the beginning, but it's led by Hilary Carpio, who was also runs in - She's been on SAS too, yeah. Yes, she's been on SAS too. Hillary is also a person who loves to kind of innovate with new technology. That's kind of, she's really passionate about that.

16:23So she was really the best leader to run this group. And then there were a lot of people that raised their hand. We announced, we're launching the council who would like to spend 10, 20 % of their time to really dive deep into this. And again, these folks that joined the council are similar to Hillary, right? They were really curious and excited about what AI can do for their function. So they get to spend about 20 % of their time just on looking at what are the use cases we should implement. They also collaborate as a team. Because also, you cannot just do all this in isolation. AI has impact on the entire team and also sometimes the entire company.

17:10so a lot of the new technologies we implement have to go through security review processes for instance it needs to meet you know or our governance and regulations you know you know as well so but again most of these people i would say all of them they raise their hand and we we send them to different you know conferences right they're attending events like like today so we also invest in their in in learning right for them to go out and learn from peers go to conferences and and then test new things and related question does the ai council also own all of the ai budget or do they just have a say in how like the budgeting and tools that are used yeah the good thing is that some of these tools and use cases are for free in some cases cases too so it's not always that there's a cost tied to it but there is the budgets is within the different departments so let's say if there's an application for the creative team that budget is most likely in that team so they it would be yeah got it makes sense for the agentic models that you mentioned so you guys have two one that is able to chat and with your marketing team more so and report on campaigns so you mentioned ad spend you mentioned coffee is that a proprietary model that you guys have set up or did you use any third-party tools to help you set that up at stuff like yeah that's a great question of this are proprietary models based you know using snowflake cortex and then we're using various large language models to build these agents that could be you know it could be anthropic you know it could be you know open ai you know and other models you know as well but we again we're of course a little bit unique situation that we are a data and ai company ourselves models but for that reason as well we need to push the limits more because a lot of these use case that we're developing we're also taking them to our customers as well but these models are developed by our we have our own intelligence team you know here at snowflake and we actually it's a shared team between all our go-to-market you know functions a year ago we had all we all had our own intelligence and data teams right we had one within cells there was one in marketing everyone has their own but also we saw that there were some duplication of work there as well there was some specific competencies that we wanted maybe use across the board so this team is now a shared team across all of gtm and they're really focused on helping us develop different you know agents for instance for different departments in a use cases but the other two these two are specific to marketing and Chris is going to share as well what we have on the sell side later.

20:15Yeah no that makes a lot of sense to consolidate not to duplicate efforts. I have to add I haven't looked at the dashboard in in a couple of months now I don't think I ever want to see your dashboard again the ability to just go in and interrogate your data I can now get answers to things that often had to slack someone hey can you look into why did this happen right or can you share this data with me and now i can just go in and ask absolutely any question of all our data and get you know answers you know back so it's very much like it's very much like chatty boutique but for all your internal information and we actually this product right that is a snowflake product is coming out in ga in november so there will be a big announcement on November 4th.

21:05Anyone who's using Snowflake is going to be able to use this agent for their use cases as well. Okay, yeah, I think it's great. You guys were your own first users, right? Okay, we'll build it for ourselves and then if it works and it's safe, we'll roll it out to other folks. We're always customer zero, right? Exactly. Yeah, there's a related question for that science team you mentioned that's now consolidated, that's building these in-house models. can you share a little bit more about the composition of that team is it mostly product is it more ai technical people do you also still have sales people on it to be more forward deployed can you share a little bit more about that yeah again this team is run by anita tasvi who's our chief data officer you know snowflake and and there are no there are no marketers or anyone from the sales team they work they have people assigned to work with us to truly understand right what are the business problems we're having right they need to really be they're embedded in our organization they don't report to us so i hope that answers the questions they actually worked within our teams before right so they they know the individuals on the team they have been living right you know the day-to-day work you know we do here both on the marketing and sell side but they are now centralized under one team they're mostly coming from a bi background or data scientists so it's a combination of data scientists and also product folks not snowflake product managers but product folks from from their developing data and ai applications right that we use internally so it's product data scientists and and analysts yeah to to denise's point like yeah we used to have a very siloed type of data teams within each group and and when sridhar came in one of the things he brought in was anidita as the chief data officer and we took any kind of data analyst or business intelligence people that were on my team and move that into the centralized team.

23:22And, you know, Anahita's job was still to support Denise and I from a business standpoint, but really we consolidated those sources. So there was not any siloed applications and stuff like that, which was really helpful as we scaled out the organization. Yep. A question that will take us straight into the sales use cases as well, which we'll get into next, and you guys can both speak to this, is there still a RevOps team at Snowflake or has this new intelligence layer replaced it? Yeah, yes, there is a RevOps team. There's still a ton of work for them to do, but the intelligence team is kind of there to support the RevOps team.

24:05So think of the RevOps team as the business stakeholders, the people coming to the data office saying, I need these things. And then they collaborate with the data office to make sure, and this intelligence team, they collaborate with the teams to do that. So I think that's all really been helpful because again, anything that Snowflake deployed in sales was deployed in marketing and marketing and sales were looking at the same data that maybe the CFO was looking at as well. Okay. So, you know, as Snowflake started to scale out our organization even more, you know, and I'll give Shridhar Ramaswamy, Snowflake CEO, a ton of credit on helping us, you know, pay attention to what's important from the pre-sales engineering team.

25:01We used to call them sales engineers and they're now called solution engineers. And one of the things that we did was, you know, Shridhar asked a good question. How do you know if your sales engineers are good? And so we, you know, which I couldn't really directly answer when he first asked me that question. And so one of the things we focused on was certifying every single sales engineer or solutions engineer all the way up to the senior, the person running the organization. So it wasn't just like the individual contributors in the, you know, running on all the sales calls. It was every senior leader, fourth line leaders and below had to actually get certified.

25:41And that was really important. That was an important first step on building out the organization and the technical credibility of the solution engineering teams. And then as we started to do that and we certified, you know, there were some people that were, that excelled at that certification, others that realized, oh geez, I need to, you know, up my skills. And that was, and we obviously helped them do that. I think one of the cool things is now we have a really technical pre-sales or solutions engineering team. And so in just six weeks, we were able to roll out cursor AI to create custom demos and custom content a whole lot faster across the entire solutions engineering team.

26:25And that allowed for iterations in the field. So as you all know, that AI is moving so quickly. So making sure that the people that are in front of customers, giving them tools that allow them to customize and change on the fly is incredibly important. And it's important that you have that skill set in the pre-sales team. So that was really a big thing. And again, I will give Tredar and Snowflakes, solution engineering leader, Modon, a ton of credit on really bringing that to bear and bringing that expertise in that. So, so as Denise indicated, Snowflake is always, you know, customer zero. And so, you know, as we looked at rolling out different solutions across Snowflake, you know, rolling out AI across Snowflake, we looked at areas where we could actually save time.

27:17And I think, you know, the way I view AI in a lot of ways, and I don't want to offend anyone, I view it as a task automator. I think it's something that if there's mundane tasks or things that humans have to do, I think AI is doing a great job of automating that. And I think there's, as Denise indicated in the beginning of the presentation, there has to be a return on investment. And so I think, you know, as we roll out tools, we have to actually see cost savings as well. Otherwise, it won't live in the real enterprise. Customers will buy something just because it's AI. Initially, they would, but over time, that's changing.

27:50And you feel that across the enterprises. There are some tools that are, hey, they're cool that I'm using it, but what's the actual business reason I'm using it? Is it generating more money or is it saving me more money? Those are the two things that, you know, ultimately we feel we're seeing in the enterprise. And that's important for everyone to kind of recognize. So, you know, as we rolled out, you know, our, you know, AI across our AI tools across our global support team, we saw 418 hours per week saved. That's a big deal. That's a huge deal. and allowing our support engineers to do things that matter more with customers.

28:27So again, on the task automation things, it was incredibly important. And as Denise said about dashboards earlier, Snowflake built a go-to-market assistant called Raven. And if you go talk to any sales leader, you talk to anyone that interfaces with a customer at Snowflake, Like they literally can go to Raven and ask questions of Raven of like, Hey, I'm going to see XYZ customer. Tell me about what's happening. Give me a 360 view on, you know, how are they? How's their consumption going? Are they a happy customer? Is there a bunch of support tickets? What opportunities are they looking at from a use case to adopt a snowflake?

29:13And, or what are the detractors of snowflake? And there's a lot of stuff because it's looking across structured, semi-structured and unstructured data. It's really giving the person that's interacting with that customer really a full understanding of what's happening. And so it's also a productivity gain because, you know, in prior lives, you'd have to go through multiple systems. You'd have to go to a dashboard over here. You'd have to go to Salesforce over here. And it wasn't as easy to find that information. But now instead of doing that, going to that dashboard, going to this, you know, sales force or whatever application, you're looking at this centralized tool that we call Raven.

29:50That's built on stuff like intelligence to really help give the sales team a more accurate view and real time view of what's happening in the customer. So it's important also to note that like, because I, and I talked about this earlier, because so like focused on security, like our customers are super worried about the security of their data. We also focus on making sure that these tools that we use internally, that we will roll out to our customer base. They're governed and they're highly secure. And again, I have to overemphasize that customers, especially a large enterprise care so much about that.

30:28So, and so does Snowflake, because we have these obligations to our end user customers, and we want to make sure that we're protecting their data and our sales teams are governed with these tools. So I think it's great to give a tool like Raven out to these customers and allowing them to interact with the data in real time, but in a secure and governed way. Yeah. And Raven is used across all departments by our leadership team as well. Also, another use case is, you know, our CEO, Shreder Ramaswamy, right? He probably meets with at least 10 customers every week. And the sales team would have to create this five-page briefs for him, right, before, right?

31:08What the ask is for Shreder, all the information about the customer. Shreder, he just brings up his phone, right, 30 minutes before. He asks the questions and he gets all the answers he needs about that account. and i think when a lot of people think about snowflake people probably think about you know data structure data that's for this right that was snowflake today you can query any type of data so you can query images you can query videos you can query pdfs you know documents so for with raven right you can ask any question of essentially all your intelligence in your company that's great before we go into your guys's closing thoughts so you guys both mentioned like ai tools across sales and marketing that they're using denise you've got this um this slide of you know 90 of your marketing team is using it daily how do you guys govern the tools that individuals are using because you also mentioned experiment you want to offer experimentation of how people can use ai so how much of that comes from snowflake and how much of that do you let folks implement themselves if they want to use it yeah again at snowflake first of all security is the number one thing for snowflake and keeping our data you know of course customers in a day it's the one thing that is the top priority you know here so like snowflake and many other larger enterprises we cannot just go out and implement you know any application right they have to go through security reviews here and that's why again it cannot happen in isolation here maybe they can go and experiment with different applications implement in production at scale it takes a lot longer time here and it's very similar right to many other larger companies and i just came back from ad week in new york yesterday the number one priority and that everyone talks about is of course you know privacy right the trust you have with your consumer in regards how you're using your data when they are actually being so generous with you giving the data to you absolutely that's really that trust you have between vendors and consumers there's gonna be there's nothing important here from an ai perspective so So no, we cannot just implement any application in production at scale.

33:33It takes time here to get those reviewed. There are many AI applications that are already built on top of Snowflake, right? Those we can use immediately. And what we're seeing more and more now is that those AI applications are being developed directly on the data. So there are hundreds of different applications just in marketing that are developed on top of Snowflake. And that is what customers are asking for, because especially for larger companies, they don't want to now their marketing departments, sales departments just to go loose and bring in hundreds of applications, right? It's a nightmare to manage from a governance standpoint.

34:13So we're seeing more and more of those applications, again, being built directly on Snowflake and distributed through our marketplace. One more question for both of you before you get into final thoughts, which is how have the AI era impacted hiring across sales and marketing?

34:34I think what's most important today is to look for people that are really always eager to learn, that are really curious about trying new things. things if we have learned something over the last couple years adaptability is a superpower of business today you need to be able to adapt fast you need to embrace you know change you need to be a lifelong learner and curious since everything look at skill sets it's more important to hire for aptitude you can learn the skills today if you're a curious life learner you can learn everything right it's changing so rapidly anything so the things you know five years ago are not even relevant today is that I'm looking for aptitude.

35:21I'm always asking people, how do you learn? How do you go out and advance your craft, right? How do you innovate yourself? Those are the questions I ask. Those are the type of people I'm looking for. And I would just add that I'm now officially an advisor to different companies now, and I advise a lot of AI companies. And what I'll tell you is that there's a ton, especially the younger people, they're super interested in working in companies that are AI relevant. And I think like, you know, of the folks that used to work for me at Snowflake, they're all, you know, Snowflake is a really AI centric focus place there.