Integrating AI Into the Customer Success Tech Stack

While customer-facing teams are racing to implement AI and start leveraging it throughout the customer journey, it falls on operations teams, like revops and CS ops, to help make sure they are doing so thoughtfully and carefully. A new technology like AI can add tremendous value, but it also requires an objective approach that considers how it’s used and potential implications. Hear what ops leaders from Okta, Greenhouse, Gong, and LinkedIn have to say about integrating AI into tech stacks and where they see the opportunity being unlocked.

Speakers

Bo Sun - Head of Customer Success - North America at LinkedIn [MODERATOR]

LinkedIn

Andrew Lefevre - Head of Talent Insights at LinkedIn

LinkedIn

Andrew Lefevre leads the Talent Insights function at LinkedIn, a global team that unlocks actionable ideas using data from LinkedIn’s 850 million members. Prior to LinkedIn, Andrew held a variety of leadership roles in HR teams, focusing on data-driven talent strategies that link to better business outcomes. Earlier in his career, Andrew was a finance and operations specialist, providing him the analytical toolkit and bias for impact that he brings to the talent domain.

Calvin Multanen - Manager, Customer Success Operations at Greenhouse

LinkedIn

Melissa Allen - Sr. Manager, Customer Success Operations at Okta

LinkedIn

Melissa Allen is a Senior Manager, Customer Success Operations at Okta. She has spent the last 4 years as Okta's Gainsight admin and CS strategic partner and has grown a team to support the ~450 Customer Success Managers and Customer Success Tech Strategists. She also partners with Okta's Digital Growth Customer Success team to enable successful business outcomes at scale through personalized, one-to-many programs. Prior to Okta, Melissa spent the last 15 years in operational roles spanning Customer Success, Professional Services and Development. Melissa and her husband live in Waynesboro, VA, where they also own and operate their own brewery.

Sonam Dabholkar - Director, Customer Success Operations at Gong

LinkedIn

Sonam Dabholkar is a versatile and results-oriented leader, with a skill set spanning Customer Success, Strategy & Operations, and Change Management. As Director of Customer Success Operations at Gong, she owns the processes, technology, and data that enable the Customer (Post-Sales) organization to create raving fans throughout the customer journey. Prior to Gong, Sonam led global programs for the Enterprise Customer Success organization at Autodesk. She was also an early employee at Gainsight, where she built and scaled their first Corporate & SMB Customer Success team. Sonam started her career in consulting at PwC, implementing Change Management strategies for Fortune 500 companies undergoing company-wide transformations.

Watch Integrating AI Into the Customer Success Tech Stack

Additional Q&A

would love to hear more about how Andrew's team built that account prioritization model -- what tools were used? what was the project plan?

Andrew Lefevre: At this link (https://lnkd.in/gpumYq6S) there's a summary of the project, and can download the PDF for more details.

‍

‍Transcript

This transcript was created by AI. If you see any mistakes, please let us know at marketing@matik.io.

Matik MC: Okay. Well, everyone, welcome to Matik, AI customer success summit. You are currently watching the integrating AI into customer success tech stack. I'm excited to introduce you to all of our speakers and moderator. And as I introduce them, I'm going to have them answer our ice breaker question today, which is what would you use AI to automate in your life outside of work? The first person is Bo Sun head of customer success, North America, LinkedIn, and he is also our moderator today.

Bo Sun: Amazing. Thanks so much Hannah. So for me, I am, an aspiring poker player, basically very highly interested but not very good. And today, the game of poker is evolved quite a bit where there's an aspect of game theory associated where you can play perfect poker. And so if I can magically turn into this coach, that helps me understand what I'm doing, right? What I'm doing wrong? I think I get a lot out of it. So that's what I would do focus on my hobbies and get better at poker hopefully.

Matik MC: That's a great answer. All right. Next step is Andrew Lefevre head of talent and insights at LinkedIn.

Andrew Lefevre: Right. Thanks for the good question, Hannah, for me, I really enjoy traveling outside of work and actually just got back from a great two week vacation in advance of those trips. I often spend a ton of time researching to try to get the itinerary just perfect and find some off the beaten path kind of restaurants or spots and takes a lot of time. And I would love AI to help simplify that. I know that there are actually a few AI powered travel planners already seems like early days though. I think AI is kinda trained on what you feed it. So maybe if I was to share some of my own personal travel photos, they might come back with like the perfect itinerary for me. So I'm excited for the future there.

Matik MC: Yes, I would love to find new restaurants like that. Thank you. Next step. We have Calvin Multanen manager of customer success and operations at Greenhouse.

Calvin Multanen: Thanks Hannah. Andrew. I think I might have some tips for you with the ChatGPT because that was actually the use case I was gonna use for travel. And actually I should show my manager Rosa gambler for giving me this, but you can basically use like ChatGPT or any other general I use ChatGPT, but you can basically say, hey, I want a 10 day itinerary for Athens, Greece and give me some restaurant recommendations, give me some, you know, things to do on each day. Price ranges, all that stuff. And so I use that actually for a trip that I'm playing later this month and I think it's still early days but it's pretty cool. And then you can obviously tweak it as one like say, I actually want to do this on this day instead. No, but I find it, really helpful especially if you're going to somewhere really, I have no idea what to do because there's so much to do so, but it's great for travelers.

Matik MC: Thank you. That's great. Next step. We have Melissa Allen senior manager of customer success operations at Octa.

Melissa Allen: Thanks Hannah. Yeah. So that's an amazing one travel. I went super simple and I was like, okay, what would I use? It? Just like every single day? Text messages, I know that sounds really simple but sometimes like you don't know exactly like the right way to word something or if you don't want to go and meet somebody, but you don't wanna say in like a nice way, I would just be really great to be like, okay, ChatGPT or whatever like say no, but in a really nice way and in a way in which they're not going to be upset with me like something super simple, but totally a daily use case for me, is, so I kinda went the more simple route on that.

Matik MC: No, that's a great answer. All right. Thank you up. Next. We have Sonam Dabholkar director of customer success operations at com.

Sonam Dabholkar: Thanks, Hannah. Andrew and Calvin also took my hands or mind was going to be travel but I'll add an interesting spin on it. So I was just in Italy in the summer which as we all know was like super crowded and had never been to Italy. And so I signed up for all these tours and went at like the worst possible times where the line was the longest and there was just so many people like crammed into all these museums. And so I would love for AI to suggest like tours and times for tours and days for tours where like no one has signed up. So I can just kinda leisurely walkthrough museums on my own and kinda take my time.

Matik MC: A great answer. It sounds like there's an idea for a company around this somehow with travel. All right. Thanks everyone. Those are great answers before we start the session today, just a reminder to add any of your questions into the Q and a section at the bottom of your screen and your Zoom controls, and we will answer in the last 15 minutes, any questions that come up. And I'm gonna hand it off to be our moderator.

Bo Sun: Great. Thanks so much, Hannah. And thank you all for joining for such an interesting topic. I think we've got a good discussion planned ahead and looking forward to your and a, so let's start with this question. Would love to hear from the group first what's your role at your company including what you're responsible for, and maybe share some of the top priorities for your department. Maybe we can start with Andrew. Alright. So I, I'm at LinkedIn and I lead the talent insights team. And so we're a team of data driven story tellers for the talent business. And so our customers are primarily corporate recruiting teams and corporate learning and development teams. And we say that we're a story tellers because, the team is technical and that they're querying LinkedIn vast proprietary data set with SQL or python. But then we package up, our insight is for a customer audience, an executive audience. And so, the mandate for the team is really twofold one is like ROI conversations and making sure the customers understand the value that they're getting from LinkedIn. And then secondly, it's to identify best practices using that large proprietary data set about how customers could hire better or up skill their employees in a better way. In terms of priorities for this year, we're going to be actually focusing on trying to map our customer objectives to our insights. And so what we've discovered is that our customers come to us not just to hire or up skill but to do a variety of things within that realm, whether it's diversifying their workforce, building competitive skills, like AI, skills, promoting internal mobility, et cetera. And that the team has developed a really large library of assets over the years. And so we want to do is kinda map the objectives to, the insights so we can more rapidly serve our customers, identify which insights are working better than others et cetera. So that's me and are year ahead?

Sonam Dabholkar: Great. I'll go next. Hi, everyone. I'm so am, I am currently the director of CS ops at Gong and in my role, I'm really the ops business partner for our chief customer officer as well as the teams that he manages. And so really kind of focused on strategic initiatives and those key priorities across the whole kind of post sales function and then cross functionally as well. In terms of, you know, what we're focused on, I would say there's a couple of different things. So one is really driving meaningful customer adoption of what we call our advanced capabilities. So these are really our core kind of stickiest features, our most strategic features that we believe will generate the best outcomes for customers using our product. And so our whole kind a post sales motion is really focused on really up leveling these motions and delivering against those advanced capabilities. So customers are kind of adopting those in a healthy way. The next thing we're focused on is really kind of up leveling our customer journey to support those advanced capabilities. So looking at what are the stages that we need in the customer journey? What is like the exit criteria for each of those stages? What kind of account teams are involved in the different roles that are involved in each of those stages? And then meaningfully kind of creating those motions to really drive, those milestones that we want customers to hit to then ultimately renew and expand with us. So really kind of taking a holistic look at revamping that customer journey. And then also this year we, we've launched some really big products and are really becoming like a multi product company. And so, looking at our post sales motion and defining, you know, as we roll out a new product, how do we embed that in our motion? What are the playbooks that our account teams can use? And then how do we really measure, you know, the milestones and the metrics to make sure that customers are adopting that specific product in a healthy way and then really becoming a true kind of multi product company, and managing that playbook for future years where we're launching other products.

Melissa Allen: As well. I love that stuff about the customer journey. I feel like that is just so so important. So that's awesome. Send. All right. So, I'll go next I'm Melissa, Alan, I'm a senior manager of customer success operations at acta, my team and I support our customer success team, our digital growth team. We kinda have a large scope. We also support our technical account managers that we call Tam and our tech strategists, and throughout how we support them is we help us comp plans, headcount, modeling, forecast planning, and we also own an admin our GainSight instance in which we try to utilize it and create call to actions, and those playbooks, and try to define our own customer journey as well. Some of the top parties of my department this upcoming year is scale. We have a huge focus on the digital customer success growth team, along with automating motions for CS and our team. You know, how can we be more efficient and give them time back to really work on those relationships? And another party is providing them with clean and helpful insights, so they can focus on those relationships with the customers with the knowledge that they're supported with the right data and the tools, and have things like a customer journey that they can view upon and then go in and talk to the customer and have that support to do so. So, we have quite a few things, on our docket for this year as well. So.

Calvin Multanen: Well, and I'll wrap things up, Melissa. We should definitely chat about some of the scale stuff. Yeah, call multi manager of CS ops at Greenhouse very. I think pretty similar role, the kind like laid out but at Greenhouse, I support our customer success team. So our managers or directors or VP basically helping them with all the kinds of day to day operations. You know, whether that be stuff like capacity resourcing tools, day to day admin, and then more strategic initiatives as well. Some of the big I think focuses for next year. And also, I guess the rest of this year is, you know, obviously customer retention, right? Like it's a tough market especially in tech and especially in hiring. So how can we make sure that you know, our CSM or have what they need to make sure that their customers are successful and, you know, in a challenging environment and also like, how can we work with our customers, make sure that they can kind of weather the storm? And, you know, that we're still a strategic partner with everything that's going on. And I think also another big initiative kinda tied to that is how can we give our CSM better signals, better kind of insights into when they should be engaging with customers? So… about a year ago, we kinda revamped our customer journey and we've really got that kinda nailed down but, you know, how can we continue to iterate on it, improve upon it? And so that our CSM are really gauging with our customers at the right moment providing, you know, the right, you know, recommendations, best practices and ultimately make our customers successful.

Bo Sun: So, I think a common theme across the board is that there's quite a bit of things that we can do to help our companies be more effective and efficient with AI. We talked about, you know, improving our customer journey, we talked about understanding customer objectives. We talked about scale. Is there a particular part of the CS workflow that you see AI having the most impact and maybe even peeling it back a little bit more. What are you most excited about? And what are you most skeptical about? So, Andrew, maybe we can go back to you to kick us off.

Calvin Multanen: Yeah.

Andrew Lefevre: I think the one use case that I would highlight in terms of a particular part of the CS workflow that I'd wanna spotlight something hat we've partnered on with our data science team and that was trying to assess our customers health at like huge scale and with a lot of accuracy. And so we've been using AI scores and predictions to identify which customers are going to churn or which customers are right for an upsell or cross sell. And so this account prioritization work was both like a huge time savings for CS ops teams or even our sales reps and CS reps versus kind of like the traditional way of doing a lot of this manually and spreadsheets and things like that. And then the accuracy was just so much higher because you could train the model on years and years of product usage data and data. And so it's been, you know, a big win for the team because it helps our reps use their time more wisely, they get in front of the right customer at the right time, and then we get really good commercial outcomes as a result of it. So, so that's good. So that's like what's exciting for me is like the ability to drive really, you know, efficient use of time and talk to customers about the right things that they're interested in. In terms of skeptical. I don't know. There's some doom and gloom articles out there about how, you know, AI is just going to be kind of a net negative around for jobs and talent and I think that might be overblown to someone that's been through a number of tech cycles. I think AI is gonna certainly be able to delegate and take over some elements of, our roles. But I think it'll also play a role and supplementing the type of work we do. And then there's always, I think some element of our jobs that will not be going away to a.

Matik MC: Yeah, I'll…

Sonam Dabholkar: Plus one, the health score and prioritization answer that is a huge kind of area of potential, for AI and CS. I think the other area especially in CS ops, we get really excited about data and how to use data to make CSM lives more efficient and help them be more effective at driving customer outcomes. And digital customer success is a very big topic these days. And I think in using data to inform what a customer's digital journey can look like and how that digital journey becomes very personalized and very targeted and very prescriptive and actually complements, you know, the human lead journey that the CSM is taking the customer on. How can we create the super effective digital customer journey that, you know, drives to customer outcomes even faster? Something that I think is really exciting and I see huge potential for especially now, I think the thing that I'm most skeptical of is like there's so much importance and I don't think this will ever go with like importance on human connection and especially in the customer success role. You know, you're responsible for really building that trusted adviser relationship with your customer, you know, getting to like that place of vulnerability where they tell you their business challenges, their pain points, they trust you to solve them. And I don't think like AI could ever replace that. I think AI could be the foundation on which a CSM could get certain pieces of information that would inform, you know, how they become a strategic advisor to their customers. But I hope that AI would never replace that human element that I think we still need, in customer success.

Melissa Allen: And I'm definitely gonna plus one what you said about digital customer and about Andrew, and the health score because, those are such big things and they can make such a movement with those two use cases. I think sometimes when I think about stuff, I almost think of like there's this huge big wins. But then what about like a daily win? And I know a lot of our CSM, they get questions from their customers saying, hey, how's another customer similar to me? Similar to my use case with, you know, how much I spend with Octa, what, how are they doing what their journey look like and just trying to dig and find comparisons? And almost like a benchmarking is super tough. And a CSM can spend hours days, weeks even to make sure that if they're gonna compare and talk to that customer say actually you're like you're on track man. You're doing great or, to actually have the ability to tell them how they're doing it's. Really tough. So, I'm almost thinking just even, on a daily use case of being able, to utilize AI to say, find me a similar customer with a similar use case and show me what that looks like and how my customers doing against at least 10 different ones. You imagine that would be like in a second and versus days or weeks that it would take a CSM. So then you've got a CSM trimming down weeks of their time to have that conversation, that really valuable conversation so much more quickly. So I'm thinking like these huge wins. And then you just got these almost like daily wins that you could use as far as kind of what I'm a little skeptical about. I'm almost afraid that we might get too dependent, right? Not not so much is like, you know, it might take certain pieces of the jobs or something like that. But what if we get too dependent on it and potentially we lose that relationship feel we're utilizing AI to answer certain emails or, to have these analysis drawn up. But we're not actually connecting with that analysis anymore because we don't have to because it was kinda built for us. So, so I do worry that we might get too dependent and lose that connection that we have with the customers versus it kind of taking over. So, I think if we're just aware and, we maintain and have it as a supplement and help it help utilize our jobs, but not take over that relationship piece, of the CS team and stuff like that. I think that would hopefully help keep that. From happening.

Calvin Multanen: Yeah, big plus one to everyone. Everything that everyone already said about kind of, the big areas. I think continuing the theme of kind of the smaller ones. Obviously, there's big wins around stuff with like, you know, digital success and, you know, using, you know, better smarter models. I also think like there's a huge opportunity for something like, the day to day admin, right? Like I think stuff like we use GainSight, we also use Gong. I think there's a ton of really awesome like time savings that we've already saved from like just a GainSight the Gong integration. But also stuff like that will automatically like track stuff around like sentiment or types of, you know, instead of having a CSM like have to like live a bunch of fields are, okay. They talked about this. They talked about that, talked about this having something that can actually just grab all that instantly put it in the right report, put it in the right structure. I think it would be a huge win or it is already kind of a huge win for our team because that's time that they don't have to spend just like fiddling around with our systems that we set up so we can report on it. It's more time they can spend on like actually talking to customers, understanding their needs, understanding their challenges and really providing more strategic value. So, I think that's a huge thing. I think also on the ops world like there's a ton of just like admin stuff that we all have to deal with. And I would love to be able to automate some of that as well, right? But I could just be like, I want to report that has xyz in it and like it's just kinda puts all together instead of having to like drag and drop columns. That would be huge. I think my, the thing that I'm skeptical about is just like accuracy, right? Like, we see stories all the time about like something about, you know, somebody asked ChatGPT, and like I gave a crazy answer. I think, you know, especially when it comes to dealing with like internal company data or especially customer data, you know, how do we make sure that whatever tools we're using, are giving the right outputs especially this data that we're then gonna go present to a customer, right? Because the last thing I want to do is put my CSM in a situation where they present something? And then customers like, that looks wrong because that instantly, you know, creates a, you know, a new obstacle but they have to deal with more work for them and then also more work for us and also more work for like, our data team as well, like try to understand, you know, sometimes it comes down to it and user error, right? Maybe they put something in wrong or maybe there's something wrong with their data pipeline. But, I think, you know, there, there is just kind of that barrier, right? And also kind of that black box. We don't really know, we know kind of conceptually how it's working, but we don't really know, right? So I think that is still just kind of thing I'm always skeptical of. I think would require a lot of testing on our end to make sure that we're really confident, but I think there's so much potential for tools like they have.

Bo Sun: Yeah, plus one there. So unlimited potential. But let's get there, with a hint of pragmatic skepticism. So, you know, overall, we talked about use cases across the board, but I think what people are looking for in this particular session is maybe some really specific examples of how companies have started to think about and maybe apply some of the powers of AI. So Calvin, I'm gonna go back to you. Maybe you can share specifically at Greenhouse with some of the things you all are focused on.

Calvin Multanen: Yeah, I'll come out and say, I think we're definitely taking a slower more temperate approach with AI at Greenhouse. I think part of that is just because of the types of data that are, that we deal with like around like candidate information and stuff. There's a lot of PII in there. So we're not, you know, itching to like plug that into some new tool and like see what happens. I think the, one of the things that I kind of resonated with me that Andrew talked about was, you know, using AI and kind of she learning in various algorithms to like build a better cohort model that's something that we recently did to help our customer teams understand which customers should we be focused on, like which ones are like low touch. They're really, you know, they're not writing into support. They're not really engaging with like any of our materials like kind of low risk long our.

Matik MC: Are…

Calvin Multanen: Versus what are the, you know, what are the ones we should definitely save, right? Strategic accounts like that that's or, you know, maybe it's something that, you know, they're actually more middle of the road. They don't need a lot of attention, more like a renew. I think that's something that we're now starting to play with especially on our scale team where they have like, you know, hundreds of customers in their books. But I think stuff like that is definitely something that we're really, you know, invested in and exploring to kind of roll out to our other segments later on this quarter. And I think also so I would say that's probably the biggest most recent practical use. It, I would say, we do have like a company like policy on generated I. And so we definitely encourage CSM to use stuff to like write e-mail templates, or exact summaries or stuff like that. It's really going to be anything that uses like internal company data or customer data, but we're going to restrict to existing vendor relationships. And I think one of the things we're really excited to see is a lot of our vendors are rolling out their own AI features, which makes it easier because we don't have to go out and the whole evaluation process and it's too. It's like it's already a trusted vendor. We know that they can work with our data and it makes it much easier to implement because it's already integrated into our tech stack. So we're definitely excited to kind of be able to start playing around with those. I know there's a couple in the works that I'm personally really excited to text that automatic. So we'll see how that all plays out. But yeah, definitely more to come. So I…

Bo Sun: Great, Andrew. I know that there's a few things that is on our radar. You care to share some specifics. Sure. Yeah, I mean LinkedIn been using, I did a little research for this to been using AI since 2007. So if you've been on the platform and never seen a job you might be interested in or a recommendation for a person, you might know to connect with, those are all outputs from AI models this year. What's changing at LinkedIn is we'll be a little bit more over about using AI with our members and customers to let them become more productive in using the products themselves. So for a corporate recruiter or customer, they're going to be able to use a generative AI writing assistant to be able to reach out to more job candidates more quickly, but also with more effectiveness because we know how to reach certain candidates with, of certain messages. So things like that will allow our members and customers to just use, the product a little bit more smartly by taking all the best practices that LinkedIn is developed through machine learning through AI, and providing, those features and those nudges to our customers and numbers. Thanks Andrew. Shifting gears a little bit here. I wanna make sure, we do spend some time to cover how we are incorporating AI into the rest of our tech stack. And so maybe Melissa we can start with you, we'd love to hear how you're thinking about this a doctor and maybe learn from, you know, overall how does this fit in? And maybe some of the things that you're expecting this to displace or replace as part of your current operating procedures?

Melissa Allen: Sure. So we're trying to be really diligent and our quest to use AI because, you know, we're at is all about identity and security. So we're just super like taking very small baby steps, and really trying to do our due diligence to make sure we're betting everything that we're doing. So there might not just be one AI tool kind of how we're seeing it is that potentially we might use.

Calvin Multanen: Multiple…

Melissa Allen: AI tools but in specific kinds of areas, so not like linking them or anything like that, but basically depending on because there's so many AI tools out there depending on our specific use case or need. Maybe we have a specific one that we utilize save for support cases. Can you imagine how much time that would save if we could say, has anybody ever had an issue kind of like this? You know, because maybe it's not tagged a certain way. And then I can just go and find all, the ways you've solved it in the past. So having specific use cases like that, maybe we have an AI tool just around our support cases in service cloud, right? And it's very unique to that and it doesn't reach other areas of the business. So we're really trying to be careful about where it should go, which ones fit the need of each kind of use case the best. And like I said, I don't think there's going to be one AI tool for us that fits everything. So we're being really careful to vet those tools and I don't necessarily think it's going to replace, but I think what it's going to do is say for, the support key senses. It's going to just trim down massively, that time of a support analysts trying to figure out if we've solved this issue before because maybe it's we, I mean if it's a very interesting use case, maybe it was dealt with two years ago, how you can find that quickly. So being able, to look at those different use cases, find the tool that's right for that need and plug it in that way. So I definitely think we're a little far off from actually plugging AI right into Octa, but we, we're taking that time that due diligence to see where it makes sense. So we can really do that analysis.

Sonam Dabholkar: Yeah, I love that thoughtful like use case based approach, Melissa at Gong. So Gong as been in the AI space, for a number of years since 2016 and so at Gong, we like to say, you know, we drink our own champagne and we, in the past have released a bunch of features and are now this year, you know, renewing our focus on that and releasing a bunch of upcoming features. I focused on AI really with the goal of applying AI to a lot of just the go to market teams motions. So examples are, you know, using smart trackers to track initiatives and concepts across more than just keywords, but, you know, across, you know, a number of kind of concepts and really looking at things like, you know, is the customer asking for a discount? Okay. Let's track that as a broader theme versus just those key words, looking at, you know, recommended contacts that analyze win loss data and actually suggest like, hey, you should contact this person because, you know, based on the past, like 1,000 deals, this seems like the right job title to contact. And so surfacing that ahead of time in a deal with generate, I specifically creating e-mail templates which I think is applicable for like sales and CS, right? Like we're always emailing our customers, sending follow ups after meetings, documenting action items, next steps. And with Gong, you know, a lot of, that functionality is now in place. And so it's really, I think improving the workflow, and making teams more efficient and kind of reducing the manual work. There, one thing that I would love to see is, you know, CSM are often tasked with proactive steps like, you know, making sure a customer is hitting those milestones on the customer journey. But what about kind of the reactive things, right? So, if there's a customer risk, if there's something that's unanticipated, can, you know, AI really surface a playbook for the CSM to apply that would really kind of be the most effective playbook for that scenario based on, you know, past experience, past data that's been put into the model. So just stuff like that as we start to kind of like crack the surface of like what Gong can do? I think the future is like making ae workflows and CS workflows are just a lot more prescriptive and a Gong, you know, we like to drink our own champaign as I said, so that's really our focus here is like kind of pushing our own product to the limits in terms of AI versus looking at other tools to add to our tech stack at this point.

Bo Sun: I love that approach. And overall just I wanted to think all the analysts here for sharing. I think we're at the precipice of a major, just innovate period in our space and hearing about the specific initiatives, the approaches, the philosophies that we're all taking as well as the potential risks concerns was helpfully helpful. I certainly learned a ton from each today with that for the last 15 minutes. I'm gonna pass it back to Hannah for Q and a. And thanks again everyone for the great questions.

Matik MC: Thank you both. I really appreciate it. And I'm gonna go through any questions that were submitted in the Q and a section. If you would still like to submit some, please do so and we'll go through them. The first question is for Andrew, this person would love to hear more about how your team built the account prioritization model and what tools were used and what was the project plan? Like?

Andrew Lefevre: Got it. Without getting to too many specifics. I would actually direct people to go to if you go to LinkedIn, dotcom, and then search for Crystal candle, that was actually the name of the project. Not not a person's name but Crystal candle and you can kind of learn quite a bit of detail on the different machine learning models that were used, to create that. And then we published an abstract with Cornell so that everyone can kinda learn from our success. The other kind of fascinating fact if you go there was that initially, there was some skepticism like despite us doing an AB test and saying, if you use these scores, you'll get, you know, an eight percent lift in revenue. The, the field was a little bit skeptical and said, hey, I think I know my customers better than some black box. And so what this resulted in was a project to actually explain what was going in going on inside the black box and surfacing the drivers for why those scores were showing up the way they did. And that actually led to quite a bit of improvement and adoption by the field. So really fascinating project. So check it out if you want to learn more on LinkedIn. Dotcom.

Matik MC: Thanks, Andrew. All right. The next question is from Calvin for Calvin at with Greenhouse, how, what tools is Greenhouse using to build the cohort models or accounts that you mentioned earlier?

Calvin Multanen: I would probably, I think Andrew's probably got a better real world example of that because we don't have a website for it. But yeah, it was, our data like analytics team basically help put together using machine learning and looking at like a ton of different variables and across that. So can't speak to the specifics of it. But we're not piping it into like all of our key tools like GainSight, Salesforce, Zendesk, and that's really kind of driving… really kind of behaviors and actions across all of our post sales team. So whether that be CSM like understanding which customers they should be reaching out to or like support reps, understanding, okay, this is a high priority customer versus a priority. It's actually gonna change kind of how our sites work there. And then for am as well. It'll kind of change how we approach renewal conversation, start to think potentially about automating some of those as well. But yeah, I wish I could speak to more of the particulars there, but I'm definitely gonna go check out Crystal.

Matik MC: Wonderful. Thank you. All right. Then next, we have a few questions. I'll just go through one each one separately. So with so much gen AI being built into products and used for time savings and efficiency, how might this change the competencies you look for when evaluating CSM talent? And anyone can jump in?

Melissa Allen: So, I think that's really interesting. I think it always kinda goes back to like my fear, right? Like of relying too much on it. I know a lot of our interview processes is, hey here here's, a use case with a customer, go build out a deck and I want you to present it to me and so to be able to still do that and make them feel like you're connected, make them feel like you care and that you even if the data was given to you still actually absorb that and, you have that relationship you're trying to build. I still think, that is so huge. So that's still going to be a necessity when in like the interview process to make sure even if they are given the data to make their lives more efficient by using something like AI or in an interview process just given it, can you still show that you've absorbed it that you've read it that you want that connection with that customer to help them be successful. And also that you can actually understand what's been given to you instead of just reading it off, right? So I think they're still going to be those tells they're still going to be that way to do that. And almost nowadays it's actually probably going to be nice that they can say even if the data is given to, you, show me that it still matters. You know what I mean?

Sonam Dabholkar: Yeah. I think I'll add onto that. I think so much of a CSM is a lot of like the soft skills around beyond just the business document. It's the relationship building it's you know, do you have active listening skills? Can you like understand the customers pain points? And then, you know, come up with an action plan that actually addresses those. And can you build like that trusted adviser relationship? And I don't think AI is really gonna replace that or should replace that. And I, so I think part of the interview process is gonna test for that and we'll account for cases where that's not in place. But I do think like AI is kind of the next wave of the future. And I think a big part of the CSM role is also just embracing change and innovation. And so those CSM that are really open to embracing AI and using it to be more resourceful and more effective in their jobs. I think that's also like a great skill to have and to look for in the interview process.

Matik MC: Great. Thank you both. The next question will similarly, will AI be used for candidate evaluation concerns or benefits?

Melissa Allen: I wonder if, I mean, if you end up using it for show me if there's red flags, if we've hired anybody with very similar answers with very similar resumes that have atritted, right? Or that have been put on AI think you can utilize it like that, but you're still gonna have to make those conversations you're still going to have that got you're. Still gonna go through that process, but I think utilizing it for that like show me past experience with people that have potentially similar things that have been put on a pip and all that. It will at least provide extra context that you could actually address in the interview as well. So, I think that could be a really good use of AI and that kind of interview process.

Calvin Multanen: Speaking from like Greenhouse perspective, I think obviously, we advocate kind of a structured approach to hiring. I think there is definitely a place for AI and stuff like app review and kind of helping, you know, recruiters and hiring managers like quickly find the right types of candidates and like help assist with that. I think it gets a little, the waters get a little crap when you're like reject or, you know, I don't wanna be in a world where AI is like choosing who to hire. I think it can help and like enhance, you know, kinda of the decision making process and should like, you know, to most point like kind of highlight those red flags or kind of make that more apparent. But I don't know, I don't think anybody wants to get rejected by a robot. But I also know from my experience, like if anybody was a Greenhouse customer, like if you build out a scorecard, right? Like I'd be really great if you could just like use AI to kind of like generate some of that for you and you can edit it, but I don't know if anybody from my product team listening, but, that's a personal pay point for me.

Matik MC: Thank you. All right. I believe this question is for you, Kevin, because you brought this up, you mentioned customer PII, what are your policies around that? And touching generative AI models?

Calvin Multanen: Yeah. So specifically our like we're pretty restrictive about PII, especially customers but also especially candidates like everything like our data warehouse is pretty much restricted. There's only a few things that will touch it and can actually access to it. Everything else is using like kind of anonymize identifiers we're very careful about how we handle that type of information, which systems go through it. And actually anytime we add a new system that uses PII, we have a whole like because I GDPR or something whole like process for like adding. So, I very, it's very lock down. So to that regard like AI tools, it would any I tool that we would use that would go through AI or use PII would go through the same kind of value in process. I think for stuff that we've already evaluated and gone through that whole like sub process through, you know, process and the kind of verification. I think I would hope that it's a little easier for us to get that like implemented and so we can start beta testing stuff. But at the same point, like the last thing I wanna do is, do you know, have something like that has a ton of like really sensitive information there? So I think we're okay to kinda hold off and like not necessarily jump the gun on, you know, putting all that into a tool, but I think there's still a lot of applications for stuff that isn't right? Like aggregated data, being able to like do trend analysis and like all sorts of stuff like that. Like I think you don't necessarily need that PII there. So I think there's ways around it.

Matik MC: Great. Thank you. I think this will probably be our last question for the session. But if you, there are any more questions, we can get to those after the session and get some answers for you. Which AI projects should a start up in high growth, focus on first ie, some options, churn risk, product usage, success, planning, et cetera.

Sonam Dabholkar: That's a hard question because those are all really good examples. I would say, I mean just given the macro economic climate that we're in, I think churn risk, will be the kind of highest value project to focus on because with churn oftentimes you need to get many quarters ahead of it. And so the sooner that you can work on something that helps you identify those indicators of churn and downgrade risk of your accounts, you can start working on those renewals quarters in advance and start to see the impact of that. So that one gets my vote.

Melissa Allen: Yeah, I would definitely second that, but also kind of going one step further, figuring out what potentially is causing the turn and then like literally water falling it to saying, okay, what do we need to do? Maybe we do some success plan analysis and see like which ones don't have success plans, which ones do? I mean, it could be a total waterfall. But yeah, I think churn is definitely going to be your top one because then it could lead to so many other things that then you can then do analysis on and then figure out where can I make the difference? Where can we make the change?

Bo Sun: I'll agree. Yeah. And I think that waterfall kind of analysis you're getting under the covers is what this explainable AI that I just provided the link to was so helpful for because sometimes you'll think that you're in a great relationship with that customer. But when you look under the covers, you're like, wow, that product utilization is terrible over here. That's not gonna go. So welcome renewal time. So, I definitely agree with that. I also just think from like a, after a hyper growth startup that the market opportunity there is pretty large if you could really kind of fix, the churn problem, large POC.

Calvin Multanen: Yeah. And I agree. Yeah, turn would definitely be my priority as well.

Matik MC: Great. Thank you all so much for joining. Thanks everyone for joining on the attendee side. We're at the end of our session. Like I said, any questions we did not get to, we'll get to those post summit and get those out to you along with the recording. All the sessions today will be recorded and we'll be available after the event. All attendees will be notified when this, the, that content is live through your e-mail and thank you again to our speakers. Our analysts. We really appreciate it and thank you everyone today for attending. Have a great day.

---

Click here to see all sessions & speakers, and navigate to recordings.