Putting Customer Context to Work with MCP

Awesome. Well, again, thank you everyone for joining. I think it's the last session of the entire day for some of you. I'm hoping to keep it lively and definitely interactive.

I hope that we have time to where I can actually walk you through some live skills I've created. So it goes from a little bit less slide where it's more like actual what it looks like and what you can do with it. Hopefully some good takeaways. So as you guys have all heard for the last two days, how we work, the speed at which we work, is changing drastically.

Sometimes that can feel very scary, nervous. But I think really the core piece to take away is that, what's driving all of this is the habits and the domain expertise that all the people in the room bring to that. That's what's really going to enhance the overall quality of how we use MCP or any AI tool overall. I keep on forgetting I have this down here.

AI is really powerful, but it does not have the wherewithal to understand how to use the data like the human in the process does. So again, it's siloed AI tools. We have generic outputs from individuals who are learning and uploading their skill sets. We have manual workflows that have different intervention points that are disjointed.

A lot of times it even creates more reactive CS motions than we really wanted to. Is anyone really still confused about MCP for all the things that we've heard about this week? Anyone? Everyone have a pretty good understanding?

Okay. Honest feedback there. I like to think about it as just MCP is APIs on steroids. It's a really clean view.

It's super easy for anyone to use. I used to think I'd hear API and I'd get really nervous because it's just something you can't, it's hard to see it or imagine it. I think the great thing about MCP is it really helps people have a better visualization of what they can do. But ultimately it's about understanding it helps drive an open standard for everyone.

It is live context. So we don't have to worry about the frequency of data flowing. We don't have to worry about like, "Well, I did it over here and now you've got to wait 24 hours for it to sink over here." Or it overwrites everything who's had that hyperloop happen in Gainsight, not me at all. It is secure and you can drive permissions on it.

So it gives everyone the competence that it's not going to get used in unintentional nefarious ways or people are going to bring in bad actors into the situation. And also it's a really great action layer. And I think that's the piece that every CSM is really focused on is how do I use it to take more actions or really just be a multiplier for them. I think that's the other big pieces.

How does it help them alleviate what can seem like a laundry list of things that they're good at doing, but we all know it takes so much time and we just don't want to do it overall. Why isn't this optional today? So one, I think the biggest piece is we all have heard the AI agents are everywhere and the reality is those aren't going away. Every business, every strategy for the most part is really focused about how do we take advantage of agents, whether we really know what that means internally, how do we agenitify it, and really then apply it to help the business be more efficient, but also help our individuals be more effective themselves and productive.

It also allows better data governance. It allows you guys to manage that. It allows everyone to have more clear paths on how to use context from different tools to help their jobs overall. And then it's also about really just creating that comfort that we are reducing the data silos with the different tools we have.

I would imagine every here can acknowledge like we have a lot of tools in the tech stack. It's painful for ops folks to put it all together. And for a while there, we would just buy stuff, deploy it, and then never think about how we connected it back to everything else. And so created proliferation of all these tools and everyone had their own little, I really like this one, I really like that one, and that game right there.

So now we get to remove all those data silos and really drive more efficiency in the long run and help our ops folks manage it all a lot better. Because that's also painful. And then just in general, we all know it's a huge industry convergence. OpenAI, Anthropic, Google, Gemini.

We are somewhat-- we have to move at the pace that these big dogs want to move. And that's really hard for everyone. I think Anthropic released another model today and has everyone scrambling on LinkedIn to try Opus 4.8. So they move really fast and now we all have to follow suit.

And I don't think anyone in the room has really ever experienced that, where you have to literally play the game or you get left in the dust. That's a new thing for us. I will say coincidentally, the tenure in the seat no longer matters as well. Everyone is on equal footing of understanding what AI is and how to use it.

So CCO, IC, they are equal on understanding how to drive AI effectively. And that's a really cool thing to think. You don't have to worry about what do I do in 20 years with my career. You can accelerate things really fast.

And that's encouraging for a lot of different things and different reasons overall. So I feel like at this point, you guys have seen this slide like 1,000 times already this week. But again, how does MCP work with Gainsight? So I did this deck before we announced all the cool stuff yesterday.

So now you can add on three more product sets to this, which is awesome. Very cool. But again, the idea is allowing you guys to tap into the tech stack of Gainsight, the product suite, and really be able to pull all those data points, the context together, to then really drive the necessary actions in the context that allows everyone to work more effectively, whether it be product management, support, leadership, PS services, FDE, in general, and keeps everybody on the same page. There's no more of this.

Or there shouldn't be much more of this. Well, I think it's this over here. And then you get to have that fun debate on who's right. And usually the person with the higher title wins that battle.

So again, this is what's really cool, is it keeps everyone on the same page, helps drive it more effectively, and then gives you guys actionable layers to move that data to other systems or just internal conversations. So the other piece to know is understanding what are some of the different use cases for maybe just direct CSM. And so again, a couple of them. Are any of these net new for anybody on what you can do with MCP?

OK, good. I'm not going to go into them then. But again, it's about really taking the understanding of what does a CSM do if we're talking about customer success and really unpacking that into whatever MCP tool you use. And so what do I mean by unpacking?

Well, the reality of building a skill-- we use Claude or whatever you use-- is you kind of have to think through the habit and the actual physical work you do to then drive the instructions. That's really the key part. So that's why it's really important when we say the humans in the loop, the context piece, is you have to really understand what they do on a day-to-day level. You have to have extreme, in my opinion, really good habits and really good domain expertise of the workflows overall to really help use AI as an accelerator.

Because again, if anyone's really played with it, you can spend a lot of time spinning up skills or other things that fall flat because you just don't know where the data is or how to craft it well. So again, it's really important to think about when you're doing this and developing it. You have to think like the person in the room that's going to do these things. So that's an important piece to take away.

[END PLAYBACK] So again, these are not anything radically new, but they do help speed up a lot of the internal administrative things that, if we don't do it right, there's gaps down the line. So account transitions, account briefs is a good one. If you have a huge portfolio of business and you can automate your account transitions and make sure they're really well done, that's a super helpful piece to give on to the next CSM or to help another CS leader who's coming over that book of business to get them spun up on everything. Eliminate the knowledge loss.

That's a key part of it all. I think the other one maybe I don't call out here is who loves the number of internal meetings they have on their calendar? Don't everyone jump out of your seat. But this is a great one.

I hate internal meetings like this. I'm going to put 15 minutes on your calendar to ask me a question you could have done on Slack, sent an email for, or just self-service through Timeline or whatever you like to use, right? It's super frustrating. And then more importantly, when the questions get asked, you're like, that was not a good question to ask me.

We have so many more things to tackle. So again, it helps speed up all of these things and really allows everyone to then focus on the things they really need to do. And sometimes people will say it allows you to be more strategic. But sometimes it's as simple as, I just don't have to do anything now because it's just being all done for you.

So again, these are some additional use cases we have on our website. And so one, if you're thinking about, where do I start? How do I do this? These are great things to look at, learn, take in, see how you guys do it in your own businesses, and then apply it.

The one thing I'll tell you is this stuff is not just simply a plug and play process, right? You can't take a skill created by Gainsight and just necessarily drop it in your LM and expect it to work magically. You have to intervene. You have to make the filter changes.

You have to help it set up to understand how you guys do things in your business. It's just a framework to help get you started. So I do want to call that out before you guys start playing with it, and it doesn't work. Elliott, you lied to us overall.

So again, some other ones that you can go on. I think at the end of the day, this is a really important slide for me is, again, these are key use cases you can talk about. But really, there's a new standard that gets to be set forth. So raise your hand if you're a CS leader.

Raise your hand if you guys have a hard time getting your CSMs to do it the way you want them to do it. Raise your hand if they do it, but it's wildly different. So it makes it almost impossible to consume the information appropriately. OK?

How hard is it to tell the CSM if they don't do it, they'll be in trouble? Or maybe you do. Do they still do it the way you want it? OK.

How many times you have to nudge them to do the thing you want them to do? So here's the real key part here. We don't now get to drop the bar to where people want the bar to be. You get to elevate the bar and keep it there.

Standards and expectations now. At the end of the day, I've been at gainsite for eight years. Clearly didn't have AI and MCP for eight years. If I chose not to do something, that's just called lazy, and I then had to go do it another time that maybe it was in the middle of night.

But I chose to do that. But ultimately, it's an excuse. And I know it's hard, but that's the reality of the situation. So now we get to elevate the standards and the practices, and you guys have an easier time enforcing that while the data accuracy becomes tenfold better for you guys.

Number of times you guys have heard, I want all my emails auto-logged somewhere. Salesforce, activity history. How many times did anyone actually ever use that information for any intelligent action before MCP? You can raise your hands.

If you don't, I know you're lying. So just an FYI. But now you can. Staircase is a good example of that, but all the context that's in that now becomes way more applicable and easier to surface for insights.

These are really important things to think about. Like, why were we asking for that? But now you get to really think about the why and the how and go from there. The other ones, it's also about now you know they don't have to worry about the time it takes to do something.

You've given them the format. All they have to do is say, press go, and be specific on what they're looking at to do it. But now you guys get to have these things documented really cleanly the way leadership needs it to be done so it helps you guys drive your internal meetings, whether that be a risk meeting, a forecasting meeting, right? But more importantly, it's now an expectation that it can be done the right way in a very efficient way too.

The other one is you get to allow the team to actually also experiment. They get to go out and do more things. They don't have to just live in what you give them. But it is important that we do try to have a line of centralizing the skill development because it can get really wildly out of hand.

At GainSight, we don't do that yet. We need to. But you can have four different people with four different risk skills. And they're all good.

It's just, again, we're back to the emotion of-- I think about it just like another Google Sheet. It's just being expanded and done and copied and duplicated. I think early on, some of the things I've seen as I've talked with my clients is there's definitely the notion of who creates the prompts. And I think it's good to think about who do you guys consider either the CSM with the most subject matter expertise or the one who does it a very specific way that's really well done and asking them to go through it, but then getting collective buy-in from other CSMs.

Because when more people are able to add to it, there's easier buy-in with that process. But at the end of the day, it's up to the leaders to say, hey, that's really good. We're going to use that one. We don't need everyone to create 10 more of these.

We're going to use the one, and we're going to drive it really well. We're going to operationalize it, practice the habit, get going. The other one is poor data quality. I mean, don't get me wrong.

The outage quality and quality out still applies to AI. But the reality is now you get a sharper image of where you need to enhance it. And it's not so much of a vague black hole of who gets to go tackle that big problem. So that's another really key thing, is don't let that necessarily stand in your way.

Use AI MCP to help give you guys a clearer picture of where you need to go solve the data problem, and then go work with those teams to go do it. Whether it's truly a CSM data quality, I mean, who reads their CSMs timeline notes? Anybody? From a leadership perspective?

Would you say they're excellent notes or just emails on a timeline? So again, you get to really enhance this, drive it more effectively. And then again, I think if you're looking about how do we get started going forward, it's naturally, I would say, start on your core leading activities. Focus on the ones that you know take the most time, they're painful.

Could be that QBR deck. Could be account transitions. Could be the risk management process, a forecasting process. So whatever those are that you know actually has the biggest time, consumption for the individuals, start there.

Think through it. But also, don't wait on it. The fun thing about this is you can be practicing and developing these on the side when you're doing other things. So that's another key thing is it doesn't have to be, I've got to go to my ops person.

You can just play with it, practice it, enhance it yourself, learn more about it. The other part is, again, this really just helps align individuals to leaders. Makes life a lot easier. It helps everyone just be, again, not to say it over again, but more effective.

And then also, if you're going to do it, measure it. That's the other thing is if you're just doing it in AI, and this is kind of why it's important to have your tools like Salesforce, GainSight, whatever that is, still is a repository because you need to measure the impact this has. We still don't get away from, we're doing it over here. We've helped them eliminate stuff, but we're not going to measure how it impacts gross retention or expansion.

These things should be accelerators to get to those outcomes. And I think it's important that you kind of think back to that because at the end of the day, if you lose track of that, then we're not really doing the right thing by the business in the long run. So the other kind of key thing right now is, we talked a little bit about CLI yesterday. We all have a good idea about MCP.

But CLI is kind of the next wave. Who's in ops, Brian? I know you're in ops, Pele, ops. CLI is going to be the fun one because this allows you guys to then actually take more action, right?

This allows you guys to build from the outside in, allows you guys to do all the things like, gotta go build that J.O. program, think through all those fun things. So it gives you the ability to really accelerate that without the frustration. And also gives you, I think this is gonna be a really cool one when you can actually take, like whiteboard your designs.

Like if you're on that offsite, take a photo of it, drop it in there and see if it can replicate what you've designed. Like who's actually played with designing something with like Claude or anything? It's awesome. You can take a photo, like, okay, I want this to be green.

It'll do it for you. So those are kind of the cool things that I'm hoping we always get to. But again, just to help give everyone some context on what's different here is that the CLI allows you guys to actually take more action within the tools to help you guys make the lift a lot easier overall. So again, I'm kind of keeping this fast because I know it's the last session.

I do want to get to Q&A and hopefully show you guys some of the things I've designed, just some takeaways. But one, I think the big piece here is think about what's gonna be a real differentiator when you guys go drive this internally. Let's see, MC is the connection tissue. It's not the product.

So again, think about it as what do we need in different systems that we can accelerate together through the MC pre-process, whatever LME you're using. Like I said, start with two to three use cases and then start to roll those out. Don't over inundate the team with a thousand of them because like enablement, if it's information overload, they'll kind of go reactive on it. And then they'll find all the reasons why they don't want to use it.

So start small, roll it out, get input, feedback, iterate, move on to the next one. And make sure you're using it like in the process, like that's an important thing. And then lastly, the destination is the predictive CS, not your practice. So that's the other thing is we want to make sure that everyone knows where to look for this stuff and how does it help drive better insights and overall business outcomes for you guys in the long run.

That's it, 20 minutes, keeping it fast and quick. We're gonna open up for questions. All right, here we go. Are there any data object restrictions with Gainsight MCP?

Is it able to access objects like surveys, timeline and even custom objects? I'm not sure about the custom objects yet or the survey objects, but success plans, CTAs, timeline, yes. Obviously if you plug in staircase, you get all the inbound emails that come through staircase. I'm sure over time, like a lot of things will roll out more access to those, but I think like a lot of things, we wanna see what people really need and if there's any kind of security concerns about that obviously.

So good question. What's the most impactful use case you've seen for CSMs once they connect Gainsight MCP? So internally, I think each Gainsight CSM will have a different perspective on which skill they find the most valuable. I'll show you the first one I built and why I built it.

I moved over into product management so I don't need it as much, but it's a really cool skill and I think for me, if I was still using it, it's very helpful overall. But I'll show you the one I've also created for product management. How are you managing the most, oh sorry, where'd that one go? Oh, okay.

Do you need staircase to use MCP? No, that's a great thing. You don't need staircase, MCP is through the LMs so if whatever tool you're thinking of you have internally has an access point, a server, you can use that and drive it, so that's great. How would, okay.

How would CSOps measure the usage of MCP skills in general? So and against retention impact? Well I think against retention impact is lining up those leading activities that you've designed for and measuring how they're driving renewals, forecasting and expansion plays. Are we saving time for the team?

Are they able to take on more accounts if that's part of the initiative? Or more importantly, how are they improving the experience for the customer? Those are some other ones to look for. But again, we wanna drive the leading activities back to the outcomes and so those are probably the key three ones, renewals, retention and just overall experience.

Can Gain Site MCP surface limits and costs? We often run against usage limits when exploring ways to maximize. So that one is definitely, that would be different for every client on how much you're willing to grant your individuals in terms of usage. Some clients, they don't wanna do that because they know it's gonna explode the cost.

So again, that's kind of an internal one. But you should be able to track your usage through whatever LM you use. So I can see my token usage, how that drives. So that would be something to talk about with your IT, whoever owns your LM number role.

How are you managing and sharing prompts cross-functionally without feeling like you're maintaining a manual library repository? That's a great question. I don't think really anyone has quite figured out that governance model yet. And I would say you don't wanna stifle the team and not let them test things out and learn and play with it.

But I think at some point it is good to have someone, well again, whether they're part of CS operations or if you guys do have a separate like AI GTM team, GTM team, to ask them to be the ones to house and track this and then also be the developers of it. I think once you're able to do that, you'll see it's not as hard to manage it and you guys feel like you're actually producing more valuable AI tool or MCP skills or Gemini gems that you guys can use. So I think that's a big one. Can GainSight MCP surface overall business insights or is it more focused on accountable insights?

Definitely can focus, it can highlight business insights. Again, a lot of that depends on the data you're hooked up to. So if you're connecting Snowflake, whatever data house, data warehouse, Salesforce, you can definitely pull all that information together and provide really in-depth business insights as well as account level insights. Do you have any tips for convincing internal security and IT teams to adopt the GainSight MCP?

I think it's a two-way conversation, right? We have to understand where they're coming from from the security concerns, but we also have to help them understand what the team has to do in order to do their job today and how this is gonna really help alleviate a lot of that. And so I think if you can take it from both sides, the process is the process, but I think it'll help everyone understand the why you're asking for it and how impactful it'll be for the business in the long run. Any recommendations for resources to learn?

See more examples of how other CS folks are using MCP. So again, we have the website. If you haven't seen it, we'll plug it here in a little bit. They have, we have some generic use cases on there.

To me, this is kind of comes down to a simple, just practice. Like the more you practice, you'll be able to share those resources, you'll learn more, you'll gain more. And like anything, it kind of becomes habit, right? You get really good at doing these things and you'll be able to think of things that maybe you didn't know you could do and now you can do it.

I will show an example of that here in a second. Yeah, can you demo some of the ways you use? I'm gonna get to that here in a second overall. How can Ops use MCP well if their teams, CSM's managers don't have Claude?

That one's a good one. I might need to come back to that one a little bit. Okay. Demo, yes.

The company I work for is in a highly regulated industry and only Gemini is allowed, sorry. (laughs) Let's see, so where'd it go? Hold on. Did that one get punted?

Oh, there it is. The company I work for is in a highly regulated industry and only Gemini is allowed, which requires a CLI to use an MCP? As in the gainside admin, how would I even get started using CLI? Good question.

It is, the CLI is a little bit limited to an admin-based role. If, and we do have instructions on how to hook that up so we can share that with you. I have not, I don't have many clients using Gemini so I don't have a good answer for would the CSMs also need the CLI component to it? So I can't answer that one directly, I apologize.

IT team said the third-party MCP access to our co-pilot studio may not be possible to set up and if it's there, if there's a large financial impact of the consumables associated with co-pilot studio agents that call the MCP servers, MCP is accessible in Microsoft co-pilot studio, I do not know the cost on that. I apologize. It's a great question but unfortunately I do not have enough insights on the Microsoft co-pilot to offer any credible answer to that. Gainside scorecards, do you see any potential use cases using MCP to create update scorecards?

I mean we're 100%. I've been doing Gainside for eight years. Do you know how many number of times I see clients just let that scorecard just die? Iterate every six months, use it to give you guys better insights on what the actual right measures are.

Use it to help you guys navigate the waiting thresholds. Use it to pressure test what you have built. That's the great thing, you get to fail so much faster now without the concern of the impact it has on the team, the confusion scorecards create, it'll give you guys really powerful ways to ensure that you're on the right track with what is truly deemed healthy. Has anyone ever just wanted to ask the business, do you guys even believe this is what health looks like?

Right? That's the number of times people are like, we built the scorecard in CS and I'm like, then you try to go talk to product about it and they said that's not correct. This allows you guys to bring all those together, have that conversation in private and then go bring this back to the business on why you decided to do this. So I think that's a really powerful thing to think through.

And also, yeah, you can just design the scorecards, like I'm all about using Gainsight, but we don't wanna push you places where you don't want to go. And so if creating the scorecard in the MCP gives you guys the best visibility into it, but then cross-functionally the best visibility, do it. That's fine. That's all we wanna hear is that everyone's using it to be more intelligent about what they're gonna do.

One of my concerns about rolling out the MCP is that the Gainsight's documentation says that Claude can write to Gainsight. Can you offer any guidance on controlling that? Queering, why would you wanna lock that down? That's the magic.

Do you want that living just in MCP? Like that's, to me, I mean, I'm not saying, I understand. (audience member speaking off mic) Well, this is part of the enabling piece, right? Like centralizing the development of it, the why we're doing it, how to ensure that it doesn't do mass bulk updates.

Like these are the tangible pieces that everyone has to learn and understand. I mean, I get where you're coming from. I just, I would say, I get nervous when people are saying, I wanna lock that and prevent that, because that's what helps the individuals be more efficient. It helps everyone, like everything be cleaner in the long run, but there is definitely, you wanna think through the list of like, what do we really want them to do?

That's important to track like that, so. But yeah, it's a good question. I mean, again, totally understand, but just, I would say, don't be rigid on it. Like thinking from both sides to a degree.

What's the difference between linking MCP to co-pilot and linking Gainsight API to co-pilot? I think the API is just more labor intensive for the admins. That takes more work on their side. MCP, like anyone can do that.

I think that's what's fun is like, you can have the super forward thinkers and they can get going on it. You can give people the space to do it, or you can just want it, you can do it through APIs. Do it through APIs. How can ops use, oh, I read that one.

MCP and its action layer. Is it possible to revert MCP initiated actions of mistakes? Yes, I do it all the time. I have Gainsight up, I have my MCP up.

I'm testing what gets flagged back and forth. And I'm making the changes in my skill instruction wise as I see it go in. So yes, definitely that's a good process. And it'll help you guys understand like if your skills are doing the right things you want it to do.

And if it shows up wrong, like one of the, I'll walk through this with you guys, but I quickly learned if you don't hard code, keep HTML formatting in your skill, you'll get a huge block of text in your timeline notes. And so you can update the skill and then ask it to revert the timeline entry and we'll do that for you. It's pretty slick. How do you ensure a sensitive data remains secure?

That's a good question. I think one, hopefully the individual is using it, understand your policies. They have good intentions, or not bad intentions, but they have a good feeling of how to avoid pulling in, potentially sensitive data. And hopefully also like through the work of collaboration with your security team or IT, they're able to put some of those permissions in there to help ensure that nothing gets leaked like that.

We already answered the tracking usage. How do you track efficacy of the things you build? Example, handoff summary. That's where the human comes in.

That's the context piece, right? If it, like you need to be there to help make sure it's the right information, right? That's important. So I think that's the human part that we wanna make sure is in that loop.

And that's what's helping everything become better is we're using the people to ensure that the data's right and it looks good if it doesn't revert, update, et cetera. Have you used the Gainsight PXMCP server for product? I tried yesterday morning. I ran into a blocker because I was hooking up the wrong subscription prod, but I was trying to get right after it.

I was trying to test out a rolling average of timeline utilization for just my book of business. Yeah, so I'll come back to that one after the conference. When should you use MCP versus action and Gainsight? I would say, again, this is like a little bit of my own take, but like every administrative thing I do, is a skill.

That's how I drive it today. And I think the other piece here, and there's certainly things that you want to use Gainsight to trigger off the data coming in from MCP, right? So, and some other pieces to that I'll show here as well, but that's how I just think about it early on. I think as we get more robust, as the CLI stuff comes into more play, it'll allow the ops seem to be way more curious and sophisticated about how they leverage the action from MCP and or Gainsight is just an example.

What's the use case you tell every CS team to implement first with MCP? What does it take to get started? Risk management. CSM's number job is to manage risk as it relates to the renewal coming up.

So it'll help you guys enhance a really critical piece like that. And so I would say start there. If you're looking for a starting point, that would be my hot take on it. Any other questions before I go into demo land?

Yeah. Yeah. Yeah. (audience member speaking off microphone) Yeah.

It's a good question. I think probably there, I don't know if I have like necessarily the immediate answer to that, but I think initially to your point, like as the handoff, like is the quality experience of implementation or transitions going better, right? Are your clients bringing up painful transitions or better yet, are you guys using MCP to evaluate the conversations your CSMs are having to see if they're using the information? How many times have you guys, like how do you define, how do you determine if a CSM is strategic?

How many times have you guys asked that question internally? No hands? I hear it all the time. How do I get my CSMs to be more strategic?

You have to listen to calls. You have to listen to what they're saying. You have to read their notes, right? Now you can elevate all those insights in a heck of a lot faster time to see if they're using things that you've invested for them, right?

That's frustrating as a leader. You invest in stuff, they don't use it, and then we're wondering how do we get them to be better? So those are some things to think about. It's just like, how do you like go full circle with it?

But I think to like the broader point, measuring time has always been a little bit as squishy because it's like you either need them to timestamp it so you can track it, but anecdotally, I can tell you like doing this presentation, the deck or a QBR deck, I can do a video recording and just drop the transcript into my cloud and have the presentation. So I can take what used to be a couple hours or broken down into multiple hour blocks in a week. I have it done in 30 seconds, right? Like those are some real impactful things and makes my life a heck of a lot easier.

So again, there's no like, I don't have a solid answer for you today, but I think just thinking through that is important as you guys go. You guys ready for some demos? Cool, let's get to it. Come on, there we go.

So this was my very first skill I built. So we opened it up at RKO in Austin and I got back on a Friday and I immediately started playing with it. So my initial effort here, and this took me six hours of playing with. So I wanna give everyone like a time, like time expectation when you do things, right?

And this, again, I was just doing this while I was doing other work in between calls, things of that nature. But what I quickly learned is I was trying to initially do this in all one compact process, like writing an instruction. I think like at the time it blocked me instead I had 20,000 characters in my instruction, right? Like that will happen, it will block you.

So it's about condensing it. What I realized is back to the human in the fold, I started unpacking physically how I do each thing and then writing a skill for that. So this is one massive skill with six different sub skills in it. So what this does is I can take, if you were to do this on a Friday, I can point the skill at any week in my year, have it read my calendar in Gmail, go into Gain Site timeline, Gain Site MCP, go into staircase, then craft me a bolted summary on what I should be talking about with the client based on my account plans, the key talking points of that.

If it's coming up for overdue success plans or I forgot to update my success plans, who loves getting their team to do account plans? Raise your hand please. It's frustrating, raise your hand. I can have my skill review my call transcript from the last call and automatically go and update my account plan using success plans in Gain Site.

And pull out a call transcript and plug it in the free form text on my account plan. If the account's at risk and I've missed my weekly update or I know I have to get it in, I can hit generate my risk update and it'll apply it to the timeline right here. Down here, agendas, a lot of CSMs like to send out an agenda. It'll craft me an agenda sent out in advance for the upcoming call.

And I can send this to my drafts in my inbox. So it takes care of that. Ghost prep. So at Gain Site we do a lot of writing on behalf of our executives.

And so we've all talked about using chat GBT to set the tone and the context. So I created a skill that actually pulls that already, or pulls that in and then already develops the email. So if I know I have to get an email out right away, I can do that pretty quickly. So that's one.

And this is what I was, this is, if I was still in the CSM role, I'll be doing this every Friday for the upcoming week. So again, that's one powerful use case bundled up. Any questions on this one? Yeah.

Co-work. Co-work is a little bit more expensive on Claude. So that might not be the thing that the business wants to drive into, but you can create a task that can run this every day for you if you want. (audience member speaking off mic) Agent Studio.

You could, yeah, yeah, you could. Oh yeah. (audience member speaking off mic) So this is a good example. Any other questions on something like, just like this?

So next one, who here has a digital scaled initiative to drive? Everyone's hand should go up. Everyone's hand should go up. If your CSMs have 100 accounts in their book, you have a digital scale initiative right there.

What's one of the hardest things about doing anything at scale? Content. You have to have content to drive action when you're doing one to many. So this is using Gainsite CS.

I got access to the community, MCP yesterday. I did this yesterday morning in 20 minutes. And I use Staircase. And I pull in all key themes that my book of business or our customer install base is talking about.

And I'm having it elevated here, drafting an abstract for it, coming down here, giving you an idea of who to position the content for, and then giving you guys the option of where to deploy that in the different channels of delivery. You've thrown in CLI, now you can actually just create the program and drop that all in the sequencing of a program just to put two and two together. What I like about the community one is it allows you to go straight out to the posts on community. So if you're trying to figure out how to get CSMs to look at the community or pull in interactions of their book of business, this is a really good way to do it.

They can quickly get to it, understand what the client was talking about. They can zone in on just their book of business. So there's a lot of really good ways to get more intelligence and understanding what the client's trying to accomplish overall. The next one.

So when I moved into product management, has anyone ever worked in product management, works very closely with their product management team? Like what's one thing PM has to do before they approve any builds on anything? They have to understand the impact of the request, how that's gonna scale. So I created a skill with Staircase and MCP.

So in my new rule, I kind of helped manage the product request process or risk process, excuse me, that will, I can just drop in the terminology I'm looking for, it'll use Staircase and Timeline, do an aggregate of all the data, pull in my top, come here, view. It'll pull all this in, PDF if I want. It'll give me the top five accounts by ARR and the confidence level behind the association of what they want and what the request is. So it gives me a really good likelihood of how much GR impact this has.

If I come down here, it'll also give me the GR impact at different segments and then the observable impact. So it gives me a much better insight into, if we build it, how does this help the rest of the business? And it'll help take a lot of time off the PMs because they don't have to go as deep and as wide or have 10 or 15 client calls to validate. It's a good way to help give them a starting point and then how to go run with it if they want.

And more importantly, down here, it tells them which list of customers they should go just immediately go look out for and go just talk to them. So with this one, I generate here, I have this post to a timeline to a CTA directly from here. So that's the other part is now you get to have really good operational rigor and you get to store all the information in a really effective way to make sure when we come back and review it, it's easy to review, easy to consume, easy to say yes, easy to say no. Those are the important things that we wanna get to.

Any questions about this one? And the great part is it looks just like this in timeline. So it doesn't get blocky, it doesn't get really messed up, who's experience where your email templates go wonky in programs. So you can keep the formatting and that's what's great is leaders can define what it should look like, what good looks like, and then all you have to do is just run it.

That's what's amazing about it. So this is a fun one. So before MCP, this was me playing with Cloud Code and I've always felt, it's really weird, hold on. Let's zoom out.

I was with all the AI follow up, I was like man, we have this great conversation layer that will really help, oh shoot, here we go. I was like it'd be really fun if we created call analyzer, right? We've all used Gong, we think that's pretty cool. Coaching, how do you develop CSMs are more strategic, we have to coach them, we have to enhance them, we have to do call reviews.

So I was like I'll just play around to see if it can be done. So what this is is it just basically stacks the CSM team, it's a prototype. You can put in your framework, like the different discoveries, you can parse it out. If you guys have a value realization framework, you can upload it so you can intelligently tell the AI what to look for in the calls and then define how good that is.

You can help define hiring profiles for different segmentations, right? How many times have we heard like our CSMs are tactical, they're not strategic, and it always seems to fall in the CSMs, have we ever evaluated if our clients are just more tactical in nature, and don't let the strategic conversation happen on the call? Wouldn't be able to, really cool if you could pair up, like, let's see if I get this real quickly. If you can pair up the distribution of their health in which accounts are more tactical versus strategic.

Like these are just fun things that float in my head through eight years of doing this job, because they're common questions I get asked, they're things I hear, just general curiosity. So this was me just fooling around with Cloud Code, and then MCP came in the picture, I was just able to build it as a skill for it, I wanna win. So on this one I just asked it to evaluate me, I'm moving out of roll, we gotta find a backfill for myself. And so I just was messing around with it, yeah, you can create a 13 page readout of comparing all your, I'm not saying do that, this was something I wanted to drive, but if you're curious you can do it.

And so it'll give you a full readout, look at all the call transcripts, help you evaluate what they're good at, what they're not good at, compare it all down, and again, it'll give you a score on what is good and bad. It's wrong, it's gonna only put me in an 82. But no, I think this is one of the cool things is, whatever you've been thinking about, can we do it, should we do it? Just go try it and see if it sticks.

That's the fun part about all this, is you get to bring more play into the workforce and actually have it be benefit for everyone if it sticks. Any questions on the demos, anything? Anybody wanna come back to anything? We're a little bit over time?

No more questions? Yeah. (audience member speaking off microphone) Yeah, on the website the skills they publish are my skills I've created. So they're there, they're there to help.

And yeah, absolutely. You wanna walk through something with me or think it out? But that's the great part and the takeaway here is, humans still need to be in the fold because it's their expertise that are gonna make all this stuff way better. And so that's the key part is, that's why it's important to bring, not just to think about we don't need the CSM or the CS Leader to be involved in this.

And I'm just gonna do it off the site because you really need what their experience is to help create the best piece and what good looks like ultimately. So yeah, absolutely happy to share. Go from there. Any other questions?

Awesome, well thank you guys for your time. Hopefully it was beneficial. Hopefully the demos were, you know, insightful. And safe travels.

Thank you guys.