Driving Meaningful AI Adoption: Turning Experiments into Real Impact in Post-Sales

Good morning, everyone. Hope everyone as well enjoyed some cocktails, maybe some Mexican food like I did last night. We're excited to be here. Kalpano just did a great job of teeing up the session.

The TLDR version is that we're going to cut through the fluff and really get into how we get AI to a place where it's driving business impact, value for the business, ideally, most importantly, value for customers. I'm very excited to be joined by three great leaders who in the trenches solving this equation right now and are ready to be very candid, very transparent, and share some important experience. But before we get started, we'll do a quick round of intros and maybe panelists, you can share just the information about your companies, because I think the actual nature of each of our companies is important as a key set of variables that go into our strategies around AI and how we adopt, how we translate that AI into business value, the nature of our products, the natures of our pricing model, our customer verticals, and so on. So maybe we can go around the panel for some intros and as a nice breaker, share what your dream personal use case is for AI.

Before we get into the business use cases, let's think a bit more personal dreams in this crazy world of AI. Brad, do you want to go first? Yeah, happy to. Good morning and great to kind of see everyone here today.

Brad Casemore, Chief Customer and Growth Officer at PartSource. PartSource is a B2B marketplace, as well as an enterprise software solution. And our mission is focused on keeping healthcare clinical assets available and online to be able to provide patient care. So really important mission that we have as an organization.

My dream AI use case is I love the concept of the daily brief. I don't know if anybody's explored the daily brief, helping you prepare for your day, plan for your day, but to me, there's one flaw in it. The daily brief only focuses on really your professional life and what you need to do during the work day. I'd love to have a daily brief that also expands my personal life.

I travel a lot for work. I love to get up and run in the morning. And I also love to try local coffee shops. So if my daily brief could suggest where I should go run, as well as telling me some local coffee shops, like that would be my ideal, is having kind of that daily brief that crosses the professional and personal life.

Awesome, and give us the 20 seconds on what PartSource does in the business model. Yeah, so our business model is a couple things. So from the B2B marketplace, we're essentially in N of one, our online marketplace is a place that where customers can go to buy parts that they need to be able to repair medical equipment. And so medical equipment, think of X-rays, CTs, pets, machines like that.

And those require both regular routine maintenance and scheduled maintenance, as well as they also have things that go wrong, right? And they require corrective maintenance as a part of it as well. And so our model is extensive in terms of not just the marketplace, but we also provide kind of the service contracts and things like that to help people manage their service contracts in one single place. And then our enterprise technology stack also we leverage AI and we use telemetry data off of the devices to actually be able to predict when machine and equipments will go down.

So you can therefore take it offline, repair it before it go, that ends up breaking and impacting patient care. Awesome, Christine. Hi, is everybody awake after that enthused? I hope so.

So Christine Strom, VP of Securonics and I manage customer success services and operations. We are a cybersecurity company. So we have SIM, UEBA and TIP products. So that puts us squarely in the face of everybody who's got something SaaS, are you protecting and securing your product?

So from that perspective, we wanna make sure that we are doing a great job at having our customers adopt our product. So from a personal use case, I am a big ultra runner. And what I would love to see is the ability for AI to take all of those wonderful health case studies that have been done that don't necessarily have sufficient quantity of women and expanded extrapolated to women so that we would know the health impact of women's running ultra distance, whatever that is, I would love to see AI be able to deliver more outcomes for women in that technology. Very cool, Shani.

Hi, I'm Shani Brandmiller, VPCS ops at Silverfort. For those of you who are not familiar with Silverfort, we're a cybersecurity company, focusing on the identity security, on-prem cloud, now also AI agent security. And I lead the operation and the AI strategy for the entire CCO organization. My personal use case would be, let's see if it's impossible.

Having AI kind of dealing with the arguments between, all the arguments between my seven year old twins, which happens approximately every 20 seconds. I can really imagine, you know, an automated mediator that can step in, stay calm, no emotions, no losing controls. Basically everything that I am not at 6 p.m. So that would be my dream use case.

I love that, I love that. My use case, closely associated, I would love AI to be able to feed the language to me on demand that I can use with my kids to get them to listen to me. First time, first time of asking, what is that perfect language? If AI can solve that, that will make me very happy.

Okay, so let's get into it. When we talk about and think about driving adoption of anything transformative, important, looking to drive business impact in an organization, clarity, quality, consistency of mandate is a very important baseline from the top down to have. When it comes to the mandates in your organizations from your, say CEO, maybe it's not CEO specifically, but what does that mandate around AI look like today? Christine, maybe we can start with you.

Sure, I think it's not dissimilar from the sentiments that have been said at the conference largely. So using AI on the product, using AI in every customer success tool that we have on offer, as well as making sure that we have any AI augmented tools, and not just using the tools, but also having the demonstrable metrics that show they have a positive impact. And to the degree that our board is asking, they are actually tracking on a regular basis, making sure that we have those metrics and that we have traction on them to make sure that it's not just AI purchase and then let it slide. They're really interested in making sure that we're using it and it has a positive impact.

And Brad, I know that you award winning team this morning on the work done on AI, so there's clearly a mandate, but what does that look like in parcels? Yeah, so our directive maybe is a little different. Ours starts with systems thinking, right? And so in order to have AI and have your AI experiments be effective, you have to take a systems thinking approach.

So do you understand the data signals? Do you have kind of evidence plays or motions, or do you know how to solve those things? And can you measure those, right? You have to be able to have measurable impact.

And so when you have those three things in terms of understanding the signals, right? The plays that are gonna drive the outcomes that you are looking for. And then you have this closed loop system. Once you have that closed loop system, now it's a lot easier to be able to put agentic capabilities on top of that.

And so that's the path we've been on really for the last 18 months. And now we're starting to see the results of that because we're able to apply agentic capabilities to that, and which gives us a higher degree of confidence actually in both the agent's ability to operate and to do so successfully as a part of it. So that's really the journey that we've been on, and we've been able to deploy agentic capabilities across our parts of supply chain, as well as many other areas of our business. That's great.

Shani. Yeah. For us, our CEO has been very clear. AI adoption is not optional.

It's like a company-wide strategy, and all the departments should get on board. Usually it will be the ops people who will lead it and push it forward, but the expectation is that everyone will find a way to use AI somehow in their day-to-day. Specifically for CSM, I can tell you that it means consistency, faster responses, being able to identify risks before they become an issue. And we need to remember that at the end of the day, we want to free up our people as much as possible so they can focus on the human parts that are the most important to gain relationship with the customer.

So in this panel, we're gonna deep dive into one specific use case per company that our leaders here are driving. But before we get into those deep dives, maybe give a summary of the kind of spectrum of use cases that you're exploring when it comes to AI, whether that be in the generative sense, the intelligence sense, the agentic sense, with a focus around post-sale that you're looking at in each of your companies. Maybe Brad can kick us off. Yeah, happy to.

So last year we made some significant investments really in our enterprise business and making sure that we started with this concept of saying, how do we just centralize all of our customer intelligence? And how do we really have one centralized customer intelligence hub? And that's actually what we did with GainSight, right? So we brought in all of the kind of user behavior, being an online marketplace, there's transactional data, they're being able to understand kind of feature adoption, and then being able to understand all the operational analytics as well.

Like, is the supply chain delivering on time? Are there disruptions and things like that as well? Once you have that centralized kind of customer intelligence, now we've been able to use staircase AI and various things like that to really understand how our customers are performing as well as the customer journey. And we've integrated those things into our workflows as well as a part of it.

So meaning now when there are things, when we get kind of certain risk detection, those things are automated in terms of notifications into Teams channels with the right cross-functional groups to be able to act on those really immediately. So now we have a much more proactive system, we aren't waiting for things to come up, we're be able to identify like where we need to sometimes do service recovery, right, as a part of it, especially when you have that supply chain capability where supply chains aren't always going to be perfect, they never are, we all get late shipments sometimes, but still you can acknowledge, right, the human kind of on the other side of that, right, and the disruption that that could potentially cause. We've also been huge in terms of how do we turn all this information into not just have the data, but how do you turn it into insights and then action. And so we actually use Matic in terms of we built out our plays and so having really these prescriptive DAX and making sure that it meets kind of all of our brand promises and things like that.

You know, for example, our EBR DAX, they would take sometimes three plus days to put together. Now the team has a 95% draft and the click of a button. And so it took that from three plus days down to less than a minute, they have a 95% kind of first pass of their deck by using Matic for it. And then the last thing we actually just recently went live with the Atlas renewal agent in our long tail, which has been a lot of fun to partner with Gaineside on and bringing that to life.

Very cool. Shani. Yeah, we have a few use cases that we're currently running. All of them should address like specific pain point that we discovered together with the CSMs.

The first one is meeting preparation. I think that was kind of our obvious starting point because we saw how much time CSMs were spending before every meeting. So we created kind of an automated AI agent that scans the CSM calendar in the beginning of the week and basically send them a brief summary before every meeting. Another great use case is white space and adoption, specifically for expansion opportunities, where we use AI kind to identify like common patterns between different customers.

And then we can suggest the CSM some potential apps opportunities for their customers. And the most recent one, and I think the biggest project that we've just accomplished is the sales to CS and over, which basically we aim to eliminate and provide some visibility between, to actually what happened during the sales cycle. So the CSM can step in into the kickoff meeting with more confidence and the customer doesn't feel like they need to repeat themselves and answer the same questions again and again. Great, that's exciting.

Christine. Awesome. So to add to the use cases here, we are obviously looking at common solutions, which is amazing. The other aspects that we were primarily looking at were support case offset, and how do we use the technology that we have internally to capture that knowledge base and make sure that we are using that, reproducing that for our customers and for ourselves.

So that first time to resolution then becomes less time to second and third and so on. The next part that we are looking at is looking at the digital TAM motion, looking at Journey Orchestrator, looking at AI to make sure that as we are taking care of our long tail, how do we put that in place for all of our customers so that they have this common fundamental experience we can build on top of. And then finally, the other part that I think is really huge for us is taking care of the overall knowledge base that we have as a company. And there are two parts.

One is that customer perspective that is crucial and important and in the gain side products, which we're all more familiar with, but that has to pair with a product and support case information, which is a different lens on the world. And so having this combined is a really powerful picture. And what I love about all these use cases and why we very deliberately titled this panel post sales, not customer success, is that this ability to think holistically across the customer lifecycle in the post sales context, but still holistically thinking services, support, TAM, customer success, account management and so on. I think it's so critical to really achieve that true adoption level of these AI capabilities versus exacerbate any silos that they're I say may exist in some companies here in the room.

We would love to drill into the hell of every one of those use cases, but alas, we do not have time to do that. So as I said, we're gonna go one use case per panelist. And Shani, let's start with you and the handover document or other handover capability. Walk us through the journey of taking that from kind of idea into adoption and value.

Yeah, so the sales does yes. And over, the idea is pretty simple. Basically, we want to eliminate all the blind spots between sales and CSMs. Because when you think about it, that moment when a new customer is ended over, can be pretty broken.

Like as a CSM, you're expected to lead the conversation, to gain trust, to build relationship with the customer. And sometimes you don't even know what is the real reason or story behind the deal. So data also can be spreaded around. It can be incomplete, sometimes completely missing because it was not documented anywhere.

And that caused a lot of frustration for the CSMs. So we kind of asked ourselves, what if AI could piece the entire story together for us and give us visibility into what happened during the sales cycle? We created a very structured prompt that basically goes to all the tools we use, and we use a lot of tools. If it's a Slack, says for staircase, our CS platform, Zendesk, Jira, everything, and bring in insights that we didn't have before.

For example, why the customer bought us. What is the problem they're trying to solve? Are there any commitments, specific risk we should be aware of, or the main stakeholders we should contact with? So this kind of information suddenly is visible to us, which was not before.

And the result is basically having a one-pager that would generate approximately 48 hours after the deal closes, and it is being sent automatically to our Slack internal channel, so it's visible to everyone. And we can already see our CSM feel much more comfortable. They feel like they can engage with the customer much faster, and we see the engagement and trust much faster than before. And I wanted to quickly drill into the trust, because trust is very much fundamental to driving adoption.

Your customers need to trust your products if you want to drive their adoption. Same with asking our teams to internally adopt new capabilities. How have you driven that trust? How have you ensured that the data that they're receiving in that handover summary is accurate?

Yeah, that was my biggest concern. And I think the way we position it made a difference, because we didn't say, this is your source of truth, and this is what you should look at. CSMs are expected to have a meeting with the sales before the kickoff, and walk through these bullet points and topics that you received from the handover doc. So basically we are kind of giving them some points to discuss.

Another thing is during the process itself, we had a lot of iteration with the sales rep, and we received a lot of feedback, so every time we kind of adjusted the prompt a bit until we reached to the final version, which is pretty accurate. And there's a distinction which I think is important when it comes to using AI in our teams to automate what they are already able to do, and are doing, and it's just a productivity gain, versus using AI to enable a capability that just wasn't truly feasible to get done at all by humans, because they don't have either the capacity just from a computation perspective in our human brains. How did you approach that sort of dynamic, and where did you see the leverage in terms of, hey, this AI is actually doing something that we could never have done through our human-led approach? Yeah, I'll be very honest here.

It's not that human cannot do that, it's just they don't. I mean, that's a fact. No one has enough time to review all the data, gather everything at scale, gain some insights, and what AI does pretty well here is, one, it brings a lot of data at scale and fast, and two, it also provides some visibility into topics that are easy to miss, all the side conversations, all the new ones, that sometimes are not documented anywhere. So then we basically create the full visibility without human too much involved.

That makes a lot of sense. So Christine, I'd love to drill into what you shared there around the knowledge base and how that's been connected to increasing agility inside Securonics. I think this goal of having this sort of oracle of truth about the customers that, again, going back to my cross-functional point, is so powerful. How have you been approaching that and driving success there?

So it does build on what Shani just mentioned. No CSM has the time to do the archeological digging necessary to go through Confluence, Jira, support tickets, the game site stuff. There are so many sources of information and so many people that have the information in their head, and if you're lucky, they documented it, so you can find it. So the pressure sandwich for the CSM and the TAM in our company is real, and the pressure sandwich comes in the form of two sides.

So you've got the internal perspective which says, okay, you've got this yellow customer, what are you doing about it? What's the plan? How are you gonna handle that and is it done yet? And the force behind that isn't just the executives, it's also support.

You're having to be the front end for them. Got a ticket, you've got an issue. And if you've got a product issue, you're also fronting for product. So you've got multiple organizations coming to you internally, and you're facing that internal pressure because you're having to represent the customer perspective inside.

Then you flip the coin, you say, oh, now I've got the customer coming at me. They wanna know how come my issue is not resolved yet. Where is my request that I asked for from the product team? And how come I know as your customer, there's something on the roadmap before you do.

And they also wanna know, how can I get more value out of the product? And I expect you to be telling me how to do that. So there's a lot of memory retention, and that does require two sets of information, and they're separated in the company often because Gainsight has a lot of customer information. So the lens is taking the internal inside and then internal out, right?

So that bi-directional communication and saying, how can we make the most out of understanding the sentiment, what that means, how we can apply that, how we can use the agents. And that begins from the time that they book the deal all the way to renewal. So it's the day-to-day meet conversation that the CSMs are having. And the real breakdown is connecting that information that we have within the Gainsight platform along with the product information and support that what are we building?

What are we talking about amongst ourselves? How do we surface that information so that it is trusted, it has boundaries, and you feed it this information that you know within the company is trusted. So product has this defined area that they are putting only customer-facing information. You know that that fact base is trusted.

You know that support cases are real. That information isn't up for debate, not up for discussion. And then when you have Gainsight that's using staircase information and emails and call transcripts, those are also fact bases that you can use. Those aren't up for debate.

That happened, this conversation, this thing was said, we promised that thing. All of that comes together to create a complete picture and you add on top MCPs that make that more extractable by anybody in the company. So the result that we ended up seeing is executive prep on demand. So the CSM no longer has to go in and say, spend a half an hour an hour prepping an executive in order to have a conversation.

You also don't have to prep engineering on what the issue is and why that's a problem for the customer. Then you get into making sure that you see other results. Like we saw a 15% improvement in our MTTR pretty much right away on the support side. And we saw like that research did give time back to the CSMs.

This is about 100 days of time in what we've built out so far. And so that's a substantial give back and that's just the early days adoption of it. Amazing, amazing. So Brad, we have to talk agentic, we have to talk agent.

So I wanna be very clear with Brad, what you and your team have been pioneering and part source is agentic in the context of taking action with customers, which agent, agentic, very hyped up term right now, used to cover a lot of bases, what we were describing as generative AI. Yesterday has been kind of described as an agent today. Fair enough. Where the rubber really hits the road is we are using agents to execute with customers, augmenting our humans.

You've been driving an initiative around GainSight's Atlas agentic platform. Walk us through the journey there and how you've been seeing value, lessons learned, what hasn't worked. Yeah, so I mean it really builds on the common themes here, right? I think we're all talking about, and that is how do we give our people the right information so they can really operate at top of license, right?

And have them focus on the things that really add value to the customer, which is spending time with the customer like where it matters most. But all this stuff still has to happen, right? As a part of it. And so we have a, what I'm gonna say is a long, long tail, which is where we decided to kind of initially start to explore these agentic, renewal capabilities with GainSight.

And just to set the stage a little bit. So in this long, long tail, we have 20 million and the average kind of contract value in that book of business is under $20,000, right? As a part of it. And so that's what we wanted to focus.

And just to set the stage a little bit more. So in our enterprise space, and we heard a little bit about this from Arli earlier, we had our GRR rates at 93% and we exited last year at 99.6%. And that was because of the renewal framework that we built out, right? And we built a really strong renewal framework that takes into account customer health, a clear sequence of activities that we took.

And we had a really strong governance model around that to making sure we were executing every step of the way. So then the question really became, how do we take this model? And how do we actually scale this model across this long, long tail of 1500 customers? And to do so really in a fiscally responsible way.

And I love what Chuck said around like, how do we provide that same value and same experience, to all of our customers and do so in a fiscally responsible way? That's what led us to Atlas and starting to work with GainSight. Because we have three employees that essentially are managing this long, long tail. And so the way that we've been managing it has been purely tactical.

30 to 60 days before renewal, we send out the renewal and we hope the customer signs. And now we could have said, hey, well, why don't we send out an earlier notification? But the question became, well, that's great. We know we're gonna get a lot of questions back.

How do we just manage the responses that come back? And so the way our new renewal kind of framework works and what we've built out in terms of taking this enterprise approach and scaling it, right, using Atlas, is we start six months prior to a customer's renewal. And we now have the Atlas agent reaching out for us to the customer to kind of get a pulse check on, do they plan to renew as a part of it? Now we'd previously been using a BPO service to help us with some of the processing.

So we had really good documentation in terms of, hey, here's the things that can come up during the renewal time. And here's how we respond to those things and our policies for doing so. So that was really helpful as we built out and deployed the Atlas agent. And what that allowed us to do is we essentially identified, hey, there's 31 common cases that make up 90% of the responses that we get from customers when they respond back to us on renewals.

And so we created a RACI model and we essentially said, hey, how do we take the management frameworks that we all know today and how do we start to adopt those management frameworks as we think around agentic capabilities? And so our account services team, the leaders are still accountable, but we now in these certain steps we've included the Atlas agent. And the Atlas agent is the one that is now responsible, but we still have people that are accountable for overseeing the agent. And just like any agent or just like any human kind of labor that we have as well, you have to coach, you have to train, you have to give feedback.

And so you have to treat your digital labor the same as you do your human labor and you have to build those things into the way you go about operating and rolling out agents. And that's an approach that we found to be highly impactful in doing so. And so with this initial kind of rollout that we're doing with Gainsight, we calculated the time savings in terms of, hey, if we wanted to go deploy our enterprise renewal methodology and to do so in a scaled back way, the first step would have taken us essentially 12 minutes per and the average back and forth that we estimated would have been kind of three, back and forth across the board. Some obviously higher, some having none as a part of it.

And so when you calculate that out, that's 900 additional hours of human capacity that we would have needed, which we've been able to avoid by working with Gainsight here. And I think the last thing I'll just layer in is so we've set our goals as well in terms of saying, hey, not only how are we improving GRR, but how are we improving our on-time renewal rate? And we're now tracking those metrics and working with Gainsight through the process. Amazing, so much to unpack there if we had time.

I encourage you all to connect with Brad afterwards. And he's very generous with sharing some details behind all of that. I'm looking at the Slido and we've got questions pouring in. So we're just gonna jump straight to some audience questions, which if our panelists can read that, I'm very impressed.

I've got the questions here on my phone, so I'll read them out and I'll attempt to direct them to who I think they are intended for. And I will go by popularity according to the app. So first one, I think Shani is definitely for you. The question is, for the sales to CS handover process, where are you getting the ingested data from sales?

How are the sales reps logging pre-sales calls, customer communications and meeting notes? Do they use Gainsight? Yeah, so the sales reps use Salesforce and all the calls and all the interaction are being recorded via GONG, Zoom, again, Teams and so on. We use a staircase to collect all the information and via the MCP, we basically combine all the relevant information from different sources.

So whatever is documented in Salesforce, whatever we managed to catch from GONG calls and recordings. And obviously, the output of Gains, of Staircase AI, so we can basically put together everything. So it's a combination of documentation, AI and call recordings. Awesome, very good.

This one is, I'd say, a more open question. So anyone here, please chime in. The question is, I'm curious to hear more about how you will measure the impact of AI for CS, what metrics you track, where do you see the biggest efficiency gains? Anyone like to take a stab at that?

I think generally, if you're looking at your friction points, where you're spending the most time and you typically know where that is, those are the spots to focus most because the friction and reducing that will have internal support. And that's typically what you need in order to gain traction with any AI program. And so, I think you know intuitively and can measure that and take the impact of that. For us specifically, there's a lot around support cases and getting the right information in the hands, like do you know the product?

So there's no support case, putting it out on communities so that there's self-service. There's a lot of things around support that can actually drain the whole organization. And so taking that off of our CSMs or TAMs and putting it in places where that self-service can happen was one of the major areas where we focused to take that friction away. Yeah, maybe I'll add two things onto that.

One would just be, I think we all have capacity models that we build as we think around staffing and head count planning for our CS teams. And so one of the things we do is we look at where are our teams spending time that aren't customer facing? And then how do we focus on those things and make sure that we have our team spending as much time as possible customer facing? And then the second thing we look at as well is we use kind of what we call post-sale attribution.

And how do we actually look at the touch points and the things that are happening and how are those things driving growth? So meaning what are the sequence of activities in the engagement models that we have and how is that actually leaving and generating growth, not just retention as well? I'll maybe add that we basically, we look at churn obviously and retention. And I think once we identify risk in advance, this is a point when we really want to see what's going on with this risk and how do we mitigate them?

So this is something that we track constantly. And I maybe also disclose some secrets, but as an ops person, I also ask to review exactly how many tokens or how many calls the CSMs are doing with using Cloud MCP to understand if they actually use AI, because as we mentioned, it's a top down approach and we need to adopt AI. I was gonna follow that thread and go to a question here from Roman around token usage. It's a hot topic right now in the press around incentivization within companies to drive token usage as some sort of meritocracy approach to AI adoption.

The other side of the coin is that budgets for tokens already very high, they're getting burned within a quarter when it's supposed to be intended for the year. So how are you sort of approaching balancing? So the exact question from Roman is how you enable your team to leverage AI within their workflow so that the usage of AI, I think meets the business requirements, but mitigating token usage with use case duplication, i.e. how we avoiding that duplication of use cases, which is burning tokens that could otherwise be used for more distinct new use cases.

How is that token management being considered? At least for us. You're trying to use the native inbuilt AI within the tool and let that take the work burden because Claude, chat GPT, they're probably an extra on top of other tools. And so being cautious, having a plan for how you're gonna use those as compared to letting game site that you've already paid for, take the brunt of that.

And so definitely think through what do you wanna use connected together, that's probably what you want Claude for or open AI, then use that, but think carefully, think that's use what you have already paid for, really well maximize that. Let your vendors pay for tokens. 100%. You've already paid, you've already made that investment.

Anything to add? I can add. Basically, it's a challenge because on one hand, you do want the team to adopt and have workflows and use AI, but on the other hand, you need to use it wisely. I think what we did differently is basically we created kind of a tiger teams, which includes CSMs and ops.

So they designed the solution together and not like everyone will do it by themselves as standalone, but as a team, and then you can spread the tokens and the usage between the arrived people. I think there needs to be greater token transparency to users though. I'll maybe share a story on this one. We actually tried putting a cap on token usage, but I didn't know we did this.

We rolled out Claude as a part of it. And so here I was on a Friday evening trying to process 75,000 call transcripts to identify patterns in those. And I didn't know, but I got a timeout and it was like, oh, you don't have enough tokens to go. And so I emailed the right team.

I was like, wait, when we open AI, we have chat GPT, I wasn't limited. Like now I'm limited on Claude. So we actually since removed that. And I think the short answer for us is right now, like we don't want to limit.

Like we want to give our people like that freedom to experiment. And look, I think there's good and bad that comes with that. I think there's, we probably all experienced some what I call AI slop, which is sometimes people just, oh, let me punch it into AI. I'm gonna send it out there.

And then I'll send you get it in your inbox and you're trying to parse through it. And so I think there's pros and cons, right? To kind of both sides of it. But right now, like we aren't like really pushing back.

We want to encourage experimentation. Great, I think we have time for either two very quick questions or maybe only one. And we'll go to a question from Taylor from Securonics. Many of us in the room are the change drivers for our company.

A challenging part of our job is the context switching. We are pushing for innovation across many initiatives in back to back 30 minute meetings. How do you think AI can help us be more effective in our roles? I don't know if Christine, you want to take that?

(laughing) Or maybe you would like one of your colleagues here to take it and. I would, I'm very curious what my colleagues have to say. I think that's pretty straightforward. I mean, I'm reading the question again.

Basically, if you know how to use the right data, and that's always a challenge because we all know that AI is not always 100% accurate. But if you manage to identify the right data and to bucket it into relevant topics, then you can just prepare yourself before meetings. You can be much more effective. You can have meeting summaries, you have follow-ups.

So basically you're much more prepared to every meeting and then you can focus on what you want to gain with from this meeting. I mean, for me, I think it's hard to just keep up with the pace of innovation, AI itself. And so I think it's just knowing that the technology is so far ahead of where we are, just from a human perspective, being able to understand and adopt it. And for me, it's all around how do we continue to experiment and continue to explore, be open-minded as to how you can use it.

And then it just becomes down to ruthless prioritization. Always saying, understanding what are the top priorities and making sure that you're carving out the space and prioritizing your day, to really focus on those things. Because I say the to-do list is never empty and never goes away, but it's just understanding what are those top things and then making sure that you're allocating your time and those things accordingly. And adding to that, I think the problem is trust.

You have to build in some level of trust that there are some things that you can count on the AI to handle for you. And one of those might be the follow-up, like providing that initial just cursory review and consent. And that will help you offload some of the more simplistic things that you let pile up day to day. I think that is probably also a really good thing.

And that will also teach your AI. That's a virtuous cycle that will end up helping you throughout. Let's go for the Hail Mary finish. We've got red numbers.

We're gonna do one more question because I like this question. Where did you find focusing on AI did not work? Any projects you just decided to scale back and refocus? Maybe just give sort of a headline answer of any initiatives that you tried and were like, oh, no.

I think everything that is top-down rollout or approach won't succeed. I mean, we try to involve the team as soon as possible so they can define the problem, design the solution. And I think most importantly, present it to the team. And then they adopt it and trust it much faster than any, you know, I tell you to do that and that.

So it's just a general answer too. I think for us, it just comes down back to the systems thinking. If you don't know what the process is, don't go try to use an agent to solve your process problems. You have to get intimate with your business, make sure you understand the process and design the process, and then figure out how do I use a genteck in AI capabilities and how do I get it to the top of the scale.

Friction, like a non-friction thing isn't gonna help you and you're not gonna get support for it. So if you're not suffering from friction and you try to use AI to solve it, I think you're not gonna get support and it's gonna fail both at the top and at the bottom. Amazing, thank you all for joining us today and big thank you to our panel. (audience applauding)