The AI-Powered Renewal OS: A Practitioner’s Blueprint for Forecasting and Retention at Scale

[Speaker 1]
Hello everyone and thank you so much for being here today. So as James said, today I want to show you what I built in the last few weeks and months. Now the time is a bit blurring because I'm always like building a bit on my operating system, as we called it here.

And more importantly I want to show you and discuss with you how I build it and the logic behind it, so that when you go back to your office on Wednesday after this conference or Monday, you can take hopefully some ideas and some thoughts out of this and build something yourself. But let's start from probably a recurring problem and a common problem that we all have. And I believe most renewal leaders are probably running the numbers on Google Sheets and manual calculation, manual forecasting and also a lot of like reconciliation of multiple files and multiple sheets.

And most importantly when a leader comes to you and say, oh can you tell me why you are committing this number, you might not be able to justify the number with a logic but more with a feeling. I believe that will land here because I feel it. But none of this is really a problem of negligence or not being able to do your job well.

It's probably because you don't have the right tools or the right reporting that allows you to do this kind of stuff, but in the meantime the deadline and the expectation are still there. So let me tell you a bit about my story. In August 2024 I joined a company called Own Backup, where I was leading the customer success team.

Great company, so good that six months in Salesforce decides to buy them. Amazing. What happens is that we move into Salesforce not anymore as a customer success team but as a renewals team.

So you can imagine we have now a different mandate, massive company behind us, different team. So we have basically to be very quick and be there to make sure that we are still running the business in the same way that we were doing before, although with a different role and a different kind of context around us. So where we were coming in Salesforce as an acquired business, obviously you can imagine that we didn't have the same level of reporting and dashboard that Salesforce has because it's an acquired business going into a larger business.

So most of my job was done through, as I said before, running numbers manually, downloading Salesforce reports, calculating my commit numbers manually, and at some point I was like, there needs to be some automation here. And that's where I started building something myself. Interestingly, 16, 18 months later, I moved into a different role.

You might have guessed that I'm Italian. So there was an opportunity to go and manage and lead the core business renewal for the Italian market. So I took that immediately, and it's been an amazing journey since.

And as you can imagine, the context there is completely different. There is a mature organization, a lot of reporting, a lot of dashboards, but at the same time all these reports and dashboards were not built for me. I was still needing some things that these dashboards and reports cannot give me.

At the same time, I only had a static view. So I couldn't, when I was building my commit number, my best case number, I couldn't really model and test different scenarios. And I was doing, again, the same calculation every single week.

So as you can see, I had two very similar problems, despite being in two completely different organizations in a way. In the acquired business, small business, lack of in-depth reporting, manual work, and obviously a fixed way of working, so I couldn't really create models. Similar in the core business, the reporting was there, the dashboards were there, but they were not created for me, and they couldn't really answer my questions.

So what I did, I, looking back at both scenarios, I applied the same logic to both problems, because they were very similar problems. So first of all, I defined what I was needing, right? What I need to know to be successful at my job.

Then I started identifying all the different specs, all the different rules that I needed to have clear before building something that was useful for me. And then eventually, I started building, which is what I'm going to show you in a second. So in the first framework, so the one of the acquired business, I had to provide two different numbers.

So the worst case and the best case. I don't want to go too much into details how I calculate these two, but just for your knowledge, so the worst case was basically the previous annual contract value minus the attrition. The best case was previous annual contract value minus attrition, but then I also start incorporating an element of recovery, which we call swing.

And on top of that, I also apply something called the manager forecast judgment. The manager forecast judgment is basically a probability of recovering that swing, and it's a probability that I apply to the swing based on conversation with customers, with the account team, with the renewal manager. Just bear in mind, this is not something that I use to override what my team is saying, but it's something that I'm using to make the number more credible and something that can survive a QBR or a forecast call.

So just to give you an idea of how the MFJ or the manager forecast judgment was set, I had in, which was basically something that I believed I could recover, I applied an 80% probability, up plus, 0.6, up minus, possible, like a stretch, 40%, and then out, something that I don't believe is going to come back, and it's 20%. So before building my dashboard, what I had to do, I had to go into the Salesforce report every single week, see what changes the team had made, and obviously it wasn't that visible because the report in Salesforce was just the last update, right, and we didn't have any dashboard providing the previous numbers. And then I had to download the reports and make the calculation by hand and then provide those if there was a significant change, provide those to my boss and the senior leadership.

As I said before, at some point I said, okay, this must be something that can be automated, and that's when I started building my dashboard. So this is the first dashboard that I built. It looks probably familiar to some of the dashboards that you use.

It's got the total ATR. Oh, by the way, the numbers are all fictional here because I can't show real numbers, but just for your visibility. The renewal worst case, the renewal best case, the forecast attrition, and the weighted attrition, which is what I said before, applying the swing and the manager forecast judgment.

You have your average contract term, all the filters built there, and as you can imagine, this was like a living creature that I was building and modifying going forward and using it because every time I was using it, I was finding something new that I needed. Then at the bottom you can see all the different regions that I was managing with all the relevant numbers. Then you could scroll down and you have all the opportunities, again, with the numbers that are relevant, when the contract is ending, what stage of the opportunities are, and then clicking on every opportunity, you could open a box on the right which is showing also other important information like the next step, the manager notes, when the contract is ending, and so on and so forth.

So this is giving me an overview of my book of business. The next thing that I wanted to do is how can I calculate my forecast automatically, and this is what the next tab is doing. So here, as you can see, you have the region, again, with the available to renew, attrition commit, attrition best case, and how these numbers have moved week on week.

The interesting thing was that when I was clicking on every single region, I could also see exactly was, obviously, this is just an example with one opportunity, but I could see all the opportunities that were moved by the renewal manager and were changed by the renewal manager. Clicking on every opportunity was also giving me not just the change in the number but also how the next steps had changed, how the manager notes had changed, so I could actually understand also the context behind the change, and then I could actually make my call and say, okay, actually, this is a change that is relevant. I want to call the new number and adjust my forecast.

So this is just on visibility of book of business, how the forecast has changed, how the forecast is changing week on week. The next thing that I wanted to understand is also how accurate are my renewal managers, and to be fair, this dashboard was already provided by us by Salesforce, sorry, to us by Salesforce, but me being me, I wanted something in my dashboard and customised by me, so I built something else, and what I built was a tab, a page that was basically showing me how my renewal manager were calling the attrition amount one week before, one month before, two months and three months, and that gave me an idea of how accurate my renewal managers were. I could understand whether they were calling attrition or renewal correctly. I could understand if they were systematically over or under calling, and how accurate they were in calling the attrition, and also how early and how proactive they were in updating the next steps, which for me is a proxy to understand how early they engage with either the customer or the account team or the customer success team and so on and so forth.

So this is what the renewal manager forecast accuracy looks like. As you can see here, you've got the renewal managers, their accuracy, luckily my team wasn't that bad. It was much better than this.

You can see a lot of red there. You can see the accuracy three months, two months, one month in, and then here, the other important bits that I needed, so how many opportunities don't have next steps, how many next steps are stale, so they haven't been updated within 14 days, but if I wanted, I could change that, any timeframe that I wanted, or for something above 50k, which was a threshold that I set myself, how many were not updated in the last seven days.

Why is this important? Not because I wanted to micromanage or I wanted to call out what the renewal managers were not doing, but because when I was having my one-to-ones with the team, I could change the conversation from, oh, I think you are being optimistic here, to something that was based on numbers. So I could say stuff like, oh, actually, I noticed that in the last three quarters, you are systematically calling less attrition, or your accuracy two months off is off by an average of 38 per cent, or something that I found very interesting, you don't update the next steps until a month away from the renewals, and, funnily enough, I noticed that whoever was doing that was the renewal manager with the worst renewal rate. So coaching moved from just a feeling to something that was actually number-based, and number, and evidence, right? And this is what can actually help with the actual retention number, and this is what actually can protect your number.

So, as I was saying before, 16 to 18 months later, I had this opportunity to move into a new part of the business, the core Italian business, and going into this new role was interesting, because I already built something that I found extremely valuable, so I could go into a new role with already knowing what I needed. But, funnily enough, I didn't go straight into building, because I realised I was still missing exactly the question that I needed to answer, right? So I took some time before I actually started the building, and we will talk about this later, this is the most important thing, because I think the most important time when you are building something is not about the building, but it's actually about the thinking and creating the specs of what you want to build.

So, in this new context, in this new scenario, I wasn't asked to provide only two numbers, but now I have three numbers. I had a worst case, a best case, or actually a worst case, a commit, and a best case. So this is what I focused on at the very beginning.

So my worst case was actually my full PACV, my full annual contract value minus attrition. My commit was including the swing and the MFJ, and then the best case was my ACV, so my annual contract value, minus a percentage of the swing that I could customise myself, and I'll show you how I set this up in the dashboard. So it was a configurable percentage of the swing recovery.

Then, if you remember, at the very beginning, I said that the reports that I had were static, they were not allowing me to build my own view of how the scenario can change. So I built something that I found extremely valuable, which is the team view, where basically it pulls the data from Salesforce, and then my view. It still pulls the data from Salesforce, but I can actually apply some changes so that I can test and simulate different scenarios.

So in my view now, I can change the swing number, I can change the MFJ number, so that I can say, okay, what happens if this opportunity doesn't close or close with less than we expect? How is my number going to change? Another box and tab that I created that I find extremely valuable is what moved.

I mean, we all work with a lot of clients and customers and a lot of large teams, right? So focusing on what moves the number is probably one of the most important things, right? So week on week, how can we check what is actually moving the number?

So I created a specific section of the dashboard where I can see exactly how the numbers have changed and what has changed. And then I was also able to add other elements that are answering some other business needs. So, for example, how is my model differing from the official forecast that I put into the system every week?

How can I justify my commit using what I will show you in a second, the commit waterfall? So the final number, what are the elements that are combining and that are forming my final number? And also different things that probably are specific to the business you're working on.

So how many of cycle renewals do I have? How many recapture revenue do I have? All these little things that come up while you work in a complicated business that you need to account for.

So here it is, my second dashboard, which I find a bit more sophisticated than the previous one. Here I can see you have my worst case, my commit, my best case. You still have the filters to check the different views.

So my view, sorry, quarter of a month. You can split, you can switch between team view and my view. And then you have other like elements and metrics that I find extremely valuable.

So the PACV in scope, what is the raw attrition forecasted? How much of this is of cycle? What is my swing?

How is my swing split between the various MFJ and what is the weighted swing at the end? So here in a snapshot I've got all the information that I need when I go into a call, in a forecast call, when I go into an OU leader at Salesforce. As before, I have the view of all my opportunities and I can click on every opportunity and I can either go into Salesforce directly or I can open a box and see all the additional information that I need.

And here this is what I was saying before. So here I've got the attrition number, the swing that the renewal manager is calling, and then this is the box. Let me just check that.

Yeah. This is the box where I can input my swing. This is the renewal manager MFJ, and this is where I can simulate and change the MFJ so that I can make some adjustment and changes to the final scenarios.

Any of the other stuff like manager notes, renewal manager notes, and so forth. One of the most interesting elements of this dashboard was the setting tab. In the setting tab I can do some other simulation and changes.

So I was saying before the MFJ is basically a probability. So in here I can go and change the probability depending on changes in the market, changes in the business. So I can go and change with a slider every single probability of the MFJ.

And as I was saying before, I can change the percentage of the best case as well. Let's say that I believe that the best case is 80% of the swing, I can put 80%. If I'm a bit more optimistic that day, I can switch it to 90% and so on and so forth.

And then in here I've got other elements that I was saying before. You have the off cycle, you have the recapture on revenue, the hedge that I want to apply to the numbers or the buffer that I want to apply, and different months, the notes at the end. And at the end of the month I can actually consolidate the month, which closes basically the number for the month, and I can move on to the rest of the quarter.

Before I mentioned the commit waterfall. So sometimes when I go in a forecast call they say, okay, but you are calling 7.9 million. How do you justify the number?

How do you get to the number? And this is the answer. Here I've got exactly all the elements that are making my number.

You have the closed, the open, on cycle, off cycle, which gives me the total attrition. Then you have the recovery, the swing, and then all the additional element that I forecast on top, which gives me the commit. So next time the OU leader comes to me and says, yeah, but I don't believe in this number.

How do you justify it? Say, this is how I justify it. These are my elements.

And if I click on any of those columns in the Instagram or in any of those elements in that commit waterfall table, I can then filter in the opportunity and see exactly what are the opportunities that are into the MFJ in, MFJ up plus, and so on and so forth. Here you've got the movement box with attrition swing, how many new risks came, how many opportunities have closed week on week. And then clicking on any of those box, you have, again, the details that you need.

You click on one of them. And actually this is something that I made very recently. You click on one of those opportunity, and you can also see how many times that number has changed.

So you can say, okay, this number is changing a lot. Let's see why. Let's see what is making this change.

So again, I can go and talk to either the customer or the renewal manager to understand the background. So as I was mentioning, the build is the easy part. And I'm going to tell you about how I approached this.

But the spec was the hard part. And learning from my mistakes, I probably started building a bit too early, because then you have to make a lot of adjustment going on. And you have to fix issues and problems that come up during the way.

And also, if the logic is vague, or if what you want to understand is vague, then AI will build something that is very vague and useless very quickly. So let's discuss a little bit how this was built. So first of all, I approached this as a conversation with Claude.

So I used Claude code to build this dashboard. But rather than going into Claude code immediately, which I still use a terminal, which is a UI that is obviously, if you haven't played with MS-DOS when you were little, you might find a bit baffling. So I went on Claude, actually it was my personal Claude, and I started to say, okay, this is my problem.

I want a dashboard that can give me this answer. What can I build? And then through a lot of iteration, without any data, without any connections, I started building my visual mockup.

So once I have something that I like, and that I think is answering my question, then I start building all the logic that I was saying before. So the best case, the commit, the worst case, the MJ, and all the stuff that you need to build that is specific probably to your business. Once I had that, once I had the visual mockup ready, once I had the specs and the rules and the logic ready, then is when you move into Claude code.

Also because moving into Claude code means that you're going to spend way more money, way more tokens to build something. So what I did, I went again on Claude and I said, okay, now that I've got this ready, build me a prompt that I can use and paste it into Claude code. So Claude built me the prompt.

I then went in Claude code and then I started building. After I had the basic of my application, I then connected the data with Salesforce, which is probably the most important thing, right? We talked, we had before during the keynote how customers are going crazy for the MCP of Gainesville.

This is like probably across every business, opening up the MCP and opening up the connections between your LLM and your data is opening up a lot of possibilities. So this is where I connected the data between my dashboard and the Salesforce data. And from there, I started pulling automated timestamp snapshots that are building the information that I need week on week to then build my forecast and build and understand how my team is operating.

I did play with MS-DOS when I was little and I did play with QBasic as well, never built anything. I'm not a builder. As I said, I've got a background in economics and I never built anything in my life.

So this was built entirely with a conversation, no engineering ticket, no handwritten code. So this is entirely conversational. Obviously, this is not as shiny as it seems, right?

Probably all of you here have tried to build something and it's not like, oh, let me build something and then it's ready after a day. No, it's not like this. So first of all, one important caveat is this does not replace your proper tooling.

If you have already a company that is providing you reporting and dashboards, I recommend you use those first and then you use AI only to answer those questions that are not answered by those reports. AI is great for building something fast, but if you want to have scale and governance, it's better to use enterprise-ready reports and dashboards. Also, personal tooling breaks.

It does break sometimes and it doesn't scale past you. I can't really pass it easily to somebody else in my team. There needs to be some work there.

The other thing is before I was ready to use this dashboard with my boss and with the senior leaders, I had to check the numbers multiple times because the last thing that you want is you go to your boss or to your OU leader and you give a number that is completely wrong. It's like, oh, they never trust you anymore, right? So check the numbers, check the numbers, and then check them again before you're really confident that the dashboard is really providing you the right number.

So I still do some spot checks every now and then to make sure that when I'm making major changes, I'm not breaking anything. And the last thing is really garbage in, garbage out. And it's not just about the data, which, by the way, is fundamental, right?

So having the right data behind is what allows you to build this. So we talk a lot about apocalypse, but AI is still not there to allow you to have the data order and organize in a way that can allow you to build this, right? You still need the system of knowledge behind it.

But also the logic and the specs need to be strong and resilient because you don't want to build something that doesn't make sense. So spend time on building the logic before you actually build the tool. All right.

So I want to leave you with a couple of things. As I said before, Wednesday, after this amazing conference, or Monday, I want really to encourage you and to start thinking, what is it that I'm doing every single week and is manual and is something that I can easily automate? Start from there.

Start from something that is manual and you do repetitively every single week and start thinking, how can I automate that with AI? That's number one. The other thing is using AI frees a lot of time, and it's great, right?

You can have more coffees during the day. You can go and have your lunchtime run without any problems. But what I really suggest you do with this extra time is spend the time to coach your team, right?

Now that I don't have to do the forecast every single week, I can spend more time with the team and I can use the forecast accuracy tab, which, by the way, I built also in the second dashboard. I can use the forecast accuracy tab to have a conversation with my team and talk together how we can improve the forecast. And also spend time on the deals itself.

If something is not working, if there is a renewal risk, now you have more time to spend on the and this is what actually eventually will move your number, will improve your retention rate. And with that, I want to thank you. My name is Riccardo Composto.

This is my LinkedIn. Please add me and I'm now free from some questions. And I'm here today and tomorrow.

So if you want to have a chat, I'm more than happy to chat with you. Thank you so much.

[Speaker 2]
Thank you very much, Riccardo. So we've got about 15 minutes to answer questions. So just to reiterate again, as it's the first session of the day, you can submit your questions via the app if you click on the track and then the slide over.

Also upvote any questions that may already be there. So, yeah, thanks for that insightful presentation. The first question we've got here is so you built all of this without a development team.

What was the hardest part and what would you tell someone to skip or buy instead? Great question.

[Speaker 1]
Yeah, it's a great question. The hardest part was starting. We were very lucky.

We had the opportunity to we were given Cloud Code as Salesforce to basically experiment with. Now they're limited there after the big everyone was starting building something. So I was one of the early adopters and as soon as it was available, I took Cloud Code and I started playing with it.

And obviously the hardest part is starting and understanding how to navigate Cloud Code, how to connect the MCP. It took some time. It took actually a couple of hours of iteration with Cloud Code.

So it's really like understanding what you need to build first, as I said before, the specking and logics, then getting started, like overcoming the first hurdle. This is probably the hardest part. In terms of what I would tell someone to skip or to buy instead, I don't know about buying, but definitely to skip, as I said before, don't build something that is already there.

You don't need to. It's a waste of time. It's a waste of money.

It's a waste of token. Just focus on what is not available to you and what can actually augment what you already have.

[Speaker 2]
Nice. I think this next question is talking about for a lot of people that is using AI. So how do you make sure the numbers coming out of AI, of an AI built tool are right, especially when leadership is relying on them?

[Speaker 1]
Yeah. Yeah, this is a great question. And as I said before, be very careful, right?

Because you don't want to share your numbers straight away without checking. So first of all, check the numbers yourself. You can do spot checks yourself.

The other thing is I was, AI makes us a bit lazy, right? So what I was doing, I was taking a screenshot of the filters that I had in Salesforce and in my report and say, I want to use exactly these filters, right? I want you to use exactly these filters, exactly these logics.

So until the number was right, I was going back and forth Salesforce report and Cloud Code until I got the numbers right. I was downloading the CSV from Salesforce, putting into Cloud Code and say, check the discrepancy. Why do I have a million more in my Salesforce report compared to the dashboard that I've created?

So just basically question, question, question until you get to the numbers. So check the filters, copy and paste the filter into Cloud Code, download the data, put it in Cloud Code, ask AI to check the numbers, and then check everything. And if you remember, in my dashboard, I had the PACV in scope, the attrition, the row attrition.

So these are numbers that I can check quickly, right? I can go into the Salesforce and see, okay, PACV, yeah, much. Attrition, yes, much.

So this is all stuff that you can check very quickly and you can be confident that you're not calling a number that is completely wrong.

[Speaker 2]
Nice. And I think the next question, I'm hearing a lot of this within my customer base, especially around AI, but it's, yeah, how did you get security or legal to approve this? And if so, what pushback did you encounter?

[Speaker 1]
Well, yeah, as I said, I think I was very lucky because Salesforce encouraged us to experiment with Cloud Code. And Cloud Code is all part of the trusted layer that Salesforce has. So I didn't have to get any approval from legal or from security because it was already secure.

In terms of legal approval, obviously, I'm not sharing any numbers out. It's all within Salesforce ecosystem. So I don't need to get legal approval on that.

Obviously, if I was sharing my dashboard out, yes, that would be another discussion.

[Speaker 2]
Fair. Yeah. I think a lot of...

[Speaker 1]
So, sorry, if I can go back. So just another thing is if you have the decision power, I would recommend that you start thinking and influencing the rest of the business or IT security legal to add some sort of AI into the workflow of your business. If you don't have the decision power, then I suggest you start influencing above you to get this into the business.

Because at the end of the day, it will come. In a way or another, it will come. And also, the other thing is, if you allow your team to experiment and to be a bit bold with AI, I think you are building a stronger team with skills that they can use in your business and in the future in their career.

[Speaker 2]
And, like, Riccardo, you say you're a great people leader. But, like, how did your team respond to having their forecast measured at such a kind of forensic degree?

[Speaker 1]
Yeah, it's a good question. Actually, I would say that they respond well. I use the accuracy dashboard during the regular check-in.

Sorry, the regular one-to-ones. And I think it's how you put it, right? If you go there and you say, oh, you're not accurate, you're not good.

But if you use that as a consistent tool and number to help them be better, at the end of the day, it's a good tool, right? It's not you want to help them. You want to coach them.

And what you can do now is to coach them with numbers, right? You can whenever I was telling my team or someone in my team, your next steps are not updated early enough, it never created tension between me and my direct reports. It actually opened up a conversation on to why.

Why is the case? What can we do to do better, right? And sometimes it's because they had a big portfolio.

They couldn't do better than that. Sometimes it's because they are struggling. They don't understand where to prioritize.

So it really depends on how you use it. If you use it just to be as a measure, it doesn't go anywhere. But if you use it as a tool to coach them and you incorporate that into your conversation with the team, I think it's very powerful.

[Speaker 2]
I totally agree with you. I think from an individual contributor perspective, as long as that report is being used consistently and then it's there for you to be able to ask for help or show where there may be gaps within your book of business, then it's always useful.

[Speaker 1]
Yeah, absolutely.

[Speaker 2]
Great answer. The next question here is how did you handle the snapshots for historical data and what changed, e.g. did you have the MCP pull a daily report? How do you data?

Any challenges with this part of the build?

[Speaker 1]
Yeah. So snapshots are pulled automatically every day. Yes.

So through the MCP, I pull the data every single day. And it's stored locally in my drive, which is then backed up. So I have all the data safe.

And any challenges with this part of the build? Not really. As I said, the MCP was a bit challenging, but just because probably I'm not experienced.

I'm not an expert. So I remember when I did it the first time to connect Cloud Code through an MCP, it took a while. I followed some instruction that the team had set up for us, and it took some time, a little bit of iteration.

But then once the connection was ready, sometimes you have to refresh it. But other than that, it's very straightforward. So yeah, this is very simple.

Pulling the report daily. If I want, I can pull an additional report. So if there's been a major change in the forecast, I can click a button that says pull the report.

It connects to Salesforce, pulls the data, and updates my dashboard straight away. So it's not live, but it's refreshed automatically.

[Speaker 2]
Nice. We've probably got enough time for a couple more questions here. One from Harry in the audience.

Have you conducted any regression tests to see how accurate your renewals forecast is now versus previously?

[Speaker 1]
That's a great question, and it can only come from a great person. No, I haven't done that, but I think it's a great idea. And I think with more data available, this is definitely the next step.

So I was saying, use your time now to do something more useful, the time that you save. And definitely trying to understand how my numbers are improving now is probably the next important step. One thing that I'm building, though, is a stability tab.

So I'm still working on it, and I want to understand how stable my forecast call is and how much it changes. I'm still not there yet. Maybe next year I can come and present that.

[Speaker 2]
Nice. And the next question is great in terms of cross-functional collaboration. So how would you advise your team to work best with support teams and other client-facing teams to collect that more granular data?

How can you build this into a dashboard to increase the impact of your build?

[Speaker 1]
Another great question, I think. And, well, if you can, right, if you have multiple data sources in your business, or even if you have only one system, right, I think if you can make the connection with all the different data sources that you have, you can really start building something that can become even more useful, right? So imagine you are a customer success leader, so not just a renewal leader, and you want to start incorporating data from gains, for example, from other data sources.

I think you can really build a dashboard that is giving you an understanding of not just the renewals motion, but also how healthy the customer is, how they're using the product. So in terms of how I advise the team to work best with support and other client-facing teams, I think in my experience building this is probably I would recommend that you have an executive sponsor of this, and you create a task force of all the different client-facing roles and all the different teams that can be involved in using this dashboard and start modeling and thinking and brainstorming what exactly you need to create a dashboard that is more impactful, as you said, and more usable for the rest of the business. But with a caution, though, if you already have that, you don't need to build it again. Only if you don't have it, then I recommend you build this.

It's going back to the keynote and what Chuck was saying is think about what you can build, what you can buy, right? So there is a compromise there.

[Speaker 2]
I think we've got time for maybe one last question here. So how do you leverage customer health as an input to your forecasting and long-term readiness of the renewables to build durable?

[Speaker 1]
Yeah, great question. I've been thinking about that as well. We do have something at Salesforce called the attrition risk insight, which is very similar to, in a way, to the Gainsight Health Score.

There is something that I'm thinking I would like to incorporate that into my dashboard, because at the moment my dashboard is great for the transactional part of the business, the day-to-day, but it's not telling me in advance how at risk the renewal is. So that is definitely something in my mind. And I think, again, if you have the data available, then you can start pulling the data into the dashboard.

If you don't have something like a health score already built or a customer health already built, then the first step would be building that, thinking about that. And that goes back to my previous answer, which is get together the leaders of the main customer facing teams and start thinking about what are the elements that would make my customer health. Once you've got that, then you can pull them into your dashboard.

[Speaker 2]
Well, Riccardo, thank you very much for the interesting presentation, answering the questions. Thank you for submitting the questions. If everyone could put their hands together again for Riccardo, and we'll see you within the next session.

Thank you, Riccardo.