This Month at Salesforce | Ep 3 – Xero, Legora, and Replit

Hey, folks. Today is a very big day at Salesforce for 2 very important reasons. Now, first, Salesforce just reported our highly anticipated earnings. This is huge market-moving news. And spoiler alert, it looks like the whole SaaS-pocalypse thing, well, it was just a myth. But the other reason that this is a big deal is because it means that you get a bonus episode of This Month at Salesforce. Welcome back to a very special bonus episode of This Month at Salesforce. I'm your host, Shivani Ahuja, and we are about to get all up in the C-suite today. We've got a CEO, a CFO, and a CRO from 3 incredible companies. These 3 companies, Xero, Lagora, and Replit, all have very different approaches to AI. Now, Xero, All in with Agentforce. And Legora, well, they work with Claude and ChatGPT, and Replit builds with Claude Code. So you've got 3 different reasoning engines, 3 different interfaces, but all 3 run on the same trusted foundation. And that, of course, is Salesforce. So let's get started with Xero. Now, Xero helps small businesses around the world manage their finances. The company serves nearly 5 million subscribers, and that includes accountants and bookkeepers. Now, Xero's been a Salesforce customer since 2014, running on Service Cloud and now Agentforce. Joining me here in the studio live today, right now, this moment, is Xero CEO Sukinder Singh Cassidy. Thank you so much, Sukinder, for joining us today. We are thrilled to have you on the show.

Well, thank you for having me.

Now, Sukinder, how critical is Salesforce to running Xero today? And, you know, not to lead the witness, but put simply, what would become harder or stop working without it?

Well, first of all, as you noted, we have been a Salesforce customer since 2014, and we have scaled subscribers multiple-fold since then. Of course, we've also scaled geos multiple times.

Right.

And we use Salesforce not just as Service Cloud, but also Marketing Cloud and Sales Cloud. So it underpins all of our customer interactions, both outbound and inbound. So it is critical infrastructure for us. And so what would happen if, if we stopped using it? Well, even before we get into AI, I think our GTM would grind to a halt and our customers would be pretty disconnected from their their experience of Xero. But it is a very critical part of our tech strategy and stack, including in an AI-forward world.

I love that. And so you've got sales, you've got service, you've got marketing. It's almost like the pipes in which all of this rich data is flowing through. And now Agentforce. So talk us through a bit about your experience with Agentforce. What agents have you built? How quickly did you move from idea, from like this idea of an agent doing work to the agent actually doing real work?

Right. Well, I think at Xero, I think Agentforce has been, I'd say, foundational and instrumental to changing our customer experience in our support area in particular. And there are a couple of things I want to call out because I think they were things that were never before possible at Xero. So, Xero is built entirely as a digital-first customer support experience. So, believe it or not, we've never had call centers.

Wow.

Customers would have to self-service, which is great, on a site called Partner Central, which houses all of our repositories of answers and knowledge base. But there was no way to ever talk to anyone live.

Yeah.

And there was no way to get real-time support. So, as the team was building Agentforce or looking at Agentforce and thinking about what we could do with it, real-time support became the critical opportunity. About 9 months ago, maybe 10 months ago, we started talking about what if we could give 100% of our customers real-time support. I think the first fear was like, well, if we had an agent, and internally we codenamed this Lucy, and we offered it to 100% of customers, would actually— and we were willing to route people to an agent, if we could— if to a live agent, if the AI agent could not answer their questions, would we see explosion in staffing costs and volume and even the ability to talk to somebody?

Right.

Would that blow our gross margins? And instead what happened, of course, is we set ourselves a goal and Agentforce was the enabler. We built a custom workflow and a custom journey for our customers on the chatbot. We now just call it our customer service agent. We don't call it Lucy anymore. And I challenged the team when they had rolled out to 25% and we're seeing good results. I said, take it to 100. Their fear was, well, what if CSAT is not good enough, what have you? I said, look, this is an experience we've never had before. For our customers, it is something that hasn't existed and could only exist in an AI world. So we rolled it out to 100% of our customers just a couple of months ago. We're already seeing, of course, many, many questions answered instantaneously. I think it's a 62% deflection rate.

Wow.

We've had over 200,000 interactions, tens of thousands of interactions going on in any given month. And what we have really done is given them the ability to, A, get most of their questions answered instantaneously. For those that aren't, they are able to be routed to an agent. So now, for the first time, there is a human talking to them faster than would have previously been possible. And of course, we've been able to do this while maintaining, you know, what is a pretty high margin of gross profit or gross margin profile for ourselves.

And this whole idea, again, CSAT would have been zero because it didn't exist before.

Today, it brings something like—

Exactly.

So how do you measure something that hasn't existed before? Yes.

And it sounds like this internally named agent Lucy is not just stopping at service because now you're starting to build upon the capabilities as well.

Yes, of course. I mean, the capabilities, once, as you know, once you have an agent and you know it can work for certain use cases, you're going to start rolling in more. So now billing, we don't just do accounting, we do payroll and payments as well on our platform. And so now we can support payroll questions as well. And so, you know, that customer service agent is just getting more and more powerful. And so that has been one of the big agent force implementations. There's one more that is worth highlighting, which is our ability to really nurture leads and procure more leads. Like any business, you know, lead gen is another great opportunity. And here we've used Agentforce to build an AI SDR. And that SDR is taking leads from the website, from accounts and bookkeepers who come in. And it also will re-nurture old leads, effectively inactive accounts and bookkeepers who we want to re-stimulate. We see that we have increased our email open rates 3x.

Wow.

our AI SDR. Of course, we are booking more leads than would have been possible previously. And of course, we're avoiding the staffing costs of trying to do this manually at scale. So another really exciting implementation.

I love these use cases because they're examples of like something you didn't have before. Like you didn't necessarily have folks that could call back all these leads to now you're seeing outcomes. You didn't have enough folks to answer all the calls.

Well, I think you're hitting it.

This is wild.

Yeah, I think you're hitting the key opportunity, right? People always think about what you lose, but what you gain is a level of scale and the ability to have new use cases that previously just simply weren't possible.

That's right.

You just thought about staffing people.

Now, Xero spent more than a decade building its customer data and workflows on Salesforce. Like, this doesn't happen overnight. What has that foundation made possible as you've moved faster with AI?

Well, you've hit the critical word, right? Data. I mean, at the end of the day, Salesforce is, of course, a system of record for us, but it also provides us clean data, proprietary data, customer data. And it also provides the infrastructure around that— security, compliance, It's all the things we care about because we ourselves are a system of record for small businesses when it comes to their financial operating data. So I appreciate what Salesforce does for us in enabling us to surface clean data upon which to build agents.

Right.

And new journeys in much the same way we do for our customers. We are their critical infrastructure and we built agentic capabilities on top. So yeah, I don't know. I think it's hard to have powerful, you know, ongoing increasing reasoning without, you know, access to great data.

We always say trust is our number one value, and it is for a reason. It's so that we can extend that trust to you. Sikinder, I really want to thank you for joining us today. We are so proud of what we've built together, and we are so excited to see what's next for Xero.

Thank you.

Thank you. And up next, another customer we've got is Legora. Now, Legora didn't just bolt AI onto legal software. Nope, nope. The company was built around it, giving lawyers a faster way to handle contract reviews, due diligence, and legal research. And it's growing rapidly, expanding from roughly 600 to more than 1,500 employees in just 6 months. Huh, that's interesting. Joining us remotely is David Eckstein, CFO of Lagora. Welcome, David. It is such a pleasure to have you here today. Thank you for joining us.

Thanks for having me, Shabani. We're very excited.

Now, David, Lagora was built in, in the AI era with no legacy systems, no bad habits, nothing to unlearn. Like, that's pretty special. As a result, what do your employees do differently?

Great question. So I have lived in the SaaS world now for over 15 years, and I can tell you the first 7 months that I've been at Lagora have been truly transformational. The clearest tell is how we're organized. We don't have one central organization that you file tickets to.

We don't.

We have a go-to-markets operations team, we have a marketing operations team, we have a product operations team, and each one owns building their first AI workflows for its own function. It seems intuitive, but it's harder to do in practice.

Right.

The people closest to the work are the ones that build the systems for the work. So let me give you a quick example. A rep at Legora, they don't write their own outbound. They don't even type up their own call notes. That's all automated. What they do is they walk into a partner meeting at a law firm or a Fortune 500 company, and they already know the practice mix. They know the deal history. They know who's in the room. because the research was done before they even sat down. This limits the number of ad hoc requests. It also enables them to become more self-service. And we bring this back to the hiring phase. Every single contributor to our organization, they need to demonstrate that they can build their own agents, not do use ChatGPT.

Right.

We are a real show me, don't tell me organization.

This is incredible what you've been able to do. And so your employees, they're using Claude and ChatGPT to be these builders every single day. So where does Salesforce fit into that environment and why is it more important now than ever?

It is the critical system of record. So we moved to Salesforce consciously because at our scale we needed that enterprise layer, we needed permissioning, we needed audit trails, we needed governance. And we've been through several generations of frontier models since we've made that move, and we'll go through several more. The AI world is rapidly changing. But some things are constant.

Right.

What can't change every 4 months is that system of record. Who do we sell to? On what terms? Who should have access to what? If that layer moves, everything sitting on top of it stops being trustworthy. So when I think about Claude and ChatGPT, that's about how people work. Salesforce is what they leverage to get that work done. I'd actually push back on any of the framing that says that that AI layer will eventually replace the system of record.

Right.

To me, it makes it more valuable Because the agent is only as good as the context that you give it, and that context has to come from an authoritative layer. So Legora is the fastest software company with a sales force to go from $1 to $100 million ARR. We did it in less than 18 months.

Wow.

And we cannot afford to change our CRM again.

No. And David, you're talking about some incredible concepts. You've grown from over 600 to over 1,500 employees in 6 months, like in addition to your ARR. Everyone talks about AI replacing jobs. Like, this is another angle and another dimension. You're proof point that it doesn't. So what are all of these new people actually doing?

Yeah, so to rewind the clock, January 1st, we had 220 employees in the organization. A few months later, we had 600. Today we have 900 Logorians. We will finish the year at nearly 1,500 Logorians. So we're going to go from 220 to 1,500 in 12 months. What are they doing? It can be boiled down to 3 main things. So first, we are investing aggressively in our go-to-market organization, which benefits Salesforce with over a 70% win rate at pilot. We need to have our reps be first to those accounts because we can close one of those accounts 3 out of 4 times. We're in over 50 markets and laws jurisdictional is very relationship driven. It's very hard to sell to a customer or prospect in Tokyo or Toronto from a seat in We need real people across our 17 global offices. Second, we are an AI company that sells software, and a lot of the change that we're driving— yes, it's software, but it's also change management. So as a result, we actually have more lawyers inside of the organization than software engineers. We call them legal engineers.

Wow.

Their job is not to do legal work. It's to redesign how a firm's work gets done given what technology can now do for that organization. That's a role that didn't exist 3 years ago. And now we're hiring hundreds of legal engineers this year alone. And then finally, and this one is near and dear to my heart, the third bucket refers to the people that are required to scale an organization to those 50 markets. Finance, security, compliance, people, operations. My own function is a really good example of a job AI created. Not destroyed. Overall, Legora sees AI as a capacity expander. In contrast to what you might think, our clients are actually adding more lawyers because they can now take on more work. Our $1 trillion TAM that we operate in, the legal services TAM, is supply constrained. We're seeing more M&A, more litigation, more regulation, more companies being formed. And our 900 Legorians, we're just getting started. The pie is growing.

Okay. Legorians, that's going to stick, David. That one's definitely going to stick. Salesforce is bringing to you that deep, trusted data, but it's actually not just the data, it's the context behind that data. While you didn't mention it, I imagine there's an element of probabilistic deterministic, but it's also bringing this new way of working, these new roles, like you said, like your own department being an entirely new function that wouldn't have existed before. There seems to be a common theme that we're starting to see. where it's less about the old traditional thoughts around AI and what AI will take away or what AI will reduce, and rather build and grow for the future. And so what, what are you thinking about for the future?

I'm thinking about how the surface area is changing. So when I look at Legora and Salesforce, everyone is logged into Salesforce, right? But if you're thinking about our Legorians logging into Salesforce in the traditional interface, The answer, it's changing every single day and it's split by role. The surface area has expanded and it's changed. And as a result, yes, ops and finance absolutely live in the traditional dashboard.

Right.

But reps likely are not logging into Salesforce in the traditional way. They're logging in via Slack, they're engaging via Slack.

Right.

And so the surface area is constantly expanding in this new AI ecosystem. Updates go into Salesforce via natural language now instead of typing. or pipeline reviews are coming directly from the system rather than building slides from scratch. To me, that's wasted mental calories. So AI is handling the busywork. Great. That's AI's job, right? But the human layer, that's adding the judgment, that's the sales strategy on top. And that is the future of Salesforce and Legora.

David, keep this up. We're going to have to bring you into marketing because you— what you are describing so well is this whole idea of headless is like you're putting it into real words to say you're— we're decoupling the Salesforce platform, from the UI, so that organizations like yours can build the way you'd like to build, what you want, build how you want to. David, I really want to thank you. That's exactly what yours is, exactly what an enterprise, a Gentiq enterprise looks like. And you've been building AI native from the ground up. Thank you.

Let's go. Thanks for having me.

Let's go. Now, for our final guest, we've got Replit. Now, Replit lets anyone build software simply by describing what they want. Can I try this on my son? No, I can't. It doesn't work that way. There's no coding required. The company is scaling fast with an ambitious goal of reaching $1 billion in ARR. Again, a common theme that I'm hearing from the last conversation and this one. Now, ARR, for the folks that are on the call, that means annual recurring revenue. And they're aiming to do this by the end of this year. Joining us is Replit's Chief Revenue Officer, Ghazi Masood. Ghazi, welcome. It is wonderful to have you here.

Thank you.

Hey, good to see you. How are you?

Great, great, great. Nice to see you too. Okay, so Replit's growth has been a little bit wild. Your enterprise adoption took off so fast that your old go-to-market systems couldn't keep pace. That speed, it broke a few things. What broke first and why was Salesforce the fix?

Yeah, yeah, it wasn't really what broke first. It was really the scale element, right? So when I got here and knowing how fast we had to go scale and build our GTM teams, it was just a common sense type conversation. In order to kind of scale at the pace that we wanted to actually scale at, we needed a hard, robust system. So the question was, yeah, could we have built it ourselves? Sure, we could have. Is that the best use of our time or energy to do it? Probably not.

Yeah.

And, you know, I've been a go-to-market leader for many, many years at different companies out there, and I'm a big fan of Salesforce. And the idea of getting a trusted, proven system of record into our ecosystem was the right, easy decision for us. So that's what we did. And, you know, not only Not only did it come on board early, it's really the foundation that's going to stay with us as we continue scaling to the levels we're scaling at over the next few years.

I love to hear that. And I love that you've actually called out that you had that contemplation, or maybe not really, to say, should we build it ourselves? I hear this day in, day out from a lot of folks to say, build it myself? Like, I've got a core business. I've got a core business of doing something. I should probably be focusing my efforts in that core business. And let alone, like, when you build it out, it's day 2. What's the day 2 implications of building something out, the maintenance of it? So thanks for calling that out. You grew Replit from 75 to 300 agent force seats, so 75 to 300 agent force seats speed, while you also increased your Slack consumption. What gave you that confidence that you could expand that quickly?

Yes, a couple of points. Like, so one, internally, if you look at our global employee base, we're all Slack users. So not in addition, in addition to growing our GTM team, we're growing as a company internationally, globally around the world. So we just needed more Slack licenses to help accommodate just our broad employee growth, which is great. You know, we're also starting on the whole Slack Code partnership with you all, in early stages with that. So the team's pretty excited about that and what's to come with that. So the Slack usage and growth was just common sense, just to scale with our employee base. We live in Slack every single day. Then on the Agentforce side of the house, you know, I had to go build a sales organization north of 200 people by the end of this year. And looking at global expansion into regions like South America, Europe, Middle East, Asia Pacific. So we needed—

Wow.

To go out and really expand our footprint of users on Salesforce. And by the way, it's not just the salespeople, right? We also have a BDR function that does both inbound and outbound. So while we use like our classic outbound dialing capabilities and tools, a lot of our systems of record all kind of pull off of Salesforce. So it's extremely important for us just to kind of make sure we have the right capacity of licenses for the folks that we're bringing on board.

That's great. And you're able to adjust and flex as needed with the— as your needs are coming along. Now, you're also running Salesforce headlessly. And we talked again with our earlier guest, David, on this notion of headless. You've got your own pipeline dashboard that you've built on Salesforce data, and it's feeding directly into Slack. Now, what does that approach make possible for your teams?

Yeah, so the philosophy is we have no shortage of Replit apps that we're building on top of. of Salesforce. And they're everything from like, we have our own CPQ, which is our order form generator, which basically pulls outta Salesforce.

Mm-hmm.

We have our own revenue dashboard that we actually run our business off of. We have our own forecasting application that our reps actually forecast their deals off of Salesforce. We also have our customer health dashboard that we actually call, which is like a Gainsight equivalent, right? That we actually go out and build on top of that. So we have no shortage of it. And our philosophy is, We, while Salesforce is important as our system of record, and that's where all the data harnesses, we want to go out and build the UI layer utilizing Replit because we're familiar with our processes, we're familiar with the look and feel, we're familiar with kind of our culture, and we're building those across the team. So in my RevOps organization, we actually have 2 functions. The classic name is go-to-market engineers, but we call them like AI engineers, and they're basically tasked with with all of the tooling that we want to go do that pulls off of Salesforce. So while Salesforce is in the background, the idea is no one's really touching it. It's kind of, you know, they're hosting the data, integrating with their data warehouse, and then everything on top of that is a Replit application.

So it's again, a common theme that I'm seeing come through in all of these conversations is Salesforce, use it as you wish. It's composable parts. We work well in the ecosystem with other technologies that you have or coding tools. And the other thing that you just talked about is a new role. It sounds like AI engineers, this idea that you've got new types of roles coming up into your organization. Maybe talk a bit about that.

Yeah, absolutely. I mean, you need somebody who is really good at dogfooding, who's really good at, you know, having the build versus buy type, common sense type conversation, and really coming up with a catalog list of libraries, right? You know, what the crazy thing is, like, everybody internally at Replit dogfooded And what that creates sometimes is, you know, duplicative applications and stuff that, you know, is happening out there. So putting that into a catalog and making sure that we actually have a source of truth of all of the value-added applications and what we can automate and what actually makes sense is like extremely important to us. So that's really what this role, you know, allows us to go out and do. And these people that we're bringing on board, I mean, they're just passionate about building with AI and tooling like this. So this is kind of what they live and breathe every single day. And in the early innings, but we've seen just tremendous growth tremendous productivity in a very, very short amount of time.

And is that building that they're doing, is it happening? Like, again, I'm having these conversations with customers where it was, this was traditionally just functions for technology roles and technology departments, whereas now this notion of citizen development where we want everyone to be a builder, is that a bit of what's happening within your organization?

Absolutely. I mean, so everybody irrespective of the roles, builders, right? So if you look at myself and my team, we're all non-developers and non-engineers, we're all builders. You look at people in the finance organization, they're builders. You'll get people in our, people in HR function, They're builders, right? So everybody is building it because the tools and capability of applications like Replit and others really allow people with like zero coding experience and knowledge to go out and bring their ideas to life in a very, very short amount of time.

And being able to use natural language to really just be able to describe very much so in line with your organization.

Yeah.

So what's next? What's next for you?

Well, next is to get to $1 billion. That's a big priority with a lot of hiring to go do. We're opening up some big international markets. You're going to see us make a big presence out in Europe during Liberty Week. That's been out on social. So if you're out in London, come see us. We're going to be making a big splash in London. And then after that is Japan and then Latin America. So we have a lot of growth to go out and do, a lot of hiring and big aspirations. And keep an eye out on our launches because we ship at just amazing velocity and speed. And you're going to continue seeing some amazing announcements from us as we close out the year.

I am looking forward to following you for many reasons. One is just all of the incredible development that you're doing. I think it's very applicable even to what we are doing. But I hope you're okay with time zones with all of that, all that jumping back and forth.

I am, yeah. Jet lag to me unfortunately does not bother me these days. So we'll see how it goes.

You just adjust. Well, Ghazi, I want to thank you for taking the time to join us today. We really, really appreciate that. And good luck with your It's very exciting.

Thank you.

Thanks for having me on.

Thank you. And there you have it, folks. Xero, Lagora, Replit. 3 very different companies that use different AI models in different ways. But here's the one thing that they have in common. None of them are throwing out their enterprise software. Not one. No, they're using it more. They're building on it, and they are unlocking value like never before. SaaS-pocalypse? I don't think so. It's more like SaaS party time. Am I right? Am I right? Well, that's all the time we have for this bonus episode. Thanks for joining us. We appreciate you. I appreciate you. Be sure to tune in next month when we'll be coming to you live from Dreamforce for all the latest news, innovation, and some of the world's most innovative companies. most influential voices. We'll see you right here on This Month at Salesforce.