From product launch to recognized revenue with Zuora AI

Hi, everyone. Um, just pulling up my screen. I think that you should be able to see that now. Kevin, does it look good on your side? Awesome. [clears throat] All right. Hi, everybody. Thank you for joining. Um, my name is Shalen. I am coming at you live from our headquarters here in Foster City, and I’m joined by Kevin Souare. Hey, everyone. I’m Kevin Souare. Uh, happy to be here with you all today. And, uh, today we are gonna take a practical look at how Zuora AI supports the work between launching a product and then recognizing the revenue that it generates for you. So basically, how a product makes you money. Um, the inspiration for this webinar is that old Schoolhouse Rock jam,

How a Bill Becomes a Law. Um, so we wanna show you, um, how a product moves through each stage of the quote-to-cash life cycle, uh, using Zuora AI. Um, so we are gonna jump in and get started, but before we do, um, just a precautionary measure, um, to share, um, I’m not gonna read it word for word, uh, but, but please just take a look. Uh, some of the things that you’re seeing in this webinar may be visionary, um, so please don’t make any purchasing decisions, um, based on what you see today. All right. Uh, let’s jump into housekeeping. So first, I would love if you could introduce yourselves in the, uh, in the chat tab. Um, say hello, let uni- let us know where you’re joining from. Um, next, I wanna point your attention to the links in the Resources section.

You can click More at the bottom of your Zoom browser. Um, third, please go ahead and ask us questions. There’s a Q&A tab for that, um, that’s also located under the More tab. Um, we will be recording this session. We’ll share a follow-up email with the recording, um, so don’t wor- don’t feel like you have to scribble down notes really quickly. Um, you’ll be able to revisit this whenever you’d like. And we are gonna have several poll questions sprinkled throughout, so, uh, please participate. Um, this is really valuable for us to hear from, uh, folks like you. All right. Uh, let’s jump into official intros. So like I mentioned, my name is Shalen. I am on the product marketing team, um, and I’m joined here by Kevin. Kevin, I’ll let you give a little intro. Yeah. Hey, everyone. I’m Kevin.

I’m from the Zuora product management team. I lead up our AI products and our platform products here at Zuora. Thanks, Kevin. So Kevin’s gonna take us through the demo. You’re gonna have to hear a little bit from me first before we can get into the fun stuff, um, so apologies in advance for that. [laughs] Um, but before we truly jump into the content, I just wanted to take a moment to celebrate, um, that Zuora is proud to be recognized again as a leader in the Gartner Magic Quadrant for recurring billing applications and placed highest in our ability to execute. Um, in my opinion, this really reflects what Zuora’s platform was built for, handling usage at scale, flexible pricing models, and clean financial closes. So if you’d like to read the report, um, there’s a link below. It will also be shared in the Resources tab if you’d like to download it.

And [clears throat] in fact, we are, we continue to be a leader in all major analyst reports. I did wanna highlight that on this Gartner report we achieved the highest score of all vendors for revenue recognition, which is connected to the story that we’re gonna tell today. So, um, just wanted to give a little plug. All right. [clears throat] We’re gonna move into our first poll. Um, so please share how frequently are you using AI in your daily quote-to-cash operations. Um, all right. I, um, can’t see the results. [laughs]

Kevin, I don’t know if you can. Um, but, uh, how do I… I can’t move on from this. Okay, I’m just gonna hit Exit. [laughs] Um, I’m not sure if anyone was able to see the results. I’m not. [laughs] Um, so apologies for that, um, but we will keep it moving. Um, so let’s talk about AI in finance. So it’s created real productivity gains in other functions, but finance really hasn’t had that same moment yet. Um, there really hasn’t been that, like, 10X breakthrough, um, “Oh my gosh, I’m way more productive now

as a finance operator.” But that’s because finance teams need AI that understands their work, the context, and the controls of the world that they’re dealing in in quote-to-cash, um, which is exactly how we’re building Zuora AI, with purpose-built agents and specialized skills built specifically for the people who run quote-to-cash. Now, across that quote-to-cash life cycle, there are distinct jobs to be done. Because Zuora AI is embedded into the system of record, and we are the system where those jobs are done, um, it can work with shared context across contracts, invoices, payments, collections, and revenue while respecting the permissions and controls that you use, uh, already, um, with full visibili- visibility

into the downstream impact that any changes that AI proposes will create. So three ideas to keep in mind today as we move through our demo are expertise, speed, and trust. Essentially the right context, faster investigation and execution, and governed action. [clears throat] All right, um, now we know that finance teams are already carrying more complexity and volume without more time and capacity. Technology is moving really quickly these days. You’re probably asked by your CEO on a Monday that we need to get a product live by a Friday. [laughs] I know we are. Um, and th-finance has to carry a lot of that downstream impact, and we know that you aren’t looking for another place to ask questions. You need to reduce that manual work

between something that understanding something has happened, and we know what we need to do next, and we have the tools available to us to take that action. So the examples here show a pattern across rate plan changes, payments, revenue actions, billing operations, and invoice write-offs with real measurable time savings in some of these key workflows. Now, the exact workflow may vary, but the goal is consistent. You want to reduce your manual investigation, accelerate the time to your next decision, and keep your people in control before any actions run. All right. So today, to make this concrete, we are gonna use a fictional company called BarkBrain AI. It is a bark translation platform

with subscription tiers and consumption tra-charges. Their operating model includes a base subscription plus usage tied to treats, park visits, and squirrels identified. So this is a company that’s very dog-centric. Every product, price, usage event, invoice, payment, and revenue outcome has to stay connected. Um, so like I mentioned, Kevin is gonna follow this one BarkBrain AI customer story through seven stages: being priced and packaged, quoted and sold, metered, invoiced, paid, recognized, and measured. And as you watch, look for three things: how quickly the team can investigate, how operational evidence connects to financial impact, and where approval remains with the user. Now, Kevin, I’ll hand it over to you to take

us through the demo. I just realized I didn’t have this screen, um, up, so here are our seven, seven workflows. Apologies for that. Um, and now I’m gonna pass it over to Kevin. Yeah, awesome. Thanks, Shaylyn. Uh, hey everyone. Great to be with you all today. Yeah, let me go ahead and share my screen. I’ve got, um… We’ll, we’ll, we’ll take you through the whole Zuora product. Uh, we’ve essentially invested quite heavily, um, in, in our AI capabilities in Zuora, so, um, from, from the whole, uh, uh, life cycle there we just saw Shaylyn talk about, I’ll take you through each, each example. Now, just a quick note on how I do the demo. Um, let me pull it open. Uh, what I’m gonna do is, uh… I, I have pre-run

some of the, uh, some of… or, or actually many of the prompts in Zuora AI, uh, because I di- I didn’t wanna spend time as, as, as we’re here today kinda, uh, having the AI respond. Um, you know, as you, you all have probably used AI quite a bit. Um, cert-certain tasks, you know, it could be a multi-step task where the AI is going off. It’s, it’s spinning off agents to go plan things, uh, and come back, and so that can take a few minutes. So, um, in the interest of time, um, the way I’m gonna do this is I’ll show you some pre-run prompts. I’ll also run some live prompts as well, so you can kind of see the AI in action. Um, and I’ll, I’ll go through the whole story. So where I wanna start is right now I’m logged into Zuora. Um, let’s start off with imagining I work at BarkBrain

AI. Uh, maybe I’m on, uh, you know, the finance team, and I want to figure out… uh, first I want to start with the pricing and packaging work. Uh, so, uh, the first area I’m gonna go into i-in Zuora is over in the products area. I’m gonna go into the product catalog. And, um, what I see here is I’ve already got a product catalog. I’m running my business already. And, um, I’m gonna go ahead and open up the AI side panel. Uh, so the AI side panel is over here. And, uh, let me just give you a brief explanation of how this works. Uh, so the side panel is available on every single Zuora page. Uh, it’s aware of what page you’re on in Zuora, and, um, it’s able to perform,

uh, either read o-o-operations where it can summarize your data in Zuora, or it can do, um, what we call supervised operations, where it’ll actually go and update data in Zuora. Uh, but it’s always human in the loop, so every, e-every different action you’re doing in, in Zuora AI, um, always has you as the user approving it. We’ve got a whole robust set of, um, uh, governance that we’ve built around this. Uh, we’ve got it documented in our, in our knowledge center. But essentially, every action that the AI takes attributes it back to the user who’s logged in. So in this case, I’m logged in. Uh, so in our audit trails and so forth, we will tag, um, any updates to data, uh, to me as the user. It’ll tag it back to the conversations. We

have a watchtower feature that allows you to, to get full visibility and observability of what’s going on with AI. Uh, but let’s, let’s start with the story, the, the BarkBrain AI story. So, uh, what I did is just a couple of days ago, I came in here and I set up, uh, a, a prompt. Um, and essentially what I said is like, “Hey, you know, um, you know, BarkBrain AI wants to launch a new Park Pro offer, uh, to our customers. Uh, it’s essentially a working tier subscription. It’s got metering. Um, so we’ve come up with these very cute units of measure around like how many squirrel alerts, um, or park visits, um, uh, the, the dogs are, are getting, uh, who are, who are subscribers of this service.”

And essentially, the, the gist of this is, uh, finance wants to create a new option for this that has the fewest downstream surprises across the whole life cycle, so across quoting, billing, revenue, reporting. And, uh, so, uh, I’m telling AI this. I’m also telling it to, you know, review our existing product catalog. I have two options in mind for how I want to price and package this. And then I’m telling the AI to go ahead and, you know, structure the answer into four sections. What we currently sell today, uh, the, um, two ways that we’re considering packaging, and then the implications of those, and then come up with a recommendation for me, and then also come up with a recommendation or, or

what steps I need to implement these in Zuora. Um, so this is an illustration of, you know, this is, this is a pretty common task for anyone who’s, who’s doing agile monetization, uh, using Zuora. We’ve, we’ve always had our traditional interfaces and APIs available for you to make these kinds of operations, but now we’re layering in AI to take care of this for you. Um, now this is a pretty big prompt, um, that I put in here. We do have many options in Zuora AI, like you can upload files if you want. Um, you can upload markdown files. You can upload PDFs, uh, Excel files, PowerPoint presentations. Um, so if, if, you know, say, say you had a meeting, uh, with your team, uh, describing the different options for how you want to price

and package, you can load that right into the Zuora AI, and it’s able to read it and, and process it. Now, one of the features of Zuora AI we’ll see, um, as this is running is that we show our various function calls. Uh, so in this case, uh, it wrote a script, uh, to go pull some, some information. Um, we do have a scripting capability built into this, so it’s able to, on the fly, based on what you’re asking, uh, generate scripts and, and write mini programs, um, to help you. Uh, we always show the thinking process in Zuora AI. So one of the principles we have is, is just full transparency in terms of the, you know, what… how the AI is thinking, what it’s doing, uh, so that you can, uh, essentially audit all of this information. And then, uh,

it also has a multi-step planning. So th-this is the one it… This, this, this prompt probably took about five minutes or so to run, um, to run through all these different steps. Um, and then it came back ultimately with, uh, what I had asked, uh, so what we sell. Um, you know, it kind of broke it down in a ni-nice summary. Uh, what I thought was really cool about this one is, you know, I did ask it in the prompt to, um, analyze these two different options we were considering. So option one, uh, was a working subscription plus, uh, metering charges and then, um, a new recurring add-on. And so what it did is, like, from each of these different angles I wanted to analyze, you know, it looked at it from a quoting perspective, uh, you know, Zuora’s CPQ product. Like, h- what effect will this have?

Straightforward. Um, mediation and billing, um, it reuses a lot of the pipeline and pricing models we’re already accustomed to. Uh, for revenue recognition, um, it, it, it uses the same, um, way we recognize daily. Um, and so it’s a very clean ASC 606 allocation. And then reporting, the, the roll-up is simple. Now, option two, it comes back with, it says, “Hey, you know, quoting is going to be more complex. Um, the billing and mediation, it’s going to require all this additional configuration.” Uh, the revenue recognition, uh, which is, you know, this is where it kind of adds a lot of help to our users. It’s, it’s, it’s flagging that this is a source of downstream surprises. So one of the things

the finance team wants here is to not be surprised later on, uh, during revenue recognition. And so option two is being flagged, and then the reporting is, like, going to be a separate track. So it goes through all this analysis. This m-this may have taken teams, you know, hours, you know, maybe longer to figure all this stuff out. Uh, but just in a matter of minutes, it’s able to kind of analyze this, um, using all the various skills we’ve baked into Zuora AI, all the context that we, we load into it. Uh, and it comes right back and says, “Hey, option one is the finance-ready choice.” Um, so this is just an example of how it can help with, um, with, uh, pricing and packaging. Uh, I’m gonna run a live prompt just so you guys can see it in action. Um,

that was a lot of text the AI came back with, so let’s just say, uh, I want a quick summary of this. Um, and you can… Um, this is gonna give you a sense of kind of like what it looks like when you’re actually using the Zuora AI. Um, so AI’s figuring out kind of exactly which, uh, capabilities to connect to, uh, and then it just, you know, essentially comes back with a very brief summary, uh, of, of what I want. Um, you can copy these. You can share your chats in Zuora AI. Uh, and then we also have an engine running that recommends, uh, next step prompts as well. So, um, this is, uh, this is often, uh, quite used by our, our users. Okay, so now let’s say we’ve got this

pricing and packaging plan set up. Uh, it’s finance ready. You know, it’s like clean. It’s gonna work across all the systems. Um, now let’s, let’s show how this can be used in a, in, in a quoting scenario. So I’m gonna pop over to an account. Uh, we’ve got, uh, Fetch Fleet, uh, which is, which is one of our customers, and we want to essentially run a simulation, a quote simulation of this new plan, uh, for this, for this, uh, account. Now, what’s nice about this is I actually didn’t create, uh, that plan in the system yet. Um, I’m able to essentially simulate it as well if I wanted to. Um, I could also have the AI create the plan for me as well.

Uh, so I have another example prompt set up here, uh, where it says, “Hey, you know, this, this particular customer, um, wants this, uh, pr- Park Pro package, uh, can you, can you go ahead and essentially, um, simulate it for me? Like, what’s it gonna look like if we, if we gave this customer this plan, uh, with a few different assumptions like, you know, a certain discount, let’s assume is, is in place. Um, and then, you know, a, a very common quoting scenario is you wanna know, hey, what are the impacts on ARR, TCV, uh, what the billing schedule looks like, and so forth. Um, so I wrote a nice little prompt on this one, um, to, to have it, have it simulate it. Uh, and then I said, “Hey, you know,

at the end of it, please structure it, uh, i-in terms of these four different, um, categories.” Uh, the key things I would say being, um, item three, which is the financial impact. Like, you really wanna understand, is this a good deal, uh, for, for us to do. Um, so in this example, uh, notice that it, it, it, it, it, it… The AI… I submit it, the AI does a b- a, a, a variety of different function calls. Um, in this case, it’s, it’s calling our preview API behind the scenes. Again, we, we make all this available through our, our, uh, watch tower, so you can see the actual calls it’s making. Um, this one also goes through another, uh, multi-step planning process and does a lot of thinking. And then, um, ultimately, it comes back with, “Hey,

here’s, here’s a, here’s like a very crisp summary of the current subscription.” Now, you can– you could view this as well on the main page, but this is just a very fast summary of it. Um, what change we were proposing to make with the new plan, and then the financial impact. So you can see, hey, okay, like this is gonna add, you know, five hundred and twenty-nine dollars of ARR. Uh, it’s gonna have a, um, you know, TCV impact of two thousand one hundred and fifty-six dollars. And, um, and so yeah. So, so you can essentially use the AI to, um, uh, to, to simulate. Um, now, what we’re looking at here, this is inside of the Zuora, uh, billing

product. We are also in the process right now of adding this into our, uh, Zuora CPQ, um, product. So coming this fall, um, natively within the CPQ, uh, sales teams and, uh, will be able to access, uh, this in their quoting processes, having all this kinda same power of, of all the data that sits with inside of Zuora. I’m gonna run like another, um, follow-up prompt here just to make it interesting. Um, what’s nice too actually is that you can download, uh, like an Excel file, right? If, if I wanted to see the simulation, um, you know, as an Excel, uh, I can, I can see it there. And I’m just gonna run, say, “Hey, in three bullets explain,

um, the quote impact, the key assumptions, and the approval b-boundary.” And so it’s gonna go ahead and, um, think about that for a second. The AI is always, you know, is always, you know, figuring out… We, uh, under the hood, we’ve developed what we call agents, uh, for different capabilities with, with inside of Zuora. So there’s like a billing operations agent, there’s a, um, a quoting agent, there’s like a revenue analyst agent. Um, they’re kinda perso– like they’re oriented to different roles, um, that we see in, in our customers. Um, and so that first step you saw is

like the AI’s figuring out which of the different agents, uh, needs to help with this. Yeah, and you can see it, it came back with, um, with that summary. Okay, so now we’ve, we’ve priced and packaged with AI. We have done a quote. Um, the next step is, uh, let’s say, let’s say the customer, you know, signs, signs this order form. They’re now subscribed to it. Um, now, like this is, this is like, um, there’s a lot of usage and consumption to, to what we’ve just quoted, right? All those different squirrel alert events, those are all feeding into Zuora, into our mediation engine. And so now we need to, um, make sure that we’re

metering and billing, um, you know, with, with accuracy. Let’s imagine one of our customers, um, came to us, very honest customer, they came to us and said, “Hey, uh, we noticed that, uh, from our engineering team, the events we’re sending into your metering, into the billing, uh, metering engine, um, doesn’t really match the, what we see in the invoice from Zuora. So can you help me troubleshoot this?” So, so this is kind of a billing operations, um, task. Uh, and let’s go ahead and go over here, uh, to the next area. So Zuora supports, um, we ha- we have a full mediation product. Uh, very high scale. It can ingest, um, high volumes of, of events into Zuora, raw events. So think of these as like, you know, if you’re doing AI monetization,

you’re pricing and packaging based on tokens consumed. You can feed all that data into Zuora, and then Zuora will, um, create these things called, um, meters, which essentially can transform and aggregate all of that raw event data in real-time into usage records that are then rated and billed. Um, so in this example, the customer’s calling, um, and they’re saying, “Hey, um, I put it into a prompt, uh, just to make it easier to investigate this.” Uh, they’re saying in their June invoice, it looked light. It l- it was like lower than they expected, um, because their engineering team is saying, “Hey, they, they fed in, you know, ninety-one hundred events, um, uh, but the invoice only

shows like sixty-two hundred.” So in cases like this, um, this is another use of the AI. It’s kind of helping, um, puzzle together, uh, uh- I-if there’s a billing dispute or a billing issue, uh, how can I go in and quickly troubleshoot this, right? There’s, there’s, you know, potentially several different things that may have gone wrong in this in terms of like maybe how the subscription was set up or, um, how the bills were run, or it could be an issue in the mediation engine. What’s nice about the Zuora AI is that it has this ability to kind of look comprehensively across all of, all of the Zuora systems and help you troubleshoot this. So I, I put in a, I put in a prompt here, um, and, um,

you know, I said, uh, what we expected versus what was billed. Like we expected it to bill off of ninety-one hundred events for these Squirrel alerts, but it only billed sixty-two hundred. Can you analyze the mediation engine, uh, which we’re looking at here in the background, um, you know, the billing impact and the, and the, uh, remediation pa-plan? Like how do we actually fix this? So in this example, um, notice the AI is, is running a lot of different, uh, looks at the system. It’s looking at the different meters, the history. We have a full audit trail within the meters, and it can see, you know, different, different information. It might write a script, um, to, to analyze this. Um, maybe this is something that in, in the past you would have done as

like an Excel, uh, model, um, where you’re pulling a lot of data into Excel and trying to do this analysis. The Zuora AI will just write a script for you, um, and help figure this out. Again, it goes through a planning process, and then ultimately, um, it comes back with its findings. And, uh, so it, um, it kind of re-summarizes the expected versus billed, and it looks at the, um, mediation pipeline. It concludes, hey, the mediation is running, and you can see it is running over here. This, uh, this, um, ID number five, uh, meter i-is ac-active. Uh, there was actually no errors. Um, and, uh, you know, it, it goes over what was flagged, and then it also goes through like how you would fix

this and what, what is the safest way, um, uh, to, to, to do this. And it’s, it’s actually saying, “Hey, go ahead and do a debit memo.” Um, so, you know, issue a debit memo in this case for a certain amount, and then you could actually ask the AI to go ahead and do that for you. Um, so this is, this is an example of how you could use the AI, um, when you’re in the process of billing, uh, if there’s some dispute or some issue, um, how it can, how it can streamline that process for you. Uh, okay. So, uh, another, another example is in collections. Uh, so, you know, let’s say you’ve been billing, uh, your customers and so forth, and

you’re trying to, um, you know, essentially look at your collections. Um, and, and so, you know, may-maybe you’re in the payment section of Zuora, and, uh, you might have another thing where, you know, you are trying to figure out which collections need priority. You’re trying to put together a recovery plan, um, by business impact. You want to focus on the highest priority customers, uh, to go out and collect. And so, um, you can go ahead and ask, uh, AI in this case to help you put that, put that plan together. Um, in this example, I’m saying, “Hey, please structure it in terms of a collections health snapshot, um, of overdue receivables,

um, open balances. Please go ahead and rank it, and then give me the recommended actions, uh, for me to go and plan my, um, collection, collections activity.” Again, it goes, it, it goes through the thinking process, and then it, it, it actually comes back with, um, a full picture of the collection’s health. Um, now there… You can do more sophisticated, uh, versions of this. Um, I’ve seen customers do churn an-analysis, where based on all the data that’s in Zuora, um, all the events and so forth, um, you can start to say, “Hey, you know, which of my customers are the highest churn risk?” Um, but here it’s, it’s listing out the, the key accounts, um, and then it’s ranking them, right? So the Squirrel Watch H-HQ customer,

um, is, is like the number one, um, collection risk. Uh, and then it’s just stack ranking them and giving you the information, um, to, to go and, and, and optimize those collections. Okay. Um, so n- so now it’s like, okay, we’ve been, uh, we priced and packaged. We did some quoting and simulations. We’ve, um, been issuing invoices and, and managing, uh, that process. We’ve, we’ve now been collecting and getting cash, um, and helping… A-AI’s been helping in this whole process. Now comes the time where we need to recognize revenue. Um, so let’s go ahead and take a look at a scenario with, um, recognized revenue. I’m gonna go over into the revenue product, and, um,

I’ve got a revenue contract here. And, you know, essentially what I’m, what I’m, what I’m looking to do here is let’s imagine you made a, a midterm amendment, um, to one of your subscriptions, right? And so you’re, you know, you’re trying to analyze, okay, well, what is the effect, uh, of that on revenue? Um, you know, and some of the, some of these subscriptions get quite comp-complex with the consumption, the recurring, uh, charges, the, you know, different, uh, usage flowing in. And so, um, one of the things AI can help with is, is assessing these kinds of contract changes, um, uh, that may-maybe happen midterm or maybe happen, um, you know, as exception processes. And so I’ve got a quote here, uh,

uh, a prompt here where, um, for the Paw Prints Labs customer, uh, uh, they, they did a midterm change- And, um, yeah, the revenue team needs to understand, uh, the allocation of the change, the close readiness, and audit support for this particular accounting period. Um, and so we’ve structured it in terms of, A, what was the change? Um, can you do a revenue trace, um, across, you know, the allocation, the recognition schedule, the expected impact? Uh, you know, I want a, um, close and audit risk analysis, and then I want, um, the evidence I need to retain, uh, for my close support package.

Uh, so again, AI goes through this, um, gathers all the information it needs, and then it comes back, uh, with, with what I asked for, um, in this case. Um, so it, it really just breaks it down in terms of, um, you know, the contract-level totals. Uh, we have, you know, full version controls on, on the revenue contract, so you can kind of see version by version what the allocation changes were, um, before and after the amendment. Um, and it does, it does, uh, just a very nice breakdown of this. Um, this is also something where, you know, if you do download the Excel… I haven’t downloaded an Excel, but let’s just take a look at what that could look like. Hopefully, y’all can see my screen.

Let’s bring this Excel up. Just resize this a little bit. Yeah, and you can see it, it does a nice… It, it creates this nice file that I can retain, uh, with, with kind of a breakdown of all the different modification events, um, the allocations, revenue allocations by version, the revenue contract lines, um, and the, um, you know, bills by invoice. Okay. Um, so hopefully that gives you a sense of using, uh, AI across, you know, across the whole Zuora life cycle. There, there’s a lot to Zuora. We-we’ve got a lot of capabilities, um, that our customers take advantage of.

And really what, what we’re doing with AI is, is just really helping streamline, streamline the work. Some of these processes I’m showing you can take, you know, can take, you know, significant portions of, uh, of someone’s day, um, to, to, to do. And, and we’re seeing customers getting real acceleration, um, free-freeing up their time to work on, um, you know, more strategic things, um, oth-other tasks. Now, the last example of the pre-prompts that I have, and then, then we’ll look at a few other things, and then I’ll hand it back to Shailen, is, um, analytics. So we’ve got, you know, very nice dashboarding and charts within Zuora, all the key subscription metrics that you want to track. Um, and let’s say I was asked by the leadership team to produce

some KPIs, um, in plain language about how Barkbrain AI is doing this quarter. Um, so this is a very nice thing where you can do a prompt, um, kind of asking for what you want. Um, maybe I dictated this one. We do have this nice little dictation feature where you can speak to the AI. It’ll listen to you. It’ll kind of, um, transcribe what you’ve said and, and put it in here, so it could be a stream of thought. Um, it goes out, it runs all the different queries and so forth, and then it comes back, um, with the analysis. Uh, you can pull down Excel files. It can create HTML files, um, if you want dashboards. Um, and then I’m gonna pop over… I haven’t shown you guys this yet, but I’m gonna pop over into full screen mode. So what we’ve been looking at in this

whole demo is our side panel. So again, the side panel has context on the, um, everything you see on the Zuora pages, and it’s got all the same capabilities. But here on, on full screen mode, I have a more fully immersive, um, Zuora AI experience. Um, and so like maybe I s- you know, after I had asked it to create these analyses, um, uh, maybe I said, “Hey, can you create a dashboard visualization of this conversation?” Right? And so then it goes off, and it builds an actual dashboard that you can download, um, showing, showing these key metrics, um, that it’s come back with. Uh, so, uh, just another one of the capabilities, um, that we have is, is like native charting,

uh, where it can actually build charts, uh, and dashboards for you, um, from the AI. And these are all, again, uh, fully shareable. So if you wanted to, you know, share this, this conversation with one of your colleagues, um, there is, there is an option here where you can share a chat, uh, share a chat to any other user who has Zuora, um, Zuora access. Okay. Um, so those are, those are all the, um, kind of, uh, life cycle, uh, prompts that I wanted to show, show you all, um, for this, uh, uh, company Barkbrain AI. Let me show you a couple just other things that we have, um, in the product. So one of the things that we have are y- Notice like at the top of this page, there’s a Save to runbook option. What we’re increasingly seeing is that, hey, as

you’re using AI, there may be things that you want to repeat and do over and over, like maybe this analysis, um, or maybe the quote simulation that we did earlier. Um, you can actually click this button and save to a runbook, and then there’s this runbook option where you can see, um, uh, essentially saved prompts. So, uh, this is a saved runbook from that mediation analysis we did, uh, where, um, based on the conversation, Zuora AI is able to create a prompt for you, and you can just rerun this. Um, the idea is you could just click this button, and it would, would go ahead and spin off a new conversation, um, for whatever next investigation you want to do. The other thing we have are what we call controls. Um,

so, uh, controls there, there’s a couple different types of controls, but these are essentially guardrails that you can put in, uh, to control what happens with Zuora AI. Um, so I had– I, I’m gonna show you guys an example of this. Um, there’s a, um, invoice, uh, that is this, uh, invoice four oh eight where I’m saying, “Hey, um, when Zora AI is asked to post, apply, correct or, you know, make some kind of an update during this webinar, um, on any of the following records, uh, block it, and then put in this, this quote, ‘Hey, this action is blocked by the BB demo,’

uh, block hero, uh, hero record rights.” So what these prompts are, you can create as many of these as you want. You can control exactly what the AI does using natural language, and it’ll control how all the end users, um, interact with, with AI. So if I start a new chat, uh, from full screen, and let me just, uh, copy and paste in the prompt. Um, so I’m gonna pick on that invoice ending in four oh eight because this is the one that the control has… should block. So if I run this, we should get that block message. Um, and this block message is coming from the control that I had set up as an admin to prevent the AI from taking any action on, on these invoices.

Now, while this is running, this can be generalized to any type of object or any type of action. You can have it… You can, uh, limit AI if, if over a certain dollar value of, of types of updates, you can say, “Hey, AI just doesn’t…” Even in supervised mode, um, I don’t want the Zora AI to make an, an update if it, if it’s an invoice over, you know, five thousand dollars, let’s say. Um, so there’s, there’s like a tremendous amount of, um, control that we give you, uh, to govern how the AI behaves. And you can see here, um, this one blocked it. Last thing I’ll show you guys, uh, on the demo is we have supervised mode. So, um, uh, anytime a user… I-if you turn on supervised mode, so it’s optional.

Uh, some of our customers run the AI only in a read-only fashion. But if you turn on AI supervised mode, uh, which is a setting that you can configure, then, um, anytime someone goes and tries to make a write update in Zora, like if you wanna, say, write off an invoice, um, the AI will ask the user to confirm, uh, what it’s about to do and that they want the AI to proceed. Um, so this is a safeguard that we have, uh, with Zora AI. So I’m gonna, I’m just gonna put in another prompt here where I say, “Apply a ten dollar unapplied credit memo to a certain invoice.” So this, this would be a, an update action. And so what the AI should do in this case is explain to me what it’s about to do and, and ask me to confirm

that I wanna make this, um, make this update. So, so notice it, it says, “Yeah, I can go ahead and do this,” and I have to say, um, “I approve this action.” Uh, so I click that, and then it’ll go ahead and, um, start, start, start the work on, on actually, um, applying that credit memo. Um, so I… Looks like I ran into something there. Um… Yeah, let me see. Uh, maybe I have the wrong invoice number.

Okay, I, I won’t, I won’t futz with this, but, um, you know, you can, you can have a conversation with the AI to, to get it to do the right thing. Um, okay. So with that, uh, I think that’s everything I wanted to show on the demo. Uh, so I will, uh, stop sharing, and we can go back to the slides. Cool. Thanks, Kevin. All righty. Um, so what you just saw was Kevin following one customer story through seven stages of quote-to-cash, but the value really that came from, from this demonstration was what connected those stages together. So there’s one platform with shared context and governed action. So you saw him investigate without rebuilding the story screen by screen. You saw him connecting operational

evidence to real financial impact and moving from insight to action while people retain approval and the work remains auditable. And that’s the larger promise of Zora AI. It’s going to reduce your manual work, accelerate decisions, and create more capacity for strategic finance work that I know you all wanna be doing. So we are gonna move into another poll. Um, so which of the use cases that we demonstrated would be more… most impactful for your daily work? I’m gonna stop sharing so that we can see the results.

Hey, Shaylin. It’s looking like a lot of people are getting back to us on saying that packaging catalog design for a new offer, um, usage metering and investigations, um, but it’s pretty much equaling out amongst all the options that we have, um, outside of packaging and catalog design for the new offer being the favorite as of now. Cool. Thank you for sharing. Um, that’s very interesting. Um, I think shows some of the pain that y’all might be feeling [chuckles] in that space. All right. Um, now I did just wanna give, uh, a little bit of a tease into what’s coming for Zora AI. So today is part of that broader direction. Um, for

the second half of the year, we are gonna be anchoring on these five themes that you see here on the screen. So firstly, what I like to say is an army of agents with more agents, more skills, more right actions, and more complete workflows across quote-to-cash. I essentially want you to think, you know, for anything quote-to-cash, any job to be done across that life cycle, there’s an agent for that. Second, trust, security, and control with more granular permissions, approvals, audit trails, and visibility into AI usage across your teams. Third, personalization and accuracy, so AI that understands your organization, your terminology, your preferences, permissions, and context. Fourth, proactive and always-on experiences, so AI that is not

just waiting for the prompt that you saw Kevin, um, Kevin, you know, enter so many times. Um, the AI will proactively surface what matters to you, monitor exceptions, and highlight anything that needs your attention, and handle routine work on a schedule for you. So it’s a little bit more hands-off for you. And fifth, we are gonna be working on agent-to-agent interaction, so connecting external agents to Zuora agents through our MCP, so your work can move across systems without forcing you to switch between tools constantly. So I believe we’re gonna move into a, um, another poll. So this one is, um, more about future

thinking. So what use cases would be most impactful for your daily finance work? So think about in the future, dream big. Um, we really wanna hear from you about what we should build next here. I’m gonna stop my share. And Sabrina will give us an update as the results are rolling in. It’s looking like fully autonomous bill run- Mm-hmm… and invoice QA within autofix for lowest risk expectations. Um, a few people have said something else. Um, please feel free to tell us in the chat. We definitely want to hear from you guys on

what you’re dealing with so we can keep iterating. Um, but the one I mentioned, Sha- Shailyn, is by far the favorite as of now. Proactive alerts before usage shortfalls, billing errors, or close risk hits are up there as well. Same with end-to-end quote-to-cash workflows that hand off across teams without switching tools. Well, Kevin, any of that surprising for you or…? Yeah, no, that– I mean, it’s great to hear. Uh, n-not too surprising. Um, but, uh, yeah, exciting. Cool. Um, and I think that we have just one last, uh, poll for you to fill out before you leave us today. Um, so before I close, I just wanna say thank you all for joining us. Um, we really enjoyed this conversation, and like I mentioned at the beginning, a recording will be available.

So please just let us know if you would like an account team to reach out to you to chat a little bit more about Zuora AI, um, and then you can be on your merry way. Thanks for joining us.