As buyer expectations rise and presales teams face increasing pressure to do more with less, demo automation is entering a new era. In this on-demand webinar, Saleo co-founders Justin McDonald and Daniel Hellerman explore how demos have evolved, from product tours and scripted walkthroughs to live, data-driven experiences, and why traditional approaches are no longer enough to scale modern sales motions.
The session introduces Saleo’s AI Demo Agent as the next evolution in demo delivery: an autonomous, multilingual agent designed to deliver always-on, adaptive product demos. Daniel Hellerman shows how the AI Demo Agent runs demos in a live product environment, supports true co-browsing, adapts in real time to buyer input, and captures rich engagement data for sales follow-up. Learn how agent-led demos help teams eliminate time-to-first demo, extend presales coverage, and preserve demo quality across the funnel.
Read the full transcript:
Hello everyone and welcome. I am Melissa Mandel, the VP of Marketing here at Saleo.
I’m really excited to have you join our webinar today, and really excited to introduce you to Justin McDonald and Daniel Hellerman, our co-founders, who will be leading the presentation today.
Thanks, Melissa. Hello.
Good afternoon all or good morning if you’re on the West Coast. Justin McDonald.
I’m the co-founder and CEO of Saleo. As Melissa said, I’m joined by Dan Hellerman, my co-founder.
And we were just joking earlier. We feel like we did these sessions all the time, years ago when we first launched the company.
So it’s exciting to get both of us back on a webinar to walk through our newest innovation and also what led up to it. The market forces that have been pushing and pulling demo automation. So Dan, excited to have you here.
Appreciate you joining. And yes, got a lot of fun stuff to cover.
All right. From an agenda perspective, we’ll give you the current state of demo automation.
And by the way, we promise not to PowerPoint you to death. I think everyone is joining today because they want to see a demo of Seleo’s new AI demo agent.
So we’ll reserve most of the time for that, as well as questions. We recognize the market, customers, prospects have questions on what it does, how it operates.
But we’ll also talk about the forces. Why does this make sense? What are the market forces? What are the day to day challenges that presales, sales, and software companies struggle with that got us to this point? And then you get to meet the agent. So we’ll talk through that.
And we’ve got several different scenarios, several different iterations of the agent, what it’s capable of doing. And then I believe we’ll have an extended Q and A towards the end.
So if you have a question right now all the way through, feel free to do that. Our marketing team is watching that closely.
So the evolution of demo automation, right? So it’s been around for technically fifteen years or so. So in 2010, there was a release of interactive videos, which is the ability to kind of jump on a video, choose your own adventure, experience products through video.
And then we got into the product tour space. There were, gosh, I don’t know, 10 to 15 different vendors that launched all at once in the 2020 era of giving just a different experience to interact with the product. On the website and through URLs, email.
So you had a slew of product vendors. Most of those were just screenshots at the time. So you didn’t get into the HTML, the DOM cloning until later 21-22.
Then there’s a couple of vendors that popped up doing that, Saleo being one of them. And then there was the evolution.
I think the biggest shift in the entire category, around the 21-22 era with the invention of data injection that Saleo created. The ability to actually inject live demo data into a live product to be able to allow SCs and AEs to demo the authentic native software product versus having to use a clone or a product tour.
And then there’s a “what next”? So it’s a fast moving category. It’s evolved quite a bit even in the past four to five years.
And so there’s just I think the thing I would highlight as well is there’s a reason all of these exist. There’s a need across the entire sales funnel and sales cycle for these products to exist.
So all of them are as relevant today as they’ve ever been. None of them obsolete.
I do think there’s legitimate use cases for these vendors, the different deployments of the product. And again, excited to talk about what’s next.
But Dan, curious of your thoughts on the market as well.
Yeah, it’s been an absolutely fascinating evolution, right? As we’ve seen vendors come in, vendors fall out of the space. To me, it’s the complete transformation of how demos and how platforms are demoed today versus even five years ago.
Where videos used to be accepted, buyers trust and their need to trust the solution is increasing. So they increasingly want their hands on product.
They want to see true live demos. And with the advent of AI, AI is part of the thing that I believe has sort of been the death of the classic screenshot based product tour market.
Because people want to interact. They want to see the true value that AI can generate, and they don’t want to just see a slapped in product.
And so as users have gotten burned with purchases more and more, we’ve seen this big move towards truly live demos, real use cases. And I think that the demo automation space is going to continue to iterate, push towards really authentic experiences as AI evolves.
Totally agree, Dan.
So you can imagine lots of consolidation in the categories evolving.
You can see some of these logos down below, some of them don’t exist anymore. Some of them may not be as relevant as they once were.
But it’s definitely evolving quickly. It’s been exciting to watch.
We’re grateful and proud to obviously be part of the category.
So today, I’m sure we have many customers, a ton of presales folks on today. But why did we even push towards an agent? I think this is the why.
Presales is being asked to do more with less. You’re asked to cover multiple AEs and regions without additional headcount.
I think there’s this natural evolution too where, with the advent of AI, there’s expectation that companies can be more efficient, that people are able to effectively do more with less. And so it’s on one hand creating this baseline of high expectations, but it’s also the solution.
I think, because of product tours in the demo automation space, there’s just more demo moments per deal. And there’s also more strain on the SE team to deliver, to be able to create these assets, these different experiences. There is pressure to be able to deliver the right demo, to be able to personalize it in a way that they can today that may not have been expected in years past.
But now that buyers know the technology exists and the competitors are showing a certain way, other companies need to level up and do the same. And then harbor tours, they can consume SEs from a timing perspective.
On one hand, the live demo is so incredibly important, that will never change. The human connection, the ability for an SE and AE to jump on a Zoom or be in person, it is the highest conversion type of event that you can do.
We fully believe in that. As you know with Saleo, Saleo Live is the underpinnings of our success.
We fully believe in the live product. But you can only do so many, right? And if you’re spreading yourself thin, if you’re doing 10 live demos in a day or in a couple of days, you’re maybe not putting as much effort or I shouldn’t say effort.
You’re not able to do five of them. You could do five better than 10 if you’re able to take your time and attention towards fewer.
So, I think, with the advent of AI, we’re going to allow you to put more focus on fewer deals and not spread yourself too thin.
The other aspect of that as well is that buyers. Buyer expectations, the whole advent of buyer enablement. They expect more. They want it more personalized. They want what they want, right now.
Obviously, we live in this TikTok culture, so this instant gratification flows over into the actual software purchasing space. They want access anytime.
They don’t want limits by calendar. They want proof often and early.
Again, there’s the expectation that you’re going to have more demo moments per deal than you’ve ever had before. It needs to be personalized even at the first touch.
We see that expectation constantly in market of just the elevation and expectation of what buyers want and what they expect has changed dramatically. And then all of that coalesces into buyers.
They’re going to keep expecting access. They’re going to keep expecting experiences that are tailored to them in real time.
They’re going to expect more interactions and more ways to interact with your product per deal or per sale cycle. And again, SEs, right, they can’t scale to match that demand because it’s just, unless you’re hiring more, it’s just very difficult to do that.
And then the bar continues to be high. It continues to have a requirement for consistency.
And as I said, AI, it’s a double edged sword, right? It’s creating the bar and the baseline for high expectations, but it’s also the very solution that can help meet that bar. So that leads us to the current state.
So this is more of a slimmed down approach of the sales cycle from top of the funnel on the left to bottom of the funnel, close sale on the right. But harbor tours can drain SE time.
It’s easy to lose deals before they start if you don’t have the right experience. And there’s always going to be a need for product tours. Don’t hear us say that.
There’s always going to be a need for live demos. They have their place. But there needs to be something more towards top of the funnel, middle of the funnel, that can relieve the pressure valve of SE time, of companies time, to be able to give buyers the type of experience that they want in a demo.
And again, I think coverage. We’re seeing that more and more.
Budgets are getting cut. SCs are being spread thin, asked to cover more AEs year in and year out.
And that just leads to very limited time and very cluttered calendars. Dan, any thoughts on this from a pure sales cycle perspective?
I think you nailed it. Everything that I hear across the market as I’ve been meeting with SEs are just, I need to do more with less.
The budget pressure and the optimization, they’re looking for this friendly, helpful coworker, right? Like how can I get help today? It’s not how can I onboard somebody and get more help in the future.
Yeah. And I think it’s really easy to get on a webinar like this, or vendors have these cliche terms. It’s a legitimate need.
It’s a legitimate comment that we hear day in and day out.
Every week I’m hearing a company saying, “I’m being asked to do more. I have less time. I’ve less resources.”
The bar has been set quite high from an FP&A, from a P&L perspective.
It doesn’t matter what type of company you’re at. So I think it’s a very, very real thing.
It’s not some cliche term that we pull off for this webinar. We hear it from our customers.
We hear it from our buyers day in and day out.
So meet the Saleo AI demo agent. It is a fully autonomous, multilingual agent that can deliver an end to end demo, fully conversational, 24/7.
It can be accessible in many different modes and mediums, which we’ll talk about. But it eliminates the time to first demo.
It keeps your buyers engaged in a way that they’re just not going to get from a product tour or from a video today. I think once you see it, you understand what that means.
It gives sales information and insights, and arms them with context in a way that they’re not going to get without the type of interaction and the data that the agent can gather. And it expands the capacity to, quite honestly, unlimited scale.
It operates 24/7. Our North Star for this product is to be able to deliver an agent that can guide a prospect through a demo end to end, just like a world class tenured sales engineer.
So our goal for that is to try to hit that particular threshold and bar. Again, this doesn’t eliminate the need for SEs.
SEs are not going anywhere. If anything, you’re even more important because if you’re a forward deploy engineer, you’re being asked to tune the agent, you’re being asked to create the best possible experience, SEs are going to be the ones doing it.
So I would say the demand for SEs is ever more present and ever more needed with this type of technology because somebody has to set it up. Somebody has to create it.
Someone has to tune it. Someone has to update it as your product evolves.
So it’s a very exciting time for the market and SEs in general. What we’re most excited about is that it’s not just an agent which has a video where you’re kind of in this mode and you’re along for the ride, it’s on rails, you have to kind of go where the agent goes.
It’s true co-browsing, because we own the live demo, because we own the demo data, we can actually create a container that allows you to interact with the product while the agent walks through the demo. Dan will actually show you what that looks like.
It has full context reasoning. It understands what’s on the screen because we control the demo data, what’s being asked, where you are in the product.
And so that type of experience, it’s next level. It is so impressive.
Again, when everyone can see it today, get your hands on it. I think you’ll understand what we’re talking about.
But it has progressive profiling as well because personalization is our DNA. The models that we’ve created, the actual agent, the way that we tune it, the way we set it up, the guardrails are set up, the entire configuration.
It actually asks discovery questions, can profile the buyers on what industry they’re from, what they care about, what role they’re in. And then it actually adjusts how it speaks to them based upon the information that it’s gathering and that it’s intaking.
And then it stays on track. But it also, and Dan, you’ve got a good scenario for this from a gaming perspective. It allows just enough leniency for the buyer to go where they want to go, but it slowly pulls them back on track.
What was that term you always used when we were developing the product?
Yeah. We call it a dynamic narrator.
So if you guys are big fans of choose your own adventure books or if you’ve played around with some of those Netflix shows where you get to choose what’s next, by hopping ahead. It’s that same idea.
So the narrator is going to adjust the demo flow, adjust the demo data, take you down different paths depending on what you’re most interested in. And the thing we wanted the most was not to have an agent that’s just trying to basically replaying a Google Slide deck.
We wanted a real agent, in real product. And we built it on the learnings that we have from Live and the data from Live to ensure that it’s a personalized end-to-end experience.
It is literally recreating the best demo experience a solution engineer could create And from our experience, it replicates in seconds what could take an SE forty, fifty hours of building complex demo data, especially if you’re a multi product company, if you’re an enterprise product company that has to hop between multiple complex products to tell a story. It’s just an absolute game changer.
Just looking at the attendees on the call, I know a number of you guys are already getting to sort of play with it and see that same experience. We’re very excited for the rest of the market to get to see that as well.
Yep. That’s great.
Thank you, man. Yeah, and I think it’s great to have an experience like that.
But then the data, the analytics, the insights that you’re going to get from the agent that can then feed downstream into your Martech stack, the type of intent data, the questions they’re asking, where they spend their time. Things that quite honestly, product tours do a good job of, right? But we have the benefit, because it’s fully conversational, you are literally interacting with it like you’re on a live demo with an SE.
The entire conversation, the output from the buyer is then able to be collected, the transcripts, the type of intent signals, things they care about, the questions they ask, are all going be pulled into the analytics and then surface to the SEs, to the AEs, to the go to market teams.
Yeah, to add to that, Justin, conversational for us, it’s a true voice agent. You can talk to it end-to-end. To us, it’s just as natural as talking to somebody on a normal Zoom call.
We want to replicate that experience end-to-end so that sales and buyers are used to going into the same flows that they’re used to for a sales cycle.
I think that, quite honestly just to open up a bit, we said we’d give you an inside look at some of the forces at play here. There’s a lot of chat prompting agents out there where you’re able to kind of interact and prompt and type it out.
That’s not what you’re going to see today. This is something that has not been released before.
It is a true conversation in the sense that it’s a voice agent that you’re talking, that you’re interacting with, just like Dan and I are today. All right.
And this is the future, right? This is the future state of what the sales cycle looks like.
It releases the pressure valve on the top mid funnel to be able to free up SEs to focus on more high value tasks, the deep dives, the technical validations, all the things that you would like to be doing on fewer accounts, on bigger deals and more qualified deals versus spending time on unqualified accounts. So harbor tourism may suck your time.
The agent allows you to provide the right experience to the buyer in a way that product tours are just not doing today. Again, expanding the coverage without the trade offs, giving you unlimited scale, and then interacting with them at peak intent. They don’t have to choose something on a calendar.
Again, SEs are limited by eight to ten hours in a day. I know some that work more than that because they’re asked to do just incredible amounts of demos day in and day out.
But it’s there to be able to support, to augment SEs. Not meant to replace.
It’s there to make their lives easier so that they can focus on the tasks that they want to do, not on the repetitive time suck components that are required today. It’s a necessary evil.
It’s obviously what pays the bills, but it can be very difficult and draining on SEs. And then the last thing I’ll say, I think we’ll be focused more from a demo aspect, top of the funnel to mid funnel.
But we’re seeing a ton of interest, a ton of use cases for different personas with our agent for post sales. So the sales enablement, the trainings, there’s a lot of utilizations of the agent post sales as well.
All right, last slide. We’re going to hand it over to Dan and we’ll jump into the product.
But what makes us different? Hopefully you’ve seen quite a bit of that so far as to what makes the agent different. It is the first of its kind.
So that in itself is different. But the reason why it’s so unique and will be unmatched is that because we’re the only vendor that controls the live demo data and the live product, you’re able to adjust the storytelling that then feeds into the agent.
It feeds into the LLM so that the agent can deliver a story personalization adaptation in a way that’s only unique to Saleo. And then AI, right? The LMs, they love data.
They love context. We feed information about the product, about the screens, about the page, about the demo data into the AI so that it’s the most lifelike.
It is the most realistic, powerful agent that could be delivered. And then it’s all on a single platform.
Whether you need a live demo, whether you need product tours, whether you need sandboxes. And then obviously the agent on a single platform.
So the SEs have one platform, fully unified, single code base at their disposal.
So that is it for my portion. We’ll hand it to Dan and let’s jump into the demo.
Awesome. Super excited to give you all a preview of our demo agent. I’m going to pull it up in just one second here.
So just as a quick frame of reference, we’re going to go through three examples of us interacting with the agent. Like I said, I know a number of you all have already gotten a chance to interact with the agent.
The rest of you will get chances over the next little bit as we get through our market launch. For us, this very first demo, we’re going to show you three back to back demos.
And each one will pause after, do a little bit of reframing, talk about a little bit of your scene. This very first demo, every one of these is going to be inside of a product.
It’s going be inside of Salesforce’s data cloud, which we’re really excited. I want to talk about two important things.
One is that this type of demo can be enabled in under thirty minutes. For us, self-serve is super important.
As we evaluate the market, we’ve seen vendors out there that go into a black box mode. They say, hey, I’ll roll out an agent for you in four to six weeks, and you have very little control.
This is completely self-serve. You as a customer will be able to create these demos, turn them on without any input at all from Saleo, and also be able to continue to add to them, continue to iterate with them, continue to adjust the data.
And so this demo is going to cover a few really key features. We have a ton in development that we’ll be excited to share at a future date.
You’re going to see light personalization. It’s going to adjust for my use cases.
You’re going to see a cursor moving around to help navigate people and show them where to click next. You’ll see co-browsing.
So you’ll see me scrolling up and down the page and interacting with it while the agent can do the same and is aware of what’s on the page. You’re going to see me ask some complex pricing questions, show modals and flows that may happen on a product, as well as interrupt the agent, ask it a question, have it answer me back, and redirect.
And so, first demo here is going to be just about two and a half minutes, and then we’ll reframe and we’ll hop in. I am going to start it.
Give me one second here. Here we go.
Nice to meet you, Dan. So tell me, how are you managing your customer data today?
Right now, we use an in house CDP.
And have you seen Salesforce Data Cloud before, or is this your first time?
This is gonna be my first time.
Got it. Let’s start with the big picture.
Data Cloud is the platform that helps you break down data silos and get a single, trusted view of every customer in real time. It pulls data from all your sources, stitches profiles together, and lets you activate those insights instantly across marketing, sales, and service.
That brings us to this screen right here. This is your command center.
If you look at the top, you see the overall health, about 160,000 data streams feeding in, and nearly 6,300,000,000 members across the system. Usually teams spend hours chasing ETL jobs and wondering if the data is current.
This view shows you live counts and the last refresh time, so you know the data is fresh. Because you can see the total records of over 500,000,000,000 and the activation numbers, you can trust that any audience you build is based on complete up to date data.
Let’s click into the details of those streams next. Does that high level health check make sense before we look at the inputs?
This makes a lot of sense. Yeah.
I’d love to see data streams as a whole and then walk me through creating a new data stream maybe from Salesforce CRM.
Great. I’m taking us over to the data streams page so you can see the full list, and we’ll walk through adding a new stream from Salesforce CRM.
On this data streams view, you get a full snapshot of every inbound feed. If you look at the table, each row shows the stream name, source, last run status, and how many records were processed.
Usually teams scramble through logs and tickets to figure out whether a pipeline is alive. Here you can see at a glance which streams are successful, which have errors, and the exact record counts.
Because the last refreshed column updates in real time, you know the data is fresh the moment you need it. This means you can trust any segment you build to reflect the latest customer activity, and you can jump straight to fixing a failing stream without hunting across systems.
Ready to start a new data stream from Salesforce CRM?
Yes, please. Let’s walk through it.
Sure. Let’s open the new data stream wizard so we can pick Salesforce CRM as the source.
To bring data in from Salesforce CRM, you start by opening the new data stream wizard. The first screen asks you to pick a source.
If you look at the tiles, the Salesforce CRM option is right up front, next to other connectors like Snowflake and Google BigQuery. Usually, teams have.
Got you, before we go on, what drives cost here? Is it gonna be, like, the number of streams I created, the records adjusted, refresh frequency, storage?
Good question. The cost model is usage based.
You pay mainly for three things. One, data volume.
How many records you ingest and store in the platform. The more rows you bring in, the higher the storage and processing fees.
Two, refresh cadence. Streams that run more frequently, say, every hour versus once a day consume more compute.
So higher frequency syncs add to the bill.
All right. Hopping back from that quick demo, hopefully I gave you guys a really good insight into it demoing a complex product.
Full awareness of what it saw on the screen wasn’t interrupted by me scrolling or me interacting. Best is the experience we’re building.
We want you to be able experience the product hand in hand directly with an agent. And so it can guide you.
I can guide my way around. It’s a really different take on the modern day demo.
Yeah, the thing I love the most is to point out too, Dan. I know we’re in a couple of different segments and ways that you interact with it.
Sometimes when you get in demos, even in live, there’s just some portions where you just don’t want to be rude to the SE. You kind of let them go down their talk track.
They’re going down an area where it’s like, Okay, I’ll see where they go. With an agent, I think it gives people a little more confidence and maybe comfort to interrupt and say, Hey, this is great, but take me over here.
So I interact all the time. The thing that is most impressive to me is that I can engage with the agent.
I can say, Hey, I’m not interested in this. Take me here.
And the agent can actually do that. They can take me immediately to where I want to go.
I can click it myself if I like, but it has full control over the product. Super impressive.
And then it allows me to scroll and interact as well. That’s the thing.
Saw one the other day. It was more of a video, but I wanted to just go down and see the dashboard or a certain pie chart that was just below the fold in the screen. And when it’s video, you can’t do that, right? If you don’t control the live product. So this allows buyers to get a better feel for the product in the quickest amount of time and for the best experience for them.
Yeah, absolutely. Before I hop into the next demo, I saw some interesting questions.
Some of them we’ll answer as we go through. There’s questions around product understanding and complexity.
So we have a full knowledge base. Saleo as a company, you’re able to feed in your websites, your knowledge base, your articles, which will only better train the AI on how to give really good complex demos.
You can even feed your own best sales calls to help train the agent on what to say on every page. So it’s very, very flexible to the type of technology that you’re working with.
Additionally, we’ve been testing it both raw, meaning no knowledge base at all, just saying, “hey, this is this product. Give me a quick demo of it.”
And it’s done a phenomenal job if you don’t have that kind of content built out. But for companies that have really detailed knowledge bases, we’ve just seen absolutely incredible results in it.
And then in addition, one of the things that we know is incredibly important with demos is the repeatability and the quality of the answers. And so one of the things you’ll see inside the Saleo UI, not on today’s demo, but as you enable these yourself, is you’ll actually be able to get a preview of what the AI is going to say on every page and tweak it, adjust it so that way you ensure it hits your marketing messages and it’s not going to go off script before you ever talk about it.
Thank you for that. I was almost about to go down a long path on mute.
Dan, Boris Lauer had a quick question. I thought you could answer this.
He says, amazing interaction. And maybe I didn’t get this yet, but what is the input for training the LLM? You obviously said that a little bit with the knowledge base, but could you go into a little bit more of all the data, the pages, training that the technology does when it creates the actual container for it, but everything that goes into training the actual agent?
Yeah, absolutely. So for us, we’ve been building this end to end.
It’s one of the things that we’ve always been nudge, nudge, wink, wink about on our sales calls. So for some of you, this may be a little bit eye opening.
But Saleo has always built Saleo Live to understand the interconnections of products that exist in the world, whether you’re one of our large customers like Salesforce or another large app or cybersecurity. We understand how data interconnects.
And so as you’re building these environments, we know natively where data ripples. So if we’re doing personalization, we need the story to go through that.
And so as we build what pages we want to include in the demo and what areas we are, it’s learning from those knowledge gaps. It knows what’s in the demo and what’s not.
So that way, not going to give you any sort of start demoing something that may not be available in the demo or not say anything outside of that. You’re also able to add your own scopes and limits.
But your knowledge base is also the other big part. In this example, we fed in previous sales demos.
We fed in public knowledge based documentation specific to those areas. And it doesn’t have to be thousands of pages.
You can literally just do, this is our knowledge base. It will search and do it in real time.
And one of the really cool things that you’re missing that our agent is doing, it’s like all the fun back end stuff, is that as I’m asking questions, it’s framing its answer. But it’s actually searching the knowledge base, putting pieces together, searching case studies.
So it can bring up anything that you’ve given it access to start using as examples. And so, we say it’s like Progressive profiling for personalization, is what we use internally, which is a mouthful.
But it’s just an incredible tech that’s constantly building the, what could I say next? The web of a tree, a dialogue tree of where am I going to go next? Based on your reactions, on your questions, it chooses and goes down that path. It’s a really, really cool complex technology that’s not just a simple Q and A.
That’s great. Yeah, just a couple more here.
I think that it says Martin Klopf. But he says, Are you able to get metrics on the interaction of the prospects? Yes.
The answer is yes to that. To know if they have a good experience and you can have some kind of a ranking based upon the flow, how far they got through it, and then just the overall propensity to purchase or sentiment from the actual agent.
And then Alan, to your point, Dan, this goes back to some of the training. But how is the training done for the agent? Does talk tracks questions? One of the things that Dan hasn’t pointed out yet is we can actually train it on your best demos.
So if you take the best demos from the SEs, can actually train the talk track and put guardrails in place. Self-service too, right Dan?
Yep, absolutely. All self-service.
You just upload the files yourself. And Alan, to the second part of your question, we do have a cursor that you’ll see in some of these videos.
Some of these were recorded as we’re at active product development, but it will actively navigate, move our cursor and do those things itself as well. It won’t just fade in and fade out screens.
Yeah. One of the reasons you saw that too, Alan, is that Dan was co-browsing with the agent.
So the modals that popped up, Dan was actually clicking that because he was interested while the agent continued his talk track. So that’s what you were saying.
Cool. Should we go to the next one? I know there’s tons of questions we have not answered.
We’ll try to get to those as well. Dan, you want to go to the next?
Yeah, I am pulling it up right now. All right.
So as I launch this second video, quick example. This is going to be very much the same demo, much shorter this time.
This time, I’m going to introduce myself as a completely different person. Pretend like I’m in charge of data at Disney.
And so you’ll see a pivot to a totally different job role. In real time, you’ll see some of those branching examples that we’re talking.
And then you’ll also see a pivot and do some things on complex questions that I ask. It’s about a minute and a half here.
Hi. I’m Alex, and I am here to give you a demo of Salesforce Data Cloud.
Before we start, could you introduce yourself and let me know how familiar you are with Salesforce Data Cloud today?
Hi, Alex. My name is Dan.
For the purpose of this demo, let’s just pretend like I’m a director of marketing tech and data platforms for Disney Parks, and this is gonna be my first time seeing Salesforce Data Cloud.
Great, Dan. Thanks for sharing that.
As a director of marketing tech and data platforms at Disney Parks, you’re balancing massive visitor data, ticketing, loyalty, and real time experiences across parks worldwide. Salesforce Data Cloud can bring all those silos into one trusted, unified customer profile so you can instantly segment guests, personalize offers, and activate campaigns across digital and on-site channels, all while keeping data secure and governed.
How do you manage customer data today? And when you pull it together, what slows you down most? Getting it in, trusting it, or getting it activated?
Right now, we use our own data management platform that we built ourself. But I struggle a lot on trusting it because of the amount of data we get, just making sure that everything’s accurate before we run ad campaigns.
Let’s start with the Data Cloud home dashboard. This is the executive control tower that gives you instant visibility into every part of the data pipeline.
Look here first at the top left KPI tile that reads 169,237 data streams. That number tells you exactly how many inbound feeds are currently configured and actively pulling data into the platform.
For a marketing tech leader at Disney Parks, it’s the first signal that your massive visitor, ticketing, and loyalty streams are being ingested at scale.
Awesome. There we go.
So really cool example, great pivoting, same exact demo. So that’s for some of your questions, what does it take? I’ve got complex customer stuff.
I’m always have to building new ones. In this example, you’re seeing the exact same demo I ran the first time personalized in real time for Disney.
And what you’re not seeing, some of the stuff that’s still in progress to come out is actually going to be the real time personalization of the environment. So as I would have self selected as Disney, all of those data feeds, things you might have seen as data streams or data points would have changed to Disney specific examples.
That’s something we’ll be previewing in a future webinar here. It’s.
Yeah, man, really quick too because it’s, Sorry about it. No, I was just going to say it’s what did not happen, which I think we have just to hammer on this on the coolness of the tech and the adaptability is we didn’t custom code the agent to just say something because Dan said it’s Disney.
It legitimately adjusted on the fly with the agent and picking up who he was, what he wanted to say. So that wasn’t forced.
We didn’t custom code it. We didn’t try to force a scenario that was all natural.
That’s legitimately how it operates. And when you start to play it, you’ll see the same.
You can tell it you’re from any company, for this industry and so on and so forth. It will adapt for you.
Awesome. Any other questions we want to take before we show the third example?
Yeah, we have one example from Boris Lauper. So this example runs on a web app.
Will this application run for demos for on premise applications as well? It will not run for on-prem. However, if your on-prem app is accessible in a web browser, which most are, it can.
But if it is just a desktop app, that it.
So there’s a couple of ways. You would be able to like do screenshots or do other parts of it that you could feed into the agent and would do it.
Obviously, it would be a little less sophisticated because we’re not going to have all the connections for the data. But we do have support for that.
Because a lot of customers have they want to show examples of maybe desktop based apps or non web apps. And so we do have support for that.
Yeah. I was trying to under promise, overdeliver that the experience would not be identical to what you just saw there.
Like the scrolling around of the actual product, it would be screenshot based. But I know a lot of on premise apps for many of our customers are still accessible in a web browser, in which case it would work exactly as you saw it.
Any other questions we want to answer? Well, I’m happy to show a third one or we can go to the list of questions.
Dan, I’m going to ask you this one. By the way, these are on the fly.
None of this is canned and these were not rehearsed. But Dan, I’ve heard you talk about this in the past.
It’s a term I’ve heard you use that you try to mitigate with our agent. But Tony Bowlin says, have you seen issues with people getting uncanny valley with the AI agent?
I love that question. So it’s definitely a concern as you go into AI.
For us, we’re doing everything possible. We’re using latest models, latest voices, latest everything.
We’re continuing to stay on the cutting edge to help avoid that. In the third video, if we get time to go to that, you’ll actually see an example of the same face, different voice.
And so as we get post launch, one of the things that we have on our roadmap is to enable you as the customer to pick what you like best, as well as to bring in third party voice providers if you are interested, which we believe will completely mitigate that uncanny valley. And then for people that may be turned off by an avatar, there’s always, and you’ll see it in that video, the ability to minimize or turn the face itself off.
And so you could do a pure voice agent for those points. I’ve seen that technology get better and better from our early tests.
And some of those customers that are on this call that are already testing it, I know they’ve echoed it. We haven’t seen that uncanny valley.
Even with this voice that you’ve seen in the past two videos, we’ve seen really natural twenty, thirty minute conversations. And we’ve even seen a general preference for early conversations to be led by the AI because it gives a chance to ask a lot of questions without that pressure of a salesperson necessarily waiting on you and really getting to explore yours and learn a lot more about a product before you do that first human led demo.
Yeah, that’s great. Greg, I’ll answer this one.
Greg Rosenberg says, Do you have any third party proof points of Avatar’s acceptance? We do. We did a survey, a large survey prior to the development of the product to say, hey, does it need to be voice? Does it need to be avatar? The vast majority, it was like 83-85% all said that the experience would be preferred with an avatar, not just voice.
So that’s why we index and leave it with the avatar. But to Dan’s point, you can turn it off.
As you saw self-service control, turning off. I want to minimize. I want to maximize. I want the product to be at the forefront. I want the agent to be at the forefront.
So you give the buyer optionality as it relates to the experience that want. So yes, third party data, massive data set that we collected and just doing our diligence.
Hey, do we build this? What kind of investment do we make? And the vast majority did say, yes, we would prefer an avatar. And as you see today, the avatars of today are very different than the avatars of even six months ago.
So technology in itself has advanced quite a bit.
Yeah, absolutely.
Tony Bowen also asked another question. Are you going to show what it takes to build out the demos?
This is an open webinar.
There’s a lot of people on here. The entire market has access. We are not going to show that proprietary IP.
Our customers will have access to that. You’ll be able to see how you can build it out all self-service.
So we’re not going to go behind the scenes to show the IP of how, because we spend a lot of time and energy and even UI. Like how do we make it so simple for them to set up the AI, to train it, to set up the guardrails, to ensure it says the right things, to make some tweaks on the actual experience? That’s part of our secret sauce.
And so we won’t be showing that to the public today, but it is available.
Yeah. So two other questions that I’d love to take, which I think will tie us into this third demo.
One, Vinesh, you asked a great question about, will context route become real? To categorize that for you all on the call, context is a big part of AI, where it starts forgetting something after it collects so much data. We haven’t seen that.
So we have over 1,100 times we’ve talked to this AI. We’ve done really, really long calls.
And we’re optimizing for twenty to thirty minute calls, but we have done hour long calls. And so we’re not seeing that.
So we’re building that progressive profile, is feeding in as we continually go through the pages that we haven’t experienced, especially with the type of things covered in a demo, like your top pains. The context just isn’t really bad.
But to us, we have seen zero instances of any sort of dropped context for longer calls. And that speaks a lot to the architecture that we’ve built.
So Dan, I’m going to ask what some may think is a controversial question, I think it’s a fair question that people are picking up. It has all to do with the structure of Zoom and multi microphones and then picking up.
Tony said, can I ask why you’re showing a video instead of demoing it? As you know, when you’re actually designed to engage an agent, it’s meant to be one to one, not over a Zoom mic and then another mic and then two audios. It’s near impossible to do well over a Zoom that’s being recorded and then being shared out.
So Dan, can you kind of speak to that? It’s not that it’s the best way that you can do it based upon the microphone and logistics.
Correct. Yeah, with the current build of the agent.
And we are working on a north star of having to actually actively join Zoom calls as like an active user or Teams and things. Right now, the shared microphone, and especially I’m on Windows, the exclusive microphone settings, cause all sorts of contradictions, can cause noise in the line, can cause it just not to respond because the data is delayed.
And so we chose to do video for this. But there’s a large number of people on this call that are already actively using it.
It’s out in market having those experiences. In the summer near future, we’re going have a public version where you’ll get to interact with it yourself.
Yeah. It just was pure Zoom webinar.
It does not play well with the agent’s designed to interact with someone audio in a setting. So in the real world, it works perfect when you’re trying to deal with multiple microphones and Zoom recording and Zoom settings.
That’s why. Alan Chan says, does it support other languages? Yes, it is multilingual.
Yeah, we are fast, fast running. We’re going to probably be launching with support for around 30.
But as we’re looking at some of the new versions of models, we’ve seen them go up to 70 plus languages. Our goal is to make sure that English first, top priority, making sure that the demo is end to end.
But we’ll be continuing to adopt a lot of them, multilingual because we service a lot of enterprise customers. But we have active tests going on with it right now, multilingually.
And so to answer, we have a lot of questions around custom avatars. So we already mentioned the forthcoming voices.
We’re going to have a library of avatars and appearances for you to pick at launch that you’ll be able to go through. But as I open this next video, it’ll be a great example of an avatar, same avatar that you saw before, but different voices.
This one is wildly different than the last one. Little a little bit of like a valley girl take, just to really hone in like this is a very different voice.
This is not a launch voice or anything for the agent, but a great example just so you guys see some of the for the futures of allowing to choose it yourself.
Hi. I’m Alex, and I am here to give you a demo of Salesforce Data Cloud.
Before we start, could you introduce yourself and let me know how familiar you are with Salesforce Data Cloud today?
Hi. Yes.
I’m Dan. And for the purposes of this demo, I am actually a VP of marketing at McDonald’s.
And this is going to be my first time ever seeing Salesforce’s Data Cloud.
Great to meet you, Dan. As VP of marketing at McDonald’s, you’re driving massive campaigns that need fresh, trusted customer insights.
Salesforce Data Cloud brings all those data silos together. CRM.
There we go. Just a little quick preview of some custom voices for y’all to see some examples.
But we’re going be launching with about 10 different physical appearances that you’ll be able to choose from. Each of those appearances will have their own dedicated voice.
And then as we expand from there, you’ll be able to choose different voices for different appearances. We’ll be adding more appearances over time.
Yeah, there’s some people asking questions about where we get access. Can we have access to the live agent? So the answer is yes.
You’ll have to go through our sales team work. There’s a large demand for people that want access to it.
Obviously it goes back to some of the pricing models, right? How is it priced out? So we’ll be by minutes and we obviously want to control. We don’t want anyone and everyone to be able to interact with it quite yet.
So if you do have interest, if you’d like to play with it, feel free to reach out to the sales team and they’ll put you into a process where we can put you on the list to be able to engage with it.
Absolutely. Yeah, I see some more questions coming in.
Yes, the agent will be able to share links to docs. You’re not seeing that in the preview today, but there is a chat functionality inside of it where it is going to be sharing links to docs that you can go it in.
And is it people switching between the product and a slide deck? Yes. That is one of the features that will be inside of it.
Yeah. And Justin Skides had asked us, say they’re looking at different specific accents of the voice agent.
I’m laughing because we were absolutely. Yesterday.
And Dan went through some Scottish accents, some British accents, some German accents, some German language. So we saw quite a bit yesterday.
So Dan, I don’t know if you have any more to say on that.
Yeah, tons. And as we enable custom voices, that’s one of the things that we think is going to be very important for multilingual, is having a voice that sounds similar to the language that the people are picking.
And so you’ll see a lot of that coming from us. And then, Sarah, you asked a great question around customer qualifying.
And I apologize. I should have answered this way sooner.
So all these agents and all these demos, as you build them, you’re actually able to give a list of questions that you want the agent to ask. You can even assign those questions to certain areas inside of the demo.
But all of those questions and all that document will sync directly back to our own analytics section and into Salesforce as well as a custom object in the future so that you’re able to see the answers to them. Because qualification is the key part of these.
And so you’ll be able to load whether you have a playbook, whether you have a dozen questions. And it’s a per demo basis.
So if you have different demos that you’ve created for different types, you’ll be able to readjust those qualification questions, the pieces of data that you want to collect, go. And then Nayson, you asked how much documentation do you need to make the agent fly.
We’ve covered we have a knowledge base. It can be as simple as just sharing your website, or it can be as much as uploading previous calls, or best sales calls or sales playbooks.
It’s really completely up to you. We’ve seen great luck across the board on very minimal training.
Hey, I did this, and I built this another twenty minutes. We talked about that self serve.
You can turn these on very quickly. And we’ve seen incredible luck with training.
Obviously, everything gets better the more context you give AI. But for most products that are publicly mentioned, have a few webinars and case studies and things on their website.
Typically, if you’re the type of company that’s joining this webinar and can buy a platform like Saleo, you have the documentation already needed. I’ve seen no cases where we’ve needed additional data unless it’s a product that is not talked about.
Maybe you have one or two employees. And then you would want to add some stuff.
That’s great. Well, I know we’re running short on time here.
We’re going to try to end it fifteen before the hour.
We’re really excited, this has been a fun season for the Saleo team.
And excited for all of our customers who are going to get access to and start using it. Appreciate you all attending.
And yeah, feel free to reach out if you have any questions. Otherwise, we’ll see you on the next webinar.
Thank you all.
