Transform Campaigns into Two-Way Conversations | Dreamforce 2025

You.

My name is Brett Bothie and I work in product marketing for the marketing cloud at Salesforce and I'm joined by my colleague Ajay and Rajesh, the CTO of Rightling.

Hey everyone.

So we have a pretty exciting session today, so we're going to dig right into it, but ultimately we're going to talk about how you can transform every message into two way conversations and how we are delivering that as Salesforce. And then we're going to go through Breitling's use case. So to kick things off, I think it's important to level set around how customer expectations are changing. With AI, we've all built really strong habits with our favorite search engines. We know exactly how to craft the right search, we know exactly how to browse those sites. Maybe you go to a second or third page, but probably you don't and you're ultimately selecting what you want. We're now in a world where you can prompt half of a thought and that thought will be finished for you and then the recommendations that you're looking for in that thought are served to you. So what does that mean for us as marketers? It matters because every customer now is expecting instantaneous and personalized interactions because that's what they're getting as a general consumer. So that's what they expect from your brand. Now I want to pose a question. What if every single interaction, every single marketing message could actually start a two way conversation in today's landscape? We send emails with the handle do not reply. We are not a customer facing sort of organization in the way that customer service is. And if someone has a problem or needs to get ahold of us, we, we send them to a different department. But what if marketing could become that department? What if marketing could be a lever for communication back with customers? That's what we're really exploring here today. So two way conversations can completely transform relationships for customers. It really can feel much more like a relationship with a brand rather than you're getting an email send that went to 1,000 people. And it's truly customer led interactions that are adapting to change in context because no two customer interactions are the same. If you are orchestrating journeys in your organization, then you understand that. And lastly, answers and actions are made immediate. So there's no more waiting, there's no more, oh, I sent this note and I'm looking to get back or I tried calling this apartment and I was put on hold and I didn't have the right contact. We can now sort of come to customers as an Entire brand and marketing can be the front door to that. So for marketers, that means you can qualify and nurture leads a lot more effectively. You can scale governance really across any channel. You can add qualitative insights and explain to the customer directly and ultimately you get much stronger results in terms of engagement, conversion and loyalty. So how are we powering this? How is this possible? Why haven't we explored this before? Well, everything is different in a world with agentic AI. So you are going to get an input from a customer and on our end you plan and run. So what does that mean Agent force is doing? It's searching for the right topics and instructions. So this is how it understands how to act and then it is taking actions which are very specific tasks, so they're task oriented. So it's going to pull all of these things and, and run the job. And what it's going to do before it has an output is it's going to ground in what you know about your customer in that exact context and it's going to route based on the appropriate activity, the appropriate action, whether that's an agentic response or potentially a human one. So what you get in terms of the outcome is a super personalized two way conversation. Customer's asking for something specific and the agent is able to understand that. It's able to look at their purchase history, it's able to look at their browsing data and pull in a extremely pertinent and relevant conversation. So how does this come to life now? We are thinking about this right now in three primary channels. So we have mobile through WhatsApp and SMS and we have web where you can make every single site feel like it was custom built for them. And that's what we're going to go through today with Breitling. And lastly, we have email, which if you saw today the keynote, the marketing keynote was at 10am if you were there, we went through two way email, web and mobile. But two way email is the concept of actually responding, getting rid of do not reply. So like I said, we're going to go into web a bit further and understand the demo and the use case from Breitling. So to do that, I'm gonna hand things over to Ajay to introduce us.

Perfect. Thank you so much, Brett. And in case all of you haven't noticed it, I have fully embraced the yellow team in solid support of Brightly.

I have the blue one by the way.

So I am joined by Rajesh who has been an absolutely wonderful partner as we have started Kickstarted the journey on adaptive websites. So simple. Where did this thought originate and where we came from is if you look at it today, every website is pretty much kind of feels like a brochure or a catalog. People have to go through multiple pages, try to find the information that they're looking for. It is hard and it's very important. As we have spoken and you can shine light on this, how important is it to meet the customer in the moment? So could you give us a little bit of an introduction about what Breitling does and what we started doing with adaptive websites here, Rajesh.

Thank you. Hey everyone. I think, I hope everyone knows Breitling simple business. We are Swiss watch manufacturer. Almost 140 years old. 140 years old. With the full of history, heritage, innovation and precision. If you don't buy one, if you don't have one, just go and buy it. We have a boating just down there. Point number one, I'm, I mean how this idea all came in. I think we wanted to. Breitling is a more emotional buying. Like every luxury watches, every luxury brand is more emotional buying. It's just not like happens like one or two transactions in our history. It's like the buying journey was talking about three to six months. It's always emotional, human based. Point number one for that you need to provide that. In a physical store it's perfectly fine because you have people in front of you, they take care of you, they give you champagne and everything is fine. But when you go to the website, how do you make it much more personalized? That's what I call come back with the hyper personalization. So that's where the whole idea came in. So I talked to Ajay and the team. That's where we come up with this idea. The two way conversation on the website. But again still focusing more on storytelling, brand product discovery, not about transactions. Yes, we do transaction online, but that's not the primary focus. More customer. Customer engagement and personalization. That's how the idea came in.

Perfect. That is awesome. Actually, since you mentioned about conversations, I'm sure all of you here are conversing with the agents that you have on your phone. And if you look at my chat history, I actually chat more with Gemini than with my wife, which she's not happy about. But no, no. That being said, talk to us from your customer's point of view, Rajesh, like an anonymous customer to a website. Right? Like how important is it for Breitling to understand who this anonymous customer is and not to lose them and be able to hyper personalize content right from that very first touch itself.

Right. I mean most of the brands, I mean we get anonymous traffic almost 90, 95%. I mean our registered users pretty less like most of luxury brand. But if you look at the top of the tunnel, the first tunnel is always the product discovery. So that's where we lose our visitors and customers, Right, potential customers. So what we wanted to do is like how do we make sure that we don't really lose all of them but try to convert a bit either at the other funnel or maybe all the way to the conversion. So where we started is, okay, so let's make sure that we address this particular product discovery as main use case because at the end it's more about finding a product. As I said before, you aren't going to buy the Breitling watch in five minutes. It's going to be a journey, is going to be a journey across online, offline, contact center, whatever you want to call it. Right. So that's where the whole journey starts. So more human element. That's what everything started.

Absolutely. No, that's very, very important. So as Rajesh was mentioning, the journey. Right. So it's not about telling customers what to do or just showing things to them, but it's about how to involve them and how we can have a conversation with them and take them along in the process. So if you look at this demo that's running behind us, so that's kind of what we have tried to do with adaptive websites or the adaptive web agent, as you see. So Rajesh, could you just walk us through? Can we just start from the.

We'll try to start from scratch maybe.

I'll probably stand a bit.

I can see it from that first. Yeah, good.

Yeah. Ingress point, please.

Thank you. So you are in the Breitling website. I mean if you see there is a small window at the bottom. So first one I'm going to ask, okay, I'm a casual visitor, I want to buy a watch, which I can afford from a casual luxury perspective. Right. So that's what I ask as a customer and then the agent thinks and the agent provides me the recommendations or the content. If you see here, you're not leaving the experience, you're still staying within Breitling website. One side you have an assistant, other side you have a completely personalized content here. If you see it's not a link, I'm not asking you to go the link and go to PDP or something. We just completely changed it. So look at all the casual luxury watches, what we see it here, but I also want to know, okay, Breitling is all about aviation and stuff. So I also like gold. Do you have gold watches? Oh, yes, of course I do have it. So that's how if you change. Again, if you see here, every information about the product and all the gold watches has been listed down. Again, the content is completely changing, as you ask. So that's a core idea of this. It's not a pure AI assistant itself. It's also about the way we change the content. Okay, okay, I see all the watches, but I know a lot about Breitling. It's more aviation brand. Can I have more Aviation? Gold, Aviation. Look at this. All the Navitimer watches. That's our core collection, right? That's the next one, right? So again, if you keep seeing it, the customer still stays within the whole journey, not even leaving this. The engagement comes into picture. The next one. Okay, fine. I have all the gold watches. Okay, fine. Let me find a red color watches. That's not my favorite color. That's my wife's favorite color, by the way. But I don't have a choice. So that's my favorite color now. So red color.

Well said.

Happy wife, happy life, guys. So that's like red color watches. Okay, so if you see what is more important every. I mean, some of you may be from retailers, some of you from the brands. The top of the funnel is where you lose your. All the marketing investments. And here the whole idea would be, how do I make sure that I get them converted in this one? Not from a transactional perspective, but again, from a customer engagement perspective. Okay, so I see. What is it? Next one. Yes. I'm in San Francisco. I mean, I see some NFL watches. I'm in San Francisco. Do you have any NFL watches? 49ers. I'm a big fan of 49ers. So there is 49ers. There's like multiple watches. We have with two or three or four watches. Right? Or two watches, actually, for NFL San Francisco, I like there are two watches. You know what, I want to really compare them. What's the difference? What's the difference between those two? Right? Then I compare the watches. So the product attributes, is it size, the price, all the information. We're comparing it again. This compared it. If you look at this one. So we have one 49ers with 40mm and one with 42mm the size. So I like more of 42mm, you know, 40 probably bit less. I like more 42 right. So that's what this is comparing again, helping me to compare the product. Product discovery all the time. Okay, so this is what comparing it. Okay, I like the second one. Do we have it in the stock? I need to know. I mean I like the product. We have it in the stock. Of course. Yes. So agent does checks, all the stock information. What we have it. Okay, I have it in the stock. Then once you have it, this is where we redirect to the actual checkout, actual PDP page, product details page where you have more information. But we don't want to convert this guy in online. But we want to bring them into offline so that we can still provide a much better personalized experience in the butching. So the journey would be that okay, I'm going to book an appointment in my boutique here. Right. So the customer journey would be okay. I ended up instead of going into Google, instead of going to all the stuff I'm comparing with all the, instead of I'm having all the experience within my Breitling website. Finding a product, getting all the information about the product, finding the right product, comparing the product, gold, red diamond, all the watches, we were there. And then the boutique appointment happens. Then what happens for the customer? Customers goes to the boutique. And when he goes it. Our idea would be to bring all this data already in front of the sales associate. So by the time the customer walks in, this guy has everything about okay, what he searched the complete history. What do you see it here? The questions, what he asked, the answer, what you provide the complete conversation history. So that's why we are more hyper personalized. But what did you achieve here as a brand? It's very simple. The visitor is supposed to be an anonymous visitor. We converted them into a more. More. A more prospect. Potential prospect. You could even buy it there as well. So that's how we're talking about the complete way conversation from a brand perspective as well. Okay. One condition does not hit life. I don't want to lie. We still wanted to improve a lot on this. We wanted to bring more innovation in this. And we are going to be live sometime in January but we are co developing it with Ajay and the team. A lot of work has been done on this. Thank you.

So. Thank you, Rajesh. No, that was excellent. So if we go take a look at, you know, the different products that were surfaced in that agent panel that you see there. So all of that content wasn't created by an agent. Right. All of this content already lives on Breitling's CMS or dam. And what we're doing is we're just rendering it in real time and you can customize it however you choose. So wish. So there's one thing Rajesh, which you mentioned which I thought would be helpful for people is even just the ordering based on a question that a user asks. And let's say if they wanted to see gold watches or if they wanted to see red watches. We spoke about what is the right order in which these recommendations should be surfaced. Can you just talk a little bit about the importance behind that and how we can actually have customizations even inside the responses from an agent.

I mean it's called visual merchandising. I mean when you go to the boutiques, when you go to the they know which product they showcase first, which product they showcase second and third boutique is perfectly fine. But how do we again do this in online? So we have a specific rule. So first we promote a Collection Avitimer Gold watches 4246 mm, 44 mm 42 mm. So the way we even show the results here has to follow this particular what we call a visual merchandising for the boutique search and dicing in online. So we wanted to follow this rule today is not the case. What we are going to expand the scope is actually implementing those rules. And not only that, we are trying to add more product information about every single information, warranty information, every single thing into this web creation or predictive website agent. So that's one. The second part one is search and rising. Second is more product attributes. And third, it's also about why don't we even ask a contact information. What if customers do not want to book an appointment? Just leave an email address. That's enough. That's already success for me. And then we can use all the online offline marketing campaigns and stuff to follow up with this. Yeah, a lot of ideas in there. But we wanted to first test this works. But it seems working. Thanks to the guys.

It seems okay.

Yes.

So no that's. Thank you very much for that. So the key thing here that we wanted to call out is that we really care about your brand. And we wanted to make sure especially Rajesh working with you at Breitling also that one whatever. However we design this panel is like on brand and brand compliant to whatever it is that how you want this to behave in front of your customers. And then there can also be customizations inbuilt on top of it to help direct the agent to give it instructions in terms of, you know, you have to Behave like a sales associate in my store. So do this, this and this in this particular order. So that's the power of actually having agentforce driving adaptive websites here. And as Rajesh was mentioning exactly right now, what you're seeing with the capability of, you know, scheduling an appointment to walk into a physical boutique. So that is something that we're working on to bring into the adaptive website agent itself in the future and being able to collect contact information. So there is a lot of innovation that we've planned for and is coming up. So we're really looking forward to that. Very important question. Rajesh, as you know, people here are looking to get started. What piece of advice do you have for people like how can they get started with this and you know, what should they think about as they get started?

Two points. Point number one, without data, there's no AI. It's like having a Ferrari without an engineer without a key. Okay? Get your data consolidated, right? Quality of data in one single place. It's absolute, absolute mandatory for any AI, what you've been hearing the whole three days. Second, start something very small. Really very small. Don't try to improve the top line, bottom line, just start something very small. Okay? If you fail, perfectly fine, it's okay. You can, you can learn and move on. So this is the only two advice which I took it back in the time. This is something successful for me so far. These are the only two things I would say if I can give it that way.

No, absolutely. That was extremely helpful. Thank you so much for your time.

Thanks, guys.

Yeah, thank you so much. I mean, so this feature is generally available. Line and drive. The guy who built it.

December of 2020. Record it. Okay.

Record it, guys.

So already recorded December.

Exactly. This feature is generally available in December. But I want to thank Rajesh and Breitling for being an awesome innovation partner and being a beta customer of adaptive websites. So how do you get started with two way conversations today? So we have amazing advice which is to focus on your data strategy and start small with a small use case. And we want to make sure that all of our marketing cloud customers, you have AgentForce, you have access to AgentForce marketing. So you want to activate it and you can talk to your account executive to do that today. But you wanted to find your use case in your channel, pick a lane, start there and essentially evaluate that channel and how you want to improve it. Have a KPI in mind that you're looking to improve with a two way conversation. So that's part and parcel with the use case suggestion and then test in a low stake scenario. You want to start small, you want to review results. And one of the things that we don't often talk about, I think, in tech adaptation is building a culture of learning and making sure that your team is feeling comfortable with this type of technology, with learning how to build these use cases, and comfortable with AI and its output. So you want to select a champion with your team, you want to celebrate all of the small wins and because they'll lead to really big and great places. So that's what we have for you today. But thank you so much for joining. Here's some guides if you'd like it.

Thank you, guys.

Thank you. Thank you very much.