Speaker 0: Alright. Hello, MarDreamin. Welcome, everyone. We are so excited to have you all joining us today. My name is Courtney Tranelia from Sercante, and I’ll be moderating today’s session. I have housekeeping items to cover. Yes. The sessions are recorded, and they’ll be available on demand after the event. You’ll also be following up with them via email. If you have a question, please post the the q and a in the tab above. And lastly, use the chat. There are emojis, gifs, and more, and we wanna hear from you as the presentation goes on. So let’s get started. I’d like to introduce you to our speakers today, Lucy Maslon and Lauren Metcalf, who have an awesome session ready for us today all about can GPT replace a Salesforce marketer. Welcome.

Speaker 1: Thank you so much, Courtney. And, uh, it’s great to return to, yes, another MaDreaming. Um, so good morning, good afternoon, good evening, where wherever you are. Um, thank you for choosing our session today. I know there’s lots going on at the same time, so we really appreciate it. Um, so my name is Lucy Maslon. Um, I’m the operations director at SF Ben. Uh, you might not recognize the new logo. We just had a rebrand this week, so we’ve gone from the classic blue, Salesforce blue, to purple. So, uh, actually, it blends in very nicely with this slide deck, um, as well. Um, so at Salesforce Ben, we are a media platform, um, where people from the community contribute, um, content, share their knowledge, share their ideas, um, but we’re always kept on our toes because the world of Salesforce doesn’t stop moving. I was, um, formally a Pardot and Salesforce consultant, um, for a few years. And, um, through that, I founded The Drip, which, um, has now been fully merged with Salesforce Ben. So all the marketing content is on Salesforce Ben now. And, um, I’m also a marketing champion, and I’m joined today with my colleague, Lauren.

Speaker 2: Hi, everyone. Um, it’s really exciting to join you today. This is my first module in, so, um, I’m excited to get involved. Um, I’m the content and services delivery lead at Salesforce Ben, um, but I also work closely with Lucy on the marketing content, um, covering everything to do with Salesforce marketing on the platform.

Speaker 1: Yeah. It’s, uh, never a dull moment.

Speaker 2: Yeah. Never. Um, so, yeah, so we first want to say thank you to the sponsors. Uh, this event wouldn’t be possible without them. So big thanks.

Speaker 1: Great. So today’s agenda is split into three main parts. So we’ve got I put upcoming reality because in some ways, it is the reality, um, with all these new things coming onto the Salesforce platform that we can get our hands on. Uh, some, um, there’s like a staggered rollout, so didn’t wanna say it is available, but, you know, that’s the little upcoming. Um, so, yeah, we’re just going to show a very brief history of Einstein GPT, GPT, which is now Einstein one, and talk, um, about what Einstein one for marketing does, um, which, you know, will be the use cases that are currently available. Then we’ll go into more can ChatGPT sorry. Can GPT replace a marketer, not specifically ChatGPT, which we’ll get to in a second. Um, so we are looking beyond just Salesforce as a technology, looking at all the different, um, areas that marketers, you know, the responsibilities that marketers, uh, juggle day to day. So we’ve chosen three of those. And then, um, couple of months ago, I kind of let my imagination run wild, uh, to see, like, oh, okay. What else could be added, um, to the platform in the future? Uh, so, yes, as I mentioned, um, let’s just clear this up. GPT stands for generative pretrained transformers. You don’t need to know that, but GPT is a technology concept, not a specific brand name. Um, I think ChatGPT has become such a household name. Um, it’s, you know, it’s easy to confuse the two, but ChatGPT is a product. GPT itself is a technology concept. Alright then. The upcoming reality.

Speaker 2: Yeah. So this slide has a lot on it, but don’t worry. We’re not gonna go through all of it. We just wanted to, um, show, like, the speed of innovation and the adoption of, like, new technology. It’s really unprecedented and Salesforce’s Einstein AI offerings have moved very fast. For example ChatGPT isn’t even a year old yet, it’s like birthday’s in November sometimes so it’s coming up to a year that ChatGPT has been around. The from this from March is when they announced GBT, uh, Einstein GBT. Sorry. Uh, and it shows how Salesforce are really heavily invested in all corners of their platform with the big switch to Einstein one in September, Juneforce.

Speaker 1: Lovely. And specifically for marketers, um, there’s, like, three, I guess, phases that we’ve been through. And, again, this is all within the space of six months. Um, so they’ve taken us on a wild ride. Um, so as Laura mentioned, in March, um, Einstein, uh, sorry, Salesforce launched Einstein GPT, which I’m calling, like, the v one. Um, so what we saw was, um, a demo that showed didn’t actually label it with it didn’t actually show what technology it was actually being used on. So it was it like marketing cloud or account engagement? But what they were just trying to show is how, um, you can generate personalized content for email, mobile webinar advertising, amending the tone, the length, etcetera. Then at connections, um, in June, um, we actually had more of a this is specific to, um, one of the Salesforce products with marketing GPT. Again, lots of use cases that we’ll dive into in a second. And then, um, as I mentioned, Einstein One um, is the new name essentially for the Salesforce platform. And, uh, so that’s why we’re calling it Einstein one for marketing. Uh, there’s actually too many to name. There were 13, um, capabilities for marketing cloud alone. Um, there was another 13 for commerce cloud, for example, and all the respective Salesforce clouds. So but a few of them to highlight is the dynamic content, advertising audiences and look alikes, and segment intelligence.

Speaker 2: So, um, here we’re just gonna share a few of the use cases, um, which Salesforce have targeted and what GPT Technology is applying for. This one is email and subject line copy, and Salesforce led with a phrase with Einstein and email is just a sentence away when they launched it. In this use case, you can generate messaging faster and have the end result reflect what was previously done well in your company’s past campaigns. You can also adjust and refine the tone, for example, generating three variations based on top performers, and you can choose from them. But, ultimately, the market has the decision, uh, making power, and they can accept the suggestions, or they can also opt to AB test.

Speaker 1: And another prime example for where Jhat GPT can really help a marketer is all in terms of segmentation. So, um, I won’t go through each of these in detail because some of them when I describe them, they’re gonna sound similar. Some are, um, available for marketing cloud. Some are available you have to have data cloud, um, actually licensed for it. Um, but, yes, as you may know, data cloud essentially underpins the whole Salesforce platform. It’s, um, providing the connectivity between clouds, so it gets the data flowing faster. However, it’s also a product that organizations can buy, um, for that, uh, profile unification and, um, segment activation, so pushing it to destinations such as marketing cloud. Um, so, yeah, with with data cloud, you’re essentially, uh, supercharging all of that. And here’s, um, how it looks. You know, it’s conversational. You’re writing prompts to get a segment, um, back from from Iceland GPT. And it just prevents, uh, us having to look through multiple reports, um, and sift through rows of data to understand what segments we can pull and what you know, you obviously want to check that when you’re campaign planning, you actually have a decent amount of people in each segment before we start going too granular. Alright. So can GPT replace a marketer? And the short answer really is no, but we all know there’s nuance and the the short answer isn’t always reflective of the whole picture. So marketers do need to be adaptable, adaptable, you know, embracing this technology, but not lose sight of their value. Um, I put ensure others don’t undermine them because there were lots of, uh, I guess, anecdotes, um, like, oh, I can replace my whole marketing team with AI. And it’s that’s just, um, not not the case. Right? Um, but you don’t want, uh, we still wanna be showing our value. And that’s, you know, with the human in the loop, intentional friction is how sales calls have developed their features, which pretty much means, like, AI won’t go off and do something unless you tell it to. You you have to, like, validate at each stage. Um, and then there’s increasing pressure that’s being placed on marketers. Um, so it’s about time that we can get support from a companion, almost like a virtual assistant, um, accelerate the most time consuming tasks, which in turn allows us to flex our creative muscles. And, um, as, uh, you know, marketers attending this conference, I’m sure that everybody has some tech curiosity, uh, and we all wanna work smarter. So, um, before we go forward, what follows, um, after we’re gonna show one slide with some survey results, um, and then what follows after that is opinion based. Um, we put together this content, um, based on Salesforce’s technology progression, as in where they’re going forward, um, anecdotes from marketers, and, um, also, uh, marketing technology thought leaders beyond the Salesforce ecosystem. Um, so just taking little tidbits and putting it all together.

Speaker 2: So, yeah, here are the stats that this research that Lucy was just talking about, they’re from Salesforce research. Um, the first list shows how marketers are using generative AI in their day to day. You can see that basic content creation and writing copy are like the two most common use cases there at the top. And then on the other side, you can it shares some of the ways, um, that AI is transforming a marketer’s job and this includes analyzing marketing performance data and creating segments for marketing campaigns. So that’s just some of the the key use cases that Salesforce themselves have identified. But there’s more to marketing than Salesforce. There are many responsibilities marketers have and this goes beyond any technology. Here we’ve picked 10 areas where marketers spend their time. We have marked whether they are covered by Einstein with a check showing yes, a cross being no, and and the wiggly line demonstrating that there’s some coverage but not entirely. We’ve also assigned a GPT impact rating. This is whether the task the task could just be disrupted by AI in the foreseeable future. But in this case, disrupted is not a negative statement. It’s been disruptive in a way that has a positive end result. So we’ve already covered content creation and segment creation. Now we’ll put a spotlight on one from each of the rating. These three are analytics with high impact, social media for medium, and user adoption for low.

Speaker 1: Great. So analytics. Um, personally, this is the one that I’m most excited about. Um, while Salesforce has done a good job of embedding that query based exploration, by that, I mean, you typing a prompt and it gives you, you know, the answer from all that reams of data that you have, um, across platforms. Yeah. So, you know, we use other systems as marketers, other data repositories. Um, so I got interested in potential for Google Analytics to query that data. Um, I don’t know about you. GA four has been a bit, uh, different to, uh, use. So, yeah, just, like, improving that adoption and also improving adoption for other people in the company to, um, get more in more interested and more in tune with the data. So how could GPT have an impact? Imagine being able to create a repository of data and using prompts to surface what you need. So, yeah, as we see on the side, that’s what, uh, segment intelligence for data cloud is getting at. Um, as I mentioned, the sifting through multiple rows of data, um, or multiple pages in Google Analytics could be a thing of the past. I’ve heard others using BigQuery and OpenAI, uh, to take that even further. Um, but, obviously, you know, you have to make sure it’s a secure environment, um, when you’re sharing data like that. Um, so why there needs to be human in the loop? Well, actioning insights. Um, there is a trust issue. Like, do the results look right? Have you is there a dataset missing? Um, is the data fresh? Um, so are anomalies being exacerbated? Um, so, yeah, there is that element of there still need to

Speaker 2: be a human in the loop. And then for social media, AI can elevate social media to determine strategic direction, foster human connection, and uncover customer perception. Two main areas of this are prospect slash customer engagement to generate new business and loyalty and social customer care, for example, customers reaching out for assistance via social networks. GPT could have an impact because you can respond faster in an accurate way with something that is valuable to each interaction. And there are a lot of use cases for this as well, including social content creation, generating suggested replies, determining optimal send times, identifying target customer segments, predicting customer behaviors, and optimizing pricing and product listings. But as always, there needs to be a human in the loop. Social media needs creativity to help brands stand out from the crowd, and this reflects back to the intentional friction approach where a user still needs to validate the outputs for accuracy on what they want their brand to portray. Now they’ll be doing this efficiently and at scale.

Speaker 1: Yeah. Absolutely. Um, user adoption. So, uh, user adoption, if you’re not familiar, measures the extent to which users have incorporated Salesforce or other tools into their day to day tasks. Um, without users actually using Salesforce, um, the platform begins to become less of a source of truth as, you know, stale data, incomplete data, breeds mistrust, um, which further deteriorates your data quality. And that that’s just you’ve got all this shiny technology. You don’t want to fall down at the last hurdle by not having good user adoption. So how GBT could have an impact, um, we also talked in in this guide that we published, we also talked about documentation. Um, so similar to that, you could, uh, leverage GPT to summarize large amounts of information on the solutions you’ve implemented. Um, you know, make the tone more approachable, um, change the length of it, like, make it more concise. Um, so if your internal communication goes unnoticed, uh, it could be a good time to look at rephrasing. Um, why there needs to be a human loop change management is, um, how organize sorry. Organizations implement and adopt changes and then iterate upon them over time. So this is a highly sought after soft skill and requires a human to human approach just so that there is no one answer to change management. There are tactics you can do, and you just really need to have a pulse on your user base and be adaptable to that. So, you know, GPT can’t tell you how to do that. Um, it’s all dependent on your organization, and as that changes and that as that user base changes over time. Great. So now we get to the final part, which is the imagination section. Um, like I said, couple of months ago, um, I let my imagination run wild. So, uh, here’s the first one. Um, account engagement optimizer was was a really nice addition. I really like it, um, which apologies. Um, which became generally available in the summer twenty three release. Um, it gives you a big picture view of your business units in account engagement. So if some if items need your attention, they’ll be flagged, um, and you’ll be presented with proactive, uh, recommendations to take action quickly. So wouldn’t it be great if you had an injection of predictive insight? Um, not only how your account is performing now or has been recently, but also running what if scenarios. Um, one example I I thought of would be, you know, if if I added all these marketing assets or prospect records, how would that impact the connector sync capacity? Um, and, sorry, before we move on, I just want to make it clear that this is totally speculative. We don’t have some magic crystal ball. We don’t we’re not taking this from Salesforce’s real road map. We just, uh, thought we’d share this with you, um, and see if you agree that these would be good additions, um, and if you would use them. And, of course, if you’ve got any more to add, that’d be awesome. Right. Sync errors. Um, so, again, this is account engagement. So this is the connector between, um, account engagement and Salesforce. Errors can happen for a number of reasons, um, like, uh, a Salesforce admin can change a pick list value, and that impacts what values account engagement can sync to Salesforce, things like that, especially if your Salesforce org is in a state of flux, lots of changes happening. Um, you could log in in the morning and find thousands of sync errors. I have been in that position before, and it’s, uh, yeah, it’s not it’s a bit grisly. Um, so, yeah, how could predictive insight help here? Like, be alerted to sync errors and maybe have ally AI analyze the root cause. Um, I mean, we don’t we can’t let GPT run wild, but you could actually fix or recommend fixes, um, for that. And then finally, testing marketing assets. So you’ve got so many assets and so many automations going on. How do you know that um, there they are working correctly? Um, you know, not broken pages, not forms form field changes, like I just mentioned, for example, with the pick list values. How do you know that they’re working and working together? Um, so this come this brings me to test automation, which is a category of technology that essentially systematically goes through each possible scenario. Um, so each possible if this happens, then this happens, and then that gives you, like, almost, like, a pass or fail, um, scenario. But it it goes through to the extent that no human could feasibly do in a in a short amount of time. Um, so, yeah, just would love to see some test automation happening. And, obviously, there are providers on the market, but, uh, uh, yeah, I think it’s, uh, a one that would give, uh, marketers and admins peace of mind.

Speaker 2: Yeah. On so on Salesforce, we’ve also got other articles about how GBT could impact that each role. Um, this includes admins, developers, and architects. And then also if you wanted to read more about this, we do have one on marketers too. So Lovely.

Speaker 1: Alright. So we’re, uh, we welcome your questions And, uh, yeah. Courtney, over to you.

Speaker 0: Hey. Thank you, Lucy and Lauren. That was an amazing session. I really like your imagination section, how we can envision the future, um, using these tools that we already have today as a platform. Um, we do have a few minutes for q and a. So if anybody watching wants to drop some questions into the q a tab or in the chat, we’ll watch for those. Um, I have one just kick us off. So I’m curious to hear from you both of what skills do you think a marketer has today that’s gonna help us adopt this new technology?

Speaker 1: Very good question. And, um, well, I wouldn’t my initial answer isn’t to do with marketers specifically, but I think it’s all a skill we can start, uh, honing. You know? So that’s prompt engineering. Like, understanding how to construct good prompts, you know, garbage in, garbage out. Um, GBT is not a mind reader. So, yeah, that that’s an interesting skill. And, um, Salesforce, um, have talked about prompts Studio prompt builder. They changed the name, but that’s gonna help admins to create prompt templates that other users can use. So, um, yeah, I just think, like, as marketers, we’re really good at portraying things and messaging, so I think that’s, uh, gonna be in our favor when writing prompts.

Speaker 0: Where do you suggest people start practicing that skill right now?

Speaker 1: Just with any, um, AI tools that they interface with. Um, so, yeah, I mean, ChatGPT is the obvious one, Bard. Um, but, you know, I, you know, have been playing around with it. Um, obviously, not putting in sensitive data per se, um, but just seeing how intelligent it is and how much it can read my mind, right, and spit out what I want. I I mean, I wrote I wrote a whole murder mystery for the sales team with characters and everything. So I was really impressed by that. Yeah. Take it. Test it. Take it to its limits.

Speaker 2: Yeah. Yeah. You really have to play around with the the tools quite a lot to figure out what what works and what doesn’t. Like, we’ve noticed things like if you put things in brackets in chat GBT, it doesn’t really like it. So there’s little little things like that, which you can as as you use the tools more, you’ll, um, you’ll get to understand.

Speaker 1: And, uh, forgive me, Lauren. I don’t wanna put you on the spot, but I think you were a skeptic for content writing.

Speaker 2: Yes. I was a bit of a skeptic. Um, I think there is some use cases, but, yeah, sometimes it’s content writing. It it can sound completely wrong, especially when you’ve got a specific, like, brand tone or tone of voice and and things like that, which is a lot of the things that we focus on at Salesforce spend. So we want to make sure that we that that stays in intact. So, yeah, at first I was a skeptic, but there I feel like there is some really good use cases now. I’ve I’ve figured them out.

Speaker 0: That’s interesting where that human element could still come in, right, to to the point of your session today. Um, there is a question in the chat from Autumn, so thanks for asking. Um, do you think AI is gonna continue to improve on itself? Currently, things like voice over can sound a little robotic.

Speaker 1: Yeah. I that’s why there needs to be human in the loop all the time because, otherwise, it will just every website would sound the same if everybody was using ChatGPT. Um, And I think Salesforce have also made a really big, um, point of saying that GPT outputs are based on your organization’s data, on your organization’s context. So, yes, I’m I’m really looking forward to seeing what Salesforce are going to, um, bring out. And the behind the scenes, they’re doing a lot of research. So, um, yeah, I mean, Lauren, do you wanna add anything about the other tools that we use internally? Any that come to mind?

Speaker 2: Um, Yeah. I mean, not not so much for content, like, for for text, like, written. But we use we use Canva, and it’s a lot of people’s favorite tool. Um, but, yeah, recently, we I’ve been playing around with the the AI tools on there for images, obviously, but, uh, some of them are really good, and I think that that that’s a really good use case for getting, like, unique images. Although, I’m still not very good at doing, like, human arms and things always look a little bit dodgy. But but yeah.

Speaker 1: Yeah. Also, to that point with, um, the typeface partnership that Salesforce have been talking about. I think it might be generally available now, but though that demo was really impressive. So if you have a chance to, uh, check that out, it is it is a paid for service, um, but, obviously, the the integration connector is free. Um, but yeah.

Speaker 0: Good. Alright. Do you think that now is the time for marketers to start looking at this? You know, obviously, they talked a lot about AI at Dreamforce. We’re hearing it in The US side in the White House this week with some of the, um, uh, I don’t know what the word is, but messaging around that and and orders going in place. I’m I’m sure there are some of us who haven’t touched the tool much. We’re like, yeah. We’re just kinda waiting to see where it goes. We hear about this. What do you think people should do today, um, to kinda take that next step?

Speaker 1: That’s a really good question. Um, I do think that we should be staying abreast of all the regulation that’s coming. We’re having similar things in Europe and the, um, and The UK, uh, because we’re no longer EU. But, yes, it staying abreast of that, and I know that Salesforce are also staying abreast of that. And, um, just do it in a safe environment, um, that you’re not, like, exposing sensitive data, like, going into the abyss, into this, like, big repository, um, that you don’t know where that’s gone. Um, I think there was some lawyer that got in trouble the other day for doing something, um, but yeah, it’s it’s these it’s like GDPR. You get these, like, horror stories, but just staying abreast of everything. Um, yeah. Mhmm.

Speaker 2: Yeah. And what yeah. Trying out new tools and and even if you’re not using them for, like, day to day work things, just play like, Lucy’s playing around with murder mysteries and things like that. Just just understanding of them, I think, is is really great.

Speaker 1: Lauren, I thought you were

Speaker 2: What?

Speaker 1: When you said me playing around, I thought you were talking about the, um, one where you can, like, clone yourself.

Speaker 2: Oh, yeah. We’re not doing that yet. But No.

Speaker 1: But yeah.

Speaker 0: It’s this murder mystery one. Sorry? Did you mention a murder mystery one? Is that what you said?

Speaker 1: Yes. Yeah. For the whole team, it spits out all the characters and the story line and yeah. So it’s it’s also about, um, I suppose, iterating on your first prompt and then iterating again. And so that really helped me, like, go forward with that, um, and just refining your prompts.

Speaker 2: Good.

Speaker 0: Good. Any other questions from our audience? Um, please pop them in. We have a couple more minutes left.

Speaker 1: Yeah. I think, um, like I was saying, this prompt builder, prompt studio, um, where admins on the Salesforce side can create these templates, I think that’s gonna be really interesting. Um, I have to check when it’s actually available. Maybe someone out there knows. And I think at this point, we’ve got New York World Tour coming up in December and, uh, then TDX. So keep an eye out. It’s going on.

Speaker 0: Good. Awesome.

Speaker 2: That’s very well. Are there

Speaker 1: any final questions?

Speaker 0: Yeah. If there’s any final questions, please feel free. Otherwise, we’ll we’ll close out here.

Speaker 1: Well, thank you everyone for attending. Um, Yeah.

Speaker 0: Great. Well, thank you, Lauren and Lucy. And that concludes today’s session. Thanks again for joining us, and a special shout out to our sponsors for their support. Without them, obviously, Marjorie would not be possible. Um, we have some great sessions coming up in a few minutes. Otherwise, head over to the agenda, check out a session to join next, and we’ll see you soon.

Speaker 1: Bye.

Speaker 2: Bye. Bye.