Andrea Hoymann  00:03

Hello everyone, and welcome to our webinar on scaling marketing with AI. I'm super excited to be sharing today with you how b2b companies can use some really new and exciting tools to grow and, most importantly, to do so in an effective and also ethical way.

Andrea Hoymann  00:27

Now before I go into the details of the webinar, I just want to spend a minute or two talking about Brand chemistry. I don't know if you all know who we are, we are a b2b marketing agency based here in Australia, mostly headquartered in Sydney but distributed across the eastern seaboard. And we've been in b2b marketing for over 15 years now and we're also a HubSpot agency partner - Platinum Partner. And in our capacity as an agency, we are always looking for new and innovative ways to help our clients grow better, but also, as an agency we're always interested in using the latest technologies to grow more effectively. And my role here at Brand chemistry is Head of Strategy and a really exciting perk or fun part of the role is that I get to play with these tools and really get hands-on and figure out how we're going to use them.

Andrea Hoymann  01:32

So that's what you need to know about us and me. Just a little bit of housekeeping. So this session is recorded, and you will all be receiving a link to the recording after this session as well. And you're more than welcome to share that as well with your colleagues. I know it's an area that a lot of people are interested in now. So feel free to share it around.

It will take about 35 to 40 minutes to go through all the content and we have an hour allocated for the session. So really, there should be plenty of time to ask questions. And you will notice that there's a little chat feature that you can use; you can either message me directly, or share more direct with everyone in the session which I would really encourage, because if you can start a bit of a conversation, I think it's more interesting than just me presenting. So there's a chat, ask questions at any point. I will answer them all at the end, or at least try to answer as many as possible at the end.

And there will also be a couple of polls popping up on the screen just to kind of get you engaged in the conversation.

Andrea Hoymann  02:42

So let's dive into it. Wow, ChatGPT. This is one of my favourite posts by a very funny influencer; I guess you could call him on LinkedIn about the launch of ChatGPT. I think, really, we can all agree that since November 2022, it seems to be everything that we can talk about on social media or on the internet in general.

Andrea Hoymann  03:07

I think what we've been seeing is really a different range of responses, right? Like people are either super enthusiastic, or they're maybe a little bit apprehensive, or really a little bit doom and gloom, like it's the end of civilisation. And that really brings me to my first poll here for the group. I really would like to see how you're feeling about AI. So you should be seeing a poll popping up now. Are you feeling excited, enthusiastic, unsure, or a little bit scared, maybe?

Andrea Hoymann  03:40

You can pop your responses in there. We're just going to leave it a couple of seconds.

Okay, just give it one more second.

I'm just gonna end it now.

Andrea Hoymann  04:05

You should be able to see the results now. So it's a bit of a split between excited, enthusiastic and a bit unsure because it's so new.

And I think that is really a good reflection of the conversations I've been having with my colleagues and friends about the topic as well. We really were all caught a little bit off the hop on the topic, right? And now we're all trying to figure out how we are actually going to use it for business, how's it going to impact the way we work and all of these things.

Andrea Hoymann  04:38

And I think one of the issues also has been that until ChatGPT really launched, most of us haven't really been able to get our hands dirty on the tool. So it was this theoretical concept, and everyone has been telling us AI is going to be this big thing that's going to disrupt business in major ways, but it was really theoretical or, as was my case, my knowledge of AI was informed by science fiction. So if I'm completely honest, before I started looking into this topic, much of my knowledge of AI was based on Westworld and Ex Machina. Really good shows and movies. But spoiler alert, it doesn't really end particularly well for the humans involved in those stories.

Andrea Hoymann  05:29

But ChatGPT definitely had me intrigued. And I wanted to figure out, back when OpenAI first launched the tool, I wanted to figure out two things. So the first one is understanding the current state of an AI and where it's headed, so really kind of understanding the technology. And secondly, how can we use it ethically and effectively for marketing?

Andrea Hoymann  05:57

Before I dive into my findings, I want to share a little bit more background of me, because it's definitely shaped the way I approached the whole thing, but also how I was feeling about AI when it first launched. So as I mentioned, I am a b2b marketer, and I have been for most of my career. And I definitely understand this constant pressure that we have as marketers to do more with less. And at the moment, in particular, the economy is a bit challenging. So I'm always interested in ways to scale our marketing more effectively, work more efficiently and get better results.

Andrea Hoymann  06:39

But at the same time, because I hadn't really had the chance to get my hands on the tools. I felt a bit sceptical about what it could actually do. Right? How good is the technology really - is it usable? The other thing you need to know about me though, is that I actually started my career as a journalist. And from that perspective, I have been watching with some trepidation the way things are heading on the internet in general with the spread of fake news and the outrage culture that we have on social media. And from that perspective, I was also a little bit fearful about the topics, just a little bit unsure, and not super optimistic about whether the positives can potentially outweigh the negatives or the downsides of the technology.

Andrea Hoymann  07:31

But as they say, knowledge is power. So I decided to really go deep on the topic and learn more about it. So let's kick things off with a definition. What does AI actually mean? So this definition is by Mike Kaput and Paul Roetzer from the Marketing Artificial Intelligence Institute based in the US - a really great resource I can really recommend for you to look into, as well, if you want to dive deeper on your own. But they're defining AI as the umbrella term for the algorithms, technologies and techniques that make machines smart and give marketers or people, in general, superhuman capabilities.

Andrea Hoymann  08:15

And I think what's really interesting here is this is about enablement. It's not about taking over or taking over altogether or replacing marketers; it's really about making us better and how we're doing things. But the question I had there is enabling how and to what extent? And for that, we really have some interesting frameworks or useful frameworks we can use to help us assess the levels of intelligent automation. So, this is, again, a framework developed by the Marketing Artificial Intelligence Institute, and it ranges from Level 0 or marketer, we all know what that is, all the way to Level 4 or machine. And just to give you some examples, So Level 1 is mostly marketer. An example of that would be, for instance, on Gmail, you know how it can complete sentences for you? You still need to do most of the work, but it can be a helpful tool. Level 2 is half and half, so it will still require quite a lot of direction from the marketer, but definitely becoming smarter.

Andrea Hoymann  09:25

Level three would be mostly machine. So that means the marketer would maybe set the objective and the framework for the initial learning, but then the machine would be able to achieve a desired outcome fairly independently over time. And the most common example of that we have currently is probably programmatic advertising, which already relies on AI quite a bit. And Level 4 is mostly machine.

Andrea Hoymann  09:55

And that actually brings me to my second poll. I would actually like to know from you where you think we are currently sitting in terms of levels of intelligent automation. What can the technology do? Do you think it's Level 1, Level 2, Level 3 or Level 4? So if you could just pop your thoughts in there, that would be great.

A couple of answers there... Just gonna give you five more seconds.

All right, I'm going to close this now.

Andrea Hoymann  10:45

So the vast majority think it's Level 2. We have a few from Level 1 and Level 3. And it's actually pretty spot on. So most of the marketing towards... intelligent automation marketing tools we have available are sitting at Level 1 and Level 2 right now. So still a way away from the total takeover that has been being talked about in the doom and gloom way.

Andrea Hoymann  11:14

So now that we know the levels of intelligent automation, let's say we wanted to look at some vendors and try to assess how good the automation is, because I think that's a really important thing too, for you as a marketer to learn because there's also a lot of hype in the space right now. So as you all know, AI is probably the darling of the venture capital world right now. So any company that has just a little bit of AI in their technology will make some maybe slightly exaggerated claims. So I think it's really important that you, as a marketer, are able to actually assess how good and intelligent a tool is before you make an investment. And so, the M2M scale is really dependent on four different variables. And the first one is information. And that is really the structured and unstructured data that you need to provide to the machine for it to be able to do its work.

Andrea Hoymann  12:14

The second variable is oversight. And that is about the constant monitoring and intervention that the machine will need to perform the work. And number three is dependence. And that's really closely related to oversight because it's actually about understanding how much of the marketer's work will need to be done and go into the machine to achieve the objective. And the fourth one, and I think it's a really important one, is improvement, because if you decide to use AI in marketing, it should really get smarter over time. So if there's no kind of self-learning involved in the technology, it might not be worth actually investing in it.

Andrea Hoymann  13:01

So when you're assessing vendors, you should really start putting some questions together that address those four variables, so you can get a good sense of how well it would help you improve your marketing. So now that you have a framework for assessing the quality of technology, I think it's good to give you just a high-level overview of the three categories of where AI can be applied in marketing.

Andrea Hoymann  13:31

The first application is around language. And that is really the ability of machines to understand and generate written and spoken language. And that's really what all the hype is, you know, with ChatGPT is all about. It's all about language and writing with AI. Here are four potential use cases for this particular category. The first one is content creation and repurposing. It's probably the area where most of us have already tried it a little bit now, with ChatGPT being open six months almost. The second one is transcription, and that one thing is really powerful. For example, we could use an AI tool to transcribe the content of the webinar now and make it really easily accessible for you afterwards in an efficient way.

Andrea Hoymann  14:23

Number three is identifying and extracting data points. And a really interesting tool to look into here is, especially if you're on HubSpot, an application called ChatSpot, which is really HubSpot's plugin. They've built an app that sits on top of ChatGPT that helps you extract and identify data points in your HubSpot interface. So it has launched as a public alpha, it's still a little bit buggy, but it's definitely a really powerful way this can be used. And the fourth potential application of the language category is social media monitoring. So, for example, you could use it to understand brand sentiment.

Andrea Hoymann  15:13

So in terms of where we're at with the technology, language is the area or the category of AI that can be applied in b2b marketing most immediately in terms of where we're at with the technology, but also the kind of work we're doing as b2b marketers, creating an incredible amount of content, and so on.

Andrea Hoymann  15:36

The second category is vision. And this is really the ability of machines to analyse and understand data from still images and videos. So one application here would be to recognise faces in photos and videos. And I think, you know, we're all familiar with that on Facebook. Actually, Facebook has been doing this for quite some time now to identify your own face in video or images that you upload. It can also be used for detecting images in photos and videos and also emotions in photos and videos. So at this point, I would say that this area of AI probably has more applications in b2c marketing. So, for example, it could be used for large-scale copyright and brand monitoring. But also, you could run some viral campaigns; we actually tried to identify faces in photos and videos and use that for viral social media campaigns. This would obviously have some pretty big privacy implications, which would need to be reviewed from a brand standpoint as well.

Andrea Hoymann  16:49

And the third and last category of AI application is prediction. It's something we all love as marketers and business leaders, the ability to predict future outcomes based on historical data. And some practical applications of these could be forecasting. So AI could be used, for example, to analyse your sales pipeline and make predictions of what you're going to close in the next quarter based on historical performance.

Andrea Hoymann  17:20

Secondly, it could help with pattern recognition. So, for example, you could look at the original source of your leads and identify which ones are converting better and which leads are converting better, and therefore you should be investing more in that channel. And thirdly, it could help with personalisation. I think that's something we've been trying, as b2b marketers, for quite some time to build lead nurture campaigns that are really relevant to people. But it's something that's really hard to do in a manual way, and AI can really help make it much more personalised, based on different factors, that we can't really programme into an automation campaign.

Andrea Hoymann  18:07

And the fourth application of prediction is recommendations. And probably the most prominent example of that would be Amazon's recommendation engine. I don't know about you, but I've definitely bought some stuff that I may not particularly need because it's been recommended by me. So it gives you an indication of how powerful that stuff can be. And actually, Amazon has built a suite of tools that allow you to build your own applications on top as well.

Andrea Hoymann  18:37

Personally, I think this area of AI is the most exciting one because it has such practical business use cases that can have a really tangible impact on business outcomes. However, the issue is that also, with a lot of b2b companies, especially in the SME space, is that you need to have a decent amount of good quality data to be able to do this kind of prediction with AI. And I think that's something where we often fall short in b2b marketing, and something you would have to work towards quite strategically to be able to allow for this kind of smart prediction.

Andrea Hoymann  19:19

Alright, so now we have a framework for assessing AI technologies. And we also have a broad understanding of its potential application. But that's quite a lot. How do you actually decide what you should be prioritising? What I would really recommend here is to work on a number of different use cases for yourself. So I would really go in with your marketing team and understand what are the areas of marketing where you are spending most of your time, and where could we gain some efficiencies and scale if we're using AI. I'm now going to take you through three really common and practical use cases. And I'll also show you some examples of extra tools that I have used as part of this test so you can get a sense. So the first use case is analytics. And the problem that we're having here is that, as marketers, we have way too much data.

Andrea Hoymann  20:22

Often, the data is also not in the most amazing shape. If we're relying on sales, for example, to fill in things manually, it's going to be inaccurate and difficult to make any meaning of. Finding those insights, it's going to be really time-consuming because you need to cross-reference and, you know, make a lot of manual things happen. And obviously, if it's larger insights, we are also slow to action. So actually being able to do anything meaningful with analytics can be quite difficult or just really too slow to be still relevant. And as I mentioned before, we are also bad at predicting behaviour, behaviour at scale, especially.

Andrea Hoymann  21:01

And how AI can help is that it can use historical data to predict future outcomes. It also can really help to find insights quickly to improve performance. And it's particularly relevant to just like day-to-day performance reporting. And it can also adapt audience targeting based on actual behaviour and help us score and nurture leads based on actual behaviour. That's another area that's quite manual these days. But again, it's really hard to kind of score leads manually, it's often guesswork.

Andrea Hoymann  21:40

An example of analytics AI in action is actually something most of you will already have access to. So this is a screenshot taken from Google Analytics. And I don't know if you're familiar, but there's actually an AI-driven insights tool available in all Google Analytics reporting now. And the screenshot here on the left shows a trend of sessions of first users that came to a website through organic search. And as you can see here, a bit of inaction in December and January. It's a quieter holiday period here in Australia, something we are expecting but obviously not want to see continue into March. And then it also tells me the impact that organic search had on our conversion path, for instance. And obviously, as a marketer, that's really good information for me to have, because it allows me to make decisions on how much I should be investing in that channel going forward.

Andrea Hoymann  22:41

So on the M2M scale that I introduced earlier; it's still a Level 1 AI tool because it does require quite a lot of input from the marketer. But it does make my life a lot easier because I don't need to go through this sometimes a bit cumbersome interface of Google Analytics to actually find the information I'm looking for. But I can actually type what I'm looking for into the tool; it was certainly the graphs; I can put in my reporting much, much quicker to insight and action.

Andrea Hoymann  23:13

Okay, the second use case is content. So, as I said already, I think it's the area we're gonna see the quickest adoption. The problem that we are facing with content is, again, we are bad predictors of content performance. So sometimes, you know, we're producing content because the stakeholder internally thinks it's really important, but it has nothing to do with what the audience is actually looking for. Content is also a really data-heavy space. It makes it really difficult to decide on a keyword focus. You're just trying to filter through a lot of data manually. You also need to check very manually how your organic search is performing and improving over time.

Andrea Hoymann  23:57

And as you all, I'm sure, know, is that search engines are extremely content-hungry, and we need to constantly create new content and improve the content we already have to actually be able to rank for the keywords that are important to our brand. And the follow on issue is that quality content is obviously difficult to scale. So you need to, up until now, usually need a team of writers, which is expensive to do as well, and especially for smaller b2b companies, not always possible.

Andrea Hoymann  24:35

AI can help us now to create content, optimise content, and also actually predict content performance. And I'm not just talking about ChatGPT actually; there are quite a lot of other tools available - paid services such as Jasper and Copy AI, that might just give you a bit more confidence also in terms of the privacy of the input data you're adding to the tool. AI tools now can also help us predict content performance in a very specific way. It can help us personalise content and actually also assist with our entire SEO strategy. And I want to talk you through a really practical example of how that can actually work in action.

Andrea Hoymann  25:24

So this is a screenshot from a tool called Market Muse, which is an AI-driven SEO research and process tool. It covers quite a lot of different things. And here you can see that I've entered information - the keyword is 'office coffee' and I wanted to see how a particular page is ranking for that keyword. And at the moment, it's not ranking at all. But you can also see in green there, it says that we should be mentioning the keyword 'office coffee' three to 10 times on the page to have a chance for ranking. And the zero in red means we're actually not mentioning it at all. So there are some really high-level insights we're getting really quickly and ways on how we could improve that page for that keyword already.

Andrea Hoymann  26:14

But it doesn't stop there. It actually also tells me how difficult it would be to rank for that particular keyword for this particular domain. And this is something that is a massive leap forward because normally, what you're getting with SEO tools is just the generalised domain authority, and it's really a difficulty ranking. But it's really hard to see how easy it will be for you to rank for it. Now, this will actually can do that based on your very own domain authority within a particular topic cluster. So much easier to get to something or to get to your tangible results with that.

Andrea Hoymann  26:54

And it does some other things as well. So it actually helps you identify questions and topics that are often asked around this keyword across the internet. So for example, we can see that ten people in Australia asked themselves the question, 'Why do office workers drink so much coffee at least once a month?' So that's a good indication that we should be starting to create some content around that. And actually, now I could go into the tool and let it help me build a brief that can be handed to a writer, or it also has an integration with ChatGPT, so you can ask ChatGPT to create the content for you from that. So it's pretty exciting. At this point, I've spent maybe five minutes to figure out what I should be writing about. So let's just put the topic right into ChatGPT. So I prompted ChatGPT to write me a blog post, 300 words long, about why office workers drink so much coffee, and asked it to optimise for the keyword 'office coffee'.

Andrea Hoymann  28:09

And the outcome here is actually pretty good. Its output was a little generic, maybe, but definitely, something that's workable. I might now spend another 20 or 30 minutes editing it, bringing in the brand voice, adding some links and things like that. But really, you can see that the process from research all the way to content creation that would normally have maybe taken me like three hours, or four hours, depending on the complexity of what you're writing about. It can be condensed to maybe an hour with a pretty good result. So that's definitely a massive efficiency getting.

Andrea Hoymann  28:46

The issue is that there's a but. So this is an example where we asked ChatGPT to write me a LinkedIn post based on one of our own case studies. Very kindly, ChatGPT says twice that we have increased leads by 200% for a particular client. That would be great - it's a great result, I would be more than happy to claim that. But the issue is it's not true. And it's actually never mentioned anywhere in the case study. So where did ChatGPT get that information?

Andrea Hoymann  29:32

And it's not the only time ChatGPT has made something up. So this is actually an example of - this is an opinion piece by a journalist called Alexander Hanff. He's a data analyst himself and a sort of reasonably public figure who publishes a lot. So he asked ChatGPT to write a brief summary about himself, like who he is as an author and his career. And ChatGPT did oblige, it wrote the summary. And it ended with the fact that Alexander had passed away. And it also included a link to an obituary in The Guardian.

Andrea Hoymann  30:11

Now, that was obviously not true. So he was alive and he was able to put his prompts into ChatGPT. But it's something that's actually a known issue with AI and language-generating AI at the moment. And it's called hallucinations. So it's actually where the AI will produce sentences or phrases that are semantically and syntactically plausible, but they're actually incorrect or nonsensical. So it is something that is a known issue. And actually, The Guardian has recently released a statement that it keeps producing fake links about the site. So they are trying to figure out as a media institution how they're going to use the technology and what it means for them. So a really interesting space to watch. But also, what I really want to stress here is that if you do want to use ChatGPT to create content - and I think it can be a really great tool - you really need to train your writers to think like journalists first and marketers second. So by no circumstances should they just copy and paste anything that's been produced by the tool and put it straight onto your blog. It needs to be checked for any copyright issues, and it needs to be fact-checked. It's really something I think we need to work on to be able to use this technology effectively.

Andrea Hoymann  31:39

And since we're on the topic of fake news, let's talk about the use case. The third and last use case to talk about is social media and AI. But don't worry; it's not going to go dark from here; I think there's actually some really good use cases for social media and AI. And so the problem you're having with social media is that social media algorithms need new content constantly. They're hungry beasts, just like search engines and creating content for multiple platforms is expensive and extremely resource intensive. It also requires a really particular skill set. Different social media channels require different kinds of content. And once again, we are really bad as humans to predict with certainty when is the best time to post and what content will perform.

Andrea Hoymann  32:31

How AI can help us here is it can actually really scale content repurposing. There are lots of different tools for really specific use cases to optimise your content for different channels. It can help us predict what content and messaging will work on different platforms. And for both of these reasons, it can greatly reduce the cost and the resources required to develop content for social media. And to give you a practical example, for that, I just want to share with you a video I've created with the help of a tool called Fliki. I've basically used a 1000-word blog post to create a one-minute video for social media. Just for context, it took me probably 45 minutes to put this together. So I'm just gonna play this now.

[On screen video plays]

Andrea Hoymann  34:31

So I think it's a pretty good result for 45 minutes of work. So it obviously could improve. We could bring in more brand; I could have spent a bit more time selecting better stock footage that works a bit better. But normally without AI, it's something that would have taken the work of a copywriter to shorten the language of the blog to something that can be used in the video. It would have taken the work of the video editor to select all the footage that we wanted to use and edit it all together and also to add the closed captions. And then potentially also, it would have needed a voice artist to create that voiceover. It would have taken a lot longer than 45 minutes or an hour and also, it would have probably been too expensive for something that you only produced to sit on social media and it's going to disappear in that world pretty quickly. So it's a really good indication of what kind of repurposing is possible here.

Andrea Hoymann  35:38

I would still say it is only Level 1 AI tool because you need the long-form blog content to actually start making it work. And you need to put a bit of effort into editing all the bits together; you can change the voice there. There are a number of different voices you can use with different accents as well. But overall, I think it's a pretty good result.

Andrea Hoymann  36:05

Next slide. The second AI social media tool I want to share with you is called Lately; again, it's a tool that helps you turn long-form content into something that can work in short form on social media. And the way it works, again, you're just dropping in the link of the long-form content, or you can copy it directly into the tool. And then, it will automatically select the sentences and paragraphs that are most likely to get engagement on social media. It does that through large-scale AI learning based on the platforms you want to post on. And then allows you to edit that content, bring it into your own voice, and the hashtags and emojis and all the bits and pieces that you do need to add to make a piece of content perform well on social media. And then, as a last step, it actually helps you schedule it in a time when it's most likely to get engagement from your particular network. And I think one thing that's interesting here is this post; it recommended me to post that at 10 pm on LinkedIn, but I would have never done that because I didn't think anyone spent as much time on LinkedIn at 10 pm. But there you go. And over time, it actually learns more about your particular network and who they are, when they are engaging and what they're engaging with. So it will get more powerful over time. So it's pretty cool stuff. But as the great sage Spider-Man says, 'With great power comes great responsibility'. I do think there's a genuine risk here that we're becoming a little bit too obsessed with what's performing well on social media and using the AI tool for that and you know, essentially using algorithms to create content for algorithms. And we all know that anything that creates outrage on social media tends to perform well. So I think there's a real risk that if you become too obsessed with just the metrics and performance, you're actually putting the reputation of your brand, but also society, at risk. I think that's where our responsibility as business leaders, as marketers comes in to set some frameworks and the expected behaviours of how we want to engage with this technology.

Andrea Hoymann  38:37

But luckily, there are actually some really practical things you can do right now; they're not super difficult. They just require a little bit of time to set up these kinds of frameworks. So the first one is to educate yourself about AI, its limitations and risks. So I really believe that, and I listened to an interesting podcast with one of the co-founders of Wired Magazine yesterday. And he said, 'You can't steer the ship if you don't use the tools'. So I think it's really important that we're going to be actively engaging with the tools, we're trying to use them and understand how they work to be able to use them effectively and ethically. Then I think you should identify your three highest-priority use cases and shortlist the number of tools that you might want to adopt. And I think the one thing that's become clear is that use cases for AI technology are very specific; they tend not to be able to do a broad range of things; you might have to select a number of different tools for specific purposes.

Andrea Hoymann  39:41

The third thing I'd like you to think about is developing an AI code of conduct for your brand and just outline what are the areas you want to use AI in. What are you comfortable with? What are we not comfortable with? That's a really good exercise to go through as a team. The fourth thing is to invest in editorial style and brand guidelines. Most of you will already have them, but they will need to be adapted to be able to work with AI. So if you're using AI, for example, for imaging and text generation, how are you going to make sure that you're not infringing any copyright? So set out some rules about how you want to go about that. You also can set some expectations in your editorial style guide, around the fact-checking procedures you want your writers to go through to make sure they're not using AI as their single source of truth.

Andrea Hoymann  40:36

And then the last thing is, depending on what areas you want to use it for, you may want to review and amend your privacy policy, as well. So if that's something you're interested in, we have actually developed a bit of a process, especially for points three and four. So feel free to drop us an email if that's something you want to discuss further.

Andrea Hoymann  41:01

So, in closing, I shared my feelings about AI in marketing from six months ago, when I first set out on this journey. So I just wanted to share how I feel now. And I would say I'm definitely 100% curious; I think it's a really exciting space; there are lots happening. The areas of application are really, really exciting and useful and interesting. And I'm also much less sceptical about the actual capability of the tools. I think they are really capable, and they can be used in a way that actually helps us rather than just, you know, have a bit of a laugh. But I'm still a little bit fearful and definitely not 100% optimistic that it's going to be used for good or that people are going to use it for good. But again, I'd like to stress that we have a responsibility as business leaders and marketers. We should get involved in the conversation and start shaping it really into a positive direction.

Andrea Hoymann  42:08

And that's the note I'd like to end things on. I'm really open to taking some questions. We will also be sharing some of the resources that have gone into producing the content for the webinar after the session, together with the recording.

Andrea Hoymann  42:28

Belinda is asking, 'What do you think about using copywriting towards creating content such as ChatGPT? Are any issues getting penalised because of bad content?'

Andrea Hoymann  42:47

The short answer is yes. I wouldn't use AI or ChatGPT to create content and copy and paste that into your blog. I think that's a really bad idea for a number of reasons. There's obviously the fact there are copyright issues, and also, sometimes, I don't know what testing you've done. You can go into a bit of a feedback loop, it can just repeat answers and can be not great. I think it's really great for a number of things like idea generation. So if you're kind of stuck on what to write about, it can really help. I think you can get create a great outline initially to get you started. So it can really make the whole process more efficient. But then ultimately, obviously, it needs to be edited for brand voice and things like that. But there are also other tools available, paid tools that actually are a bit more quality in terms of improving your writing. Actually, Grammarly is quite a good one to just help you edit. Also, Jasper AI and Copy AI are good tools to use. So yeah, definitely summary is don't copy and paste the content. But yeah, I think it's a good starting point.

Andrea Hoymann  44:15

Yeah, so there's a really good question here from Juliana. So she's interested in our thoughts on privacy concerns related to using AI, specifically when it involves sharing sensitive information for reporting purposes, for example, connecting HubSpot and their AI for reporting or providing confidential information to generate a paper or email.

Andrea Hoymann  44:38

So that's a really good question. And obviously, privacy is a really big concern. So I did actually talk to HubSpot about this just yesterday because it's a really common concern about their ChatSpot application. So the way ChatSpot works, it doesn't actually use your input information. It's not used to train AI. And that's kind of what you will need to review when you're reviewing different tools; how is your input information going to be used? So I think it's really important when you're evaluating vendors is to understand how that information is going to be used. If they're not using it to train AI. It should be okay to use. HubSpot, for example, there are no issues connecting HubSpot directly with ChatSpot. So it really depends. But it's a really, really important question to ask. And then the question to ask is, it's really about how is input information used to train AI and then, obviously, to create new information as well.

Andrea Hoymann  45:49

And Matthew is asking, 'Do you have an estimate of the percentage productivity gains a marketing department can get from using ChatGPT?' I don't know at this stage; I think it's a little bit too early. But as an agency, we are looking at it quite seriously because they're definitely massive things to be gained. In particular, we're looking at using MarketMuse across all our clients at the moment. We're assessing that as an option right now. Obviously, at the moment, we're spending quite a lot of time maintaining editorial calendars and actually making decisions on what content should be produced next. And I think there are some really strategic things the tools can help us make better decisions on and get results faster as well. So a little bit early still, but I'm really, really optimistic about them.

Andrea Hoymann  46:55

Jody is asking if there are there any existing content providers, for example, Shutterstock, implementing paid AI content generation so that marketers can gain some ownership and still support the industry? I'm actually not aware of it. So it's a really good question, but I do not know. If you do know, I'm happy to hear it, I'm still learning so much about this space as well.

Andrea Hoymann  47:26

That's all the questions, for any other questions I'll just leave you another minute. So if there are not any other questions, that's all right. So what we will do is, as I said, there's a recording of this webinar that will be shared with all of you. Please feel free to share it around. There will be a write-up, and there will also be links to the resources that have gone into creating this. Ultimately happy to hear any feedback or any other things, and if there are any other questions, feel free to email them to us. We will be writing and talking a lot more about this topic, so really keen to hear what you're interested in.

Andrea Hoymann  48:04

Thank you so much. Bye.