<img height="1" width="1" style="display:none;" alt="" src="https://px.ads.linkedin.com/collect/?pid=4374522&amp;fmt=gif">
Skip to content

Spotlight: Beyond ChatGPT – Practical use and innovation in marketing with GenAI

On September 28, 2023, the World Federation of Advertisers (WFA) hosted a webinar by Intentful – “Spotlight: Beyond ChatGPT - Practical use and innovation in marketing with GenAI.”

Watch the full video recording of the session:



Learn how GenAI is being rolled out in marketing across the globe, the use cases beyond ChatGPT, the pivotal moment for e-commerce with AI, social commerce and the search evolution, and other paradigm shifts caused by Generative AI.

The webinar transcript: 

Alice: Hello and welcome. It feels like it's been a while since I've been hosting a webinar. I hope everyone's had a lovely summer. I can't believe it's the end of September already. But thank you so much for taking time out of your day and joining us for our webinar called Beyond Chat GPT. Practical use and innovation in marketing, e-commerce and search with Gen A. Hopefully all of you here today have some idea of who the WАA is. But it might not be so familiar as this is an open webinar. So this is just a quick overview. What you see on the screen is us.

Intentful_WFA webinar

We are a not-for-profit membership organization. We've been around since 1953 and since then we've been representing the interests of our members who you can see on the screen and it's great to see a number of you connected today. So welcome to you all.

In the world of marketing, generative AI has taken centre stage in the past 12 months or so. It's ignited conversations. It sparked curiosity about its potential. But it's around that's often viewed with a mix of all, intrigue and sometimes caution, presenting both exciting opportunities and history.

So I thought it would be a great idea to invite Marina Petrova, CEO and Co-Founder, and Bruce Amick, COO and Co-Founder of Intentful to speak with us today, as not only do they have a wealth of knowledge in this space, they're a leading voice in the industry and always fascinating to speak to. So without further ado, I'd like to hand it over to Marina and Bruce. Welcome. Thank you for your time today and over to you.

Marina: Thank you. Thanks for inviting us.

So today we want to talk about the practical use and innovation and everything that's been happening in the world of Generative AI in the last several months. It is the second time that we are joining a WFA session. Our first webinar was in December of 2022 when the world didn't even know much about AI – ChatGPT was already launched, but it was still new. And back then we wanted to share all of the opportunities that AI was bringing, especially to marketing professionals and marketing procurement and media.

And today, we are taking that one step further, and we would like to share a couple of things. So first, where AI adoption is right now with CPG and other companies. The second one is the practical use cases, what you can do, and what goes really way beyond ChatGPT. And then what's next for e-commerce and search.

As Intentful – and that's the only time when I will mention Intentful – started in 2021. And since the minute we touched AI, with our first use case was bringing together data and AI, it was so obvious and so clear that it is a game changer, and it's going to change absolutely everything in how we work.

The ways of working, the results, the approaches, the efficiency and the effectiveness. And since then, we have been telling the world how great it is. Thanks again for inviting us to share this information with you today.

So three key topics today are Marketing, E-commerce, and Search.

I’ll start with one of the key things that we think is super important right now. And that's probably one of the key facts that I would like you to know when you leave after this webinar it is now possible to have the AI, the generative AI model that knows your brand and knows your business, and that there is so much more than you can do with AI now, and that all those capabilities go beyond ChatGPT.

ChatGPT is what everyone keeps talking about all the time that's why we're referencing it. And they are fantastic and they are amazing. ChatGPT is a GPT model that is trained to have a conversation, but as all of you know, it does not know your brand. It does not know your style. It does not know any of the proprietary information.

And yes, there are ways in which you can showcase context, but it doesn't yet know your brand. 

What's possible now, with the way technologies have advanced, is that in just a few clicks, you can build on top of an AI engine and teach it to understand any of your information. There are multiple ways you can teach, or better said, allow the AI to understand your brand information, business details, data insights, your voice and style, your top-performing campaigns or any other data you'd want a highly capable employee to know. 

The technology now makes it possible to build on top of the core AI model, making it extremely quick and easy. 

Let me show you what I mean. Alice was very kind to us, as she always is, and she asked if we could possibly use some content from the WFA website and teach the AI to understand that content. We chose Project Sprint, a project familiar to many of you. In those three clicks, as I mentioned earlier, we uploaded this information to the AI engine and trained the AI with everything it needed to know about Project Sprint. And I've recorded a little video, but we can also, if you prefer, go through it live. We refer to this as the WFA GPT, because that's what it is. 

The first thing we ask the AI is to tell us about Project Sprint; assuming I don't have the time to read 25 or 40 pages, I can quickly get a summary. I use this as an example, but think of it as information from your brand that you can feed to the AI for it to understand. 


And this information doesn't come from ChatGPT from the open internet.

It's your information. Your brand. Your style. In this case, we didn't add a style, it was just a single document that was uploaded.  There's another video where you can see what else you can do with it. It's not just about understanding what it's about.


You can apply to so many use cases. In this particular case, I'm asking it to extract information from the report and tell me what the key role of marketing procurement is. In natural language, I can alter the tone of the response, I can change the language, making it multilingual. Even if the original information is in English, I can pose a question in Spanish or Ukrainian or any other language. And I will receive the answer back in the language of my choice. 

So it did provide me with the answer to my question based on the report.

Keep thinking of it as your brand, when you have all the information that you want AI to comprehend and seamlessly locate that information, and, most importantly, do so much more with that data. In this specific instance, I'm asking it to create a summary of the key learnings from the report. I wanted it organized as a list; I could even request it as a poem, and it would do so. 

Here, I'm asking for a list. So, the AI condenses information from the 25 to 35 pages of the report into seven key points, all in natural language.

But it's not just about summarization, in the next request that I give to the AI, I'm asking it to transform all the gleaned information into a LinkedIn post.

When I thought about it further, I realised it's too simplistic to just convert this into a LinkedIn post. I instead asked it to create a content plan for two weeks, and also add hashtags and mentions.

We're talking about using WFA as an example here, but let's conceptualize this for your brand. So, you have an AI that understands your brand, that's aware of what you want it to know, and then you can ask it to conduct any task. You can create a social media plan, or craft e-commerce content, or generate search content, or anything else, really. You can even integrate data insights that you already have and add them on top of the engine, but we'll discuss that in a separate session.

So, in just a few seconds, I have a LinkedIn post ready. If I wanted to add a particular voice and style, I can do so by simply uploading examples, and they can be in multiple voices and styles, and in a variety of languages.

It still needs to be reviewed by a human, obviously. But as we all already know how much time AI saves, in September 2023, you can now have an AI that is familiar with both your brand and business.

Here, I'm asking it to combine post number three and post number seven into a single email addressed to WFA members.

And if it was a bigger AI model, it would also know the type of WFA members, and we could teach it what the style should be and what type of communication and the lens of the response and the format of the emails.

In this particular case, I'm saying to add a call to action to download the report. Alice, you might need to change that email a little bit, because again, it is speaking in the voice of the report in this case because that's what we gave to it initially.

Another use case here. I want more information from the report, so I'm asking which companies have been mentioned in the project’s report and look what the response is.

AI has the issue of making things up, but in this particular case, it tells us it doesn't know. It says that this information is not included and it does not make up names or companies because it is trained on a particular set of information and it will not be making things up. So it suggests that I go and ask WFA.

Another use case. Here, I'm asking it if this report was published online, what are the keywords that would rank on in search? And there is a smart answer here that offers potential keywords. So it does understand that this is not a given that it would rank on these keywords, but it does extract key points that would most likely rank. 

So hopefully, this gives you an idea of what you can build now. It’s so much more than just the standard functionality of ChatGPT that does not know anything about the brand. And you can do it in a super simple way by simply uploading and having your own AI do whatever you want for you. 

Of course, the quality of information that is being uploaded is super important, but everyone has a lot of content that can be used as a basis. So when AI knows your brand, think about how many use cases there are. There are thousands, and we are not going to go through all those thousands of use cases, but I would just like to briefly talk about the key ones.

One of my favorite ones, and actually how we started the company, is the idea of not simply using AI for content creation. You can make it data-driven and super relevant to audience interests. Given how many touchpoints there are now, given how customer engagement is changing, and the need to interact with people at multiple touchpoints and in the right context, it is super important to be able to create that content at scale.

So the use case that we have here is data-driven content creation. All companies now have a lot of data about their customers - what they say, what they do, and what their interests are - obviously, if they have provided you with their consent, that goes without saying. And the breakthrough is what happens when you blend that data with generative AI.


In this example, we're using a supermarket, where their communication is primarily built around discounts. Of course, this is just an example. You can choose the audience interest, the intent, the lifestyle, etc. This example is for a company in Southeast Asia, where Viber is frequently used. We're demonstrating that you can apply the brand voice, the knowledge, the promotions, and even do it multilingually. 

In seconds, you can generate data insights about audience interests, where they are in their customer engagement journey, what we want them to do, and quickly create content. This content can then be repurposed into other channels. 

Although we're discussing Facebook Messenger and Viber here, this approach can be applied to any type of content that needs to be created. It's incredibly simple and efficient. You get on-brand content because the AI knows your brand. This process is 90% faster and can be multilingual. This approach can be utilized across multiple channels, combining your data with an AI that knows your business and brand. 

Another common use case is in search. Instead of just asking ChatGPT to create a long-form article about a particular topic. You can incorporate all the data insights you have, for example, for search. You can use target keywords or the intent, which is increasingly important now. You can include your brand voice and style, or even multiple voices and styles in multiple languages, as well as facts. 

One significant advancement is the ability to fact-check, if you have your own AI model, which can cross-reference information it knows about your business. This minimizes the 'hallucinations' that large language models are notorious for. Although it still requires human review, it's much more advanced than it was even a few months ago. 

In our previous WFA webinar with Vice President Bob Hoffman and a client, Broadway Inbound, we discussed how we started using AI to create content for Broadway shows nearly two years ago. The client has seen consistent growth. Although some of this growth is natural, much of it is due to the data and AI approach. We were able to create relevant content for the audience of 15 million theater-goers searching for information each month. This steady growth has continued for almost two years now.

Search is the most obvious thing to get started with as soon as realistically possible. The other is e-commerce, and I'll talk a little bit later about how e-commerce is changing because of AI. But for now, talking about the use cases, the approach is similar, and relevant to many companies at the moment.

Similarly, you bring together multiple data insights that you have. It's different for every company. For some, it is digital shelf information, for others, it is also conversion data or intent data. So you get those e-commerce data insights, then on top of that, you add everyitng we discussed previously, the brand guidelines, facts, and style, but then there are also retailer requirements.

Each retailer is different, and using the same approach doesn’t work if you want to sell with a particular retailer because they all have their specific approaches, guidelines, their own search algorithms. So the content needs to be optimized for each e-commerce platform specifically. Doing this manually is often painful, expensive, and slow.

When you use AI for all of this, the time savings and the efficiency are unbelievable. You use it for fact-checking, the resulting content can be multilingual, and you can do this for multiple brands. Of course, there is a change in the amount of actual real hours put into this work. 

You still need quality assurance (QA) involved in the process, but it becomes 90% faster and the resulting content is on brand. Content created in this way is much more relevant and accurate than just opening ChatGPT and saying, ChatGPT create the PDP for Amazon. When you bring in those data insights, brand guidelines and a tone of voice then the efficiency is so much higher.

Another use case is internal documentation. I love all of the use cases, but every time I start talking about a new use case, I usually say, I love this one so much. I love this use case because all of us, all businesses, all companies have a lot of information that lives somewhere on the server and not everyone really knows what that information is. There are many existing presentations and long files that not everyone reads anymore because we don't have time to search for information. 

Similar to what I was showing with WFA projects, you can teach AI to know core guidelines across different departments. It could be used for HR, media, marketing, procurement, legal, or documents. This way you can quickly find information and then interact with it, reuse it, ask questions, get key insights, and so much more.

Customer support is another area where AI can be used. Chatbots have existed for a very long time, but today they are changing. You no longer need to have predefined responses. That is still possible, and should be the case in some situations, but you can also use the natural language based knowledge base. This way content can be turned into virtual assistants, you can build help center articles, tutorials and more.

The example from the next video that I would like to show to you is not a client that we work with. I'd like to make it clear that it was just a demo that we built in under several minutes. It is for Hill’s Pet Food. We took about 10 pages from their website, and we built a knowledge base. Disregard the interface, it's just a demo to show the functionality.

Let’s imagine that there is a user is asking in a natural language, “my cat has a sensitive stomach, what would you recommend?”


The old reality of customer support required predefined answers, and if there wasn't an answer to a question, it would respond with something like, "I don't know," or transition you to a live assistant.

With this AI, you have core information about the product, the brand, and anything else you'd like the AI to know, and the communication can occur in natural language.

Furthermore, the AI's responses can mimic the style of the user and adapt to the style of the conversation. You can also adjust it, alter the length of the response, the type of response, and even add images and links.

The examples we're using here are just to show that all users are different, and we all have different questions.

The other thing I love about this is that it's an additional way to gather data insights about what your customers or audience are interested in, and what their pain points are. Then we take that information and add it to your other pool of data insights, because realistically, most content that's produced is based on assumptions. Most of the information we provide on the website is marketing-oriented, based on what we want people to know, but we don't always pay attention to what they want to ask. This is just another way to gather that additional information.

Customer loyalty is another use case that you can start using right now, where you combine the data you have in the CDP or CRM, or any other customer engagement platforms, with any data insights you have, along with the knowledge of your brand, guidelines, facts, style, and the goals you have for a particular campaign.

You build all of this on top of the engine, and it's quite straightforward. It's not a project that would take six months to implement and cost hundreds of thousands of dollars, absolutely not.

In reality, it's just a few clicks for teaching the AI to know your business, and adding other documents, files, and data insights might take a couple more clicks, but it's a matter of days, not months.

That was on customer loyalty, one of the use cases. But what's next? How is technology evolving and how are companies adopting AI? Where is everyone in their customer journey?

We'd like to share a few thoughts that might inspire some new ideas. Firstly, we discussed how people interact with numerous touchpoints. It's always been a challenge to cut through the noise.  While there is increased personalization and customer engagement happening on an individual level, it still needs to happen at scale for us to be able to see results.

Conversational AI is not a new concept, but it's ushering in a new era of brand engagement. As humans, we navigate the world through search and conversations. That's what we always do. Therefore, the first transformation we're seeing is that search is becoming conversational; it's no longer static, and it's crucial to ensure your content is discoverable in conversational search.


It's not only about Google, but also e-commerce. Retailers have been losing a lot of money due to abandoned carts because of limited search functionality. If you sell online, you're likely familiar with the challenge of ensuring your product is discoverable on the retailer's website. It's a difficult task for retailers to provide that search algorithm because we, as humans, do not always clearly communicate what exactly we are searching for. We might say one word, but we are thinking of so many other things. This miscommunication often leads to cart abandonment. AI is changing that.

Imagine a customer who can find a product on a retailer's website without having to search through hundreds of pages, but instead using that same conversational search approach we were talking about. In natural language, the customer asks for what they want. It's not a traditional chatbot. It is a natural conversation where the AI can help connect your product with what the customer is searching for. This drives engagement, makes it easy to find information, and leads to more conversions.

Before diving into AI adoption, let's discuss e-commerce more. It's crucial to ensure that the content you have for each retailer is optimized for that retailer's platform. Most retailers are continually working on updating their search capabilities. AI advancements will take this discovery process to a new level. So, it's essential to ensure that your product detail page (PDP) content is optimized for the retailer, and with AI, you can do this quickly, easily, and increase your conversions.

As you know, ChatGPT announced yesterday that they are now connected to the live internet, so they can work as a search engine. This changes not just how Google works or how ChatGPT works; it changes how our customers interact with information. Customers will expect to have similar interactions, regardless of the touchpoints where they are. It could be a brand website, a retailer website, an app, or a support conversation. The landscape is changing rapidly. For brands and marketers, it's important to ensure you're not just providing marketing content, but content that aligns with audience interests and insights. AI now makes this possible.

In terms of AI Adoption, it's fascinating to see how companies are adapting. While I can't mention them all due to confidentiality reasons, I can tell you that most Fortune 500 companies are currently in the discovery and pilot phase. Considering how new and disruptive this technology is and how it requires new skills, attitudes, and processes, this adoption is happening incredibly fast.

So there are a few things happening. Firstly, this is where I can mention the company. Innovation Day workshops are a fantastic concept that we implemented for our friends at Carlsberg, where there was an idea of gathering over 100 marketers, around 150, from all around the world. The goal was to learn about what AI is, how it functions, how it determines what content to generate, what the potential applications are, and what its capabilities are. 

They also got the chance to interact with AI and actually experiment with it. This cultivates an understanding of the use cases. We can discuss the potential use cases, but the real use cases come from your teams because these cases need to address a problem. 

Carlsberg has always been an innovative company, and the idea for this event came from procurement. It was Lenka Laurinova’s (Note: Carlsberg’s Senior Global Procurement Lead) suggestion. I'm not sure if she's present in today's webinar, but thought I’d mention she was the one who proposed it.

The innovation day workshops with AI seem like something that will be continuously happening due to the constant changes in technology and how rapidly these changes occur. There's literally something new happening every day in this field. So, an innovation day workshop is an excellent way to start the process of adopting AI.

Many companies have already started using AI for multi-channel content creation. We do a lot of work with the travel industry, and Destination Marketing Organizations (DMOs) are using our AI to generate content for their online presence. The use cases are numerous and can include content creation, search, e-commerce, and so on, specific to that particular industry. This is being adopted rapidly, and it's progressing extremely fast. 

DMOs can connect with their travelers or visitors through content. As I mentioned, the discovery and Proof of Concept (POC) are happening across multiple companies, and the use cases are very diverse.

For some, it's about internal documentation and how to utilize the vast amount of information that was built to avoid losing that information. The goal is to simplify the time it takes to find the information and to generate new content. For others, it's all types of marketing materials. For some, it's about connecting data insights with content creation. For others, it's about using AI to find insights, read reports, or summarize information. For others, it's e-commerce and finding new ways to do it. 

We receive new use case requests almost daily, which is absolutely amazing. 

My final point is to start small but start now. If you haven't started yet, start today because AI is changing everything. You can make it super easy, and you don't need to change your existing processes. You can just start integrating it into what you have.

Alice: Thank you, Marina. That was excellent. I think everyone is still processing all the information. We don't have any questions yet, but please, everyone, ask Marina and Bruce any AI-related questions. This is a perfect opportunity.

You can post them in the chat or in the Q&A. But I'll start with a few questions. You talked a lot in the presentation about multilingualism. This is a question that our members often ask about adaptation and transcreation. So, what are the multilingual capabilities of AI?

Bruce: Six months ago, we were saying it's getting close. It's maybe 80% there and should be reviewed by not just a native speaker, but also by a subject matter expert to ensure accuracy. But now, in the last two to three months, the improvements have been so dramatic. A lot of this is due to advancements in the large language model side of the AI.

And a lot of this comes from our learnings because we have some clients for whom we do long-form translation content. So it is now better than 85% across dozens and dozens of languages. And it's also almost instantaneously usable.

We still believe that you should have someone in the mix to look at things, even if it's just in the initial setup and development stage. But the process has become so much better and faster. Just test the languages available. There are about 80 now. 

All of the most common languages are available and they are almost flawless. 

And that’s music to many people's ears.

Alice: So we have some questions coming in. The first question that we've got is from Gretchen, who asks, who in an organization uses AI, in your opinion? Is it a general business tool that marketers are using? Or do you find it's mainly IT that are using the tool?

Marina: Thank you for the question. Really, AI can be used by anyone in the company. And I think the future is that we are going to have AI as an everyday assistant, no matter what the department is. This is something that is already used by marketers today. Absolutely. As well as IT and other departments. 

Bruce: I think a lot of operational departments are beginning to discover the benefits of AI. We've had conversations with various organizations about implementing knowledge-based solutions for contract review, to make sure that certain clauses are in there, and for comparison purposes. For SOW or RFP, either generation or review. So there's a number of things that are opening up now.

And a year ago, we were talking almost exclusively to marketers. Now we're talking to different areas of the organization. The adoption is starting to spread. The virus is spreading.

Alice: Cool, thank you. So we have another question from Sasha. And they're asking, how do I start if I want to create a ChatGPT model for a specific purpose? Similar to the example of WFA that you used, how do they get started on that?

Marina: This is not a sales presentation. So I will not be promoting our services, but my response would be if we work together, we would help you with creating the knowledge base or explaining what needs to go in the knowledge base. And then it's just a one-click upload. The second click is to test it, and the third one is to start using it. So the key part is building that information that you want AI to know. 

Bruce: The toughest part in getting started is to identify your sources of information. On the marketing side, you might assume it’s easy, but we have discovered it's not necessarily simple. Most of the content that Marina mentioned earlier is marketing content, and not necessarily informational content. Where the identification happens really quickly and we're able to do this one, two, three, click process seems to be more in the operational aspects of an organization where you have policies to change your documents or you have playbooks.

So we're evolving from our relationships that we've developed with marketing divisions. Marketing divisions have their playbooks, have their guidebooks, have their approach to creaitng content for Google ads, Expedia, or other platforms. And these are documents that are hundreds and hundreds of pages long, or maybe spread out across different documents. They're potentially in the process of being updated and it becomes really difficult for other team members or stakeholders to access that information and unlock. AI does it very quickly in a kind of democratic way where everyone has equal access to that information within the organization. It makes it easier for people to make use of the information.

Alice: An anonymous person asked, with so many things happening so quickly, what are your go-to sources to get the best examples of the latest use cases for AI?

Marina: It's a great question. You can use AI to come up with use cases. But, the use cases come from within the organization. You have your own existing processes, you know where the time is spent, you know where you could use some operational efficiency. The question to ask is can the AI be used to be applied in that situation, can something be automated with the help of AI. We have a couple of articles about use cases on our website, and on our blog, but it is so individual and needs to be adjusted for individual companies.

Bruce: Would it be in bad form for us to put a link to our 100 use cases, Alice?

Alice: No, no, please share it. Anything you think will be helpful, I'm sure that'll be greatly appreciated. 

Alice: Well, we have another question that's come through from Kevin. Hello, Kevin. What would you say are the biggest watchouts or pitfalls for those who are just getting started or planning to get started? What do people need to be on the lookout for?

Marina: Another great question. So I think it is always important to start by understanding the basics of how AI works. No need to know what a neural network is or how predictability works there. But it is important to understand the core basic principles because that helps understand what's possible and what's not possible. So that's number one. 

Number two is making sure that AI knows your brand and your business. If you are using a generic open model, it will be making things up because of how large language models work.

In just 30 seconds, a large language model predicts the next best word based on what it was trained on. So if that information is not in its original knowledge set, then it would continue finding the next best word. And that’s what makes it sound so convincing. Whereas, in fact, it is made up. That's why it is so important when you do use AI. Make sure you use AI that knows your business or your brand or whatever you wanted to know.

And the third point is, of course, data privacy. Use your own secure API environment to make sure that this information does not go anywhere else. If you are using a free version of ChatGPT, there is a tick box that you can use for the privacy settings to make sure that your information does not go anywhere. It will still live there for 30 days, if I remember correctly.

But again, our recommendation is having your own AI that knows your business and that is also secure.

Alice: We don't have any other questions coming in, but leading on from Kevin's question, I'm curious about biases that exist in AI. I was at a talk a couple of weeks ago and they were talking about the biases that exist and the risks that are associated with that. They asked AI "Can you give me a picture of a doctor?" and every single time it was a white man, which just doesn't reflect society. So what are the watchouts in terms of biases and is AI improving in terms of bias? I'd love to know.

Marina: Thank you so much for this question. First of all, the reason bias exists is because of how large language models or how generative AI works. It was trained on a huge amount of data some of which goes back centuries, when life was very different. That’s why a white man is shown when asked about a doctor. AI’s knowledge comes from many years ago, that’s why biases exist. 

In order to make changes, multiple initiatives are underway. OpenAI and Google are doing commendable work. We also have our own initiative called AI in DI, where we identify biases in content during its development process. We have created a list of different biases that exist. The process has been a mind-blowing educational experience. 

The idea is to flag biased content, but not necessarily change it. It's not our place to decide what is good or bad content. However, we make sure the end user is aware of the bias, so they can pay attention to it. 

I believe we’ll see more and more content providers, like OpenAI, finding ways to prevent biased information from entering content creation. They have filters for this and have been working on it for some time. However, it's still an area that needs attention.

One more thing I would add is the reason why we focus on creating brand-specific knowledge bases. We instruct our AI to refer to these knowledge bases for content creation. This won't entirely prevent bias, but it will stop bias from the general training data being referenced, unless we specifically ask the AI to do so.

Bruce: We focus on text and not images due to the various legal and ethical issues surrounding AI image content creation. My cautionary advice is not to use AI image content for production due to concerns around intellectual property, usage rights, and ownership. 

There are also consumer concerns regarding this. Initially, it was a marvel, but now there's a sense of unease. From a brand reputation perspective, it's essential to be wary. AI is great for ideation and storyboarding and there are companies like Adobe developing fantastic AI tools within their software. However, relying solely on the generative capabilities of AI models for images is risky at this point. The advances are amazing, but the regulatory, legal, and ethical issues are still contentious. There are definitely some things to look out for. 

Marina: David has asked an interesting question about the brands I admire for their use of AI. There's a brand I admire for its innovative spirit and approach, but I can't mention it yet as their initiatives are not live. They are not using AI for the sake of PR, but to achieve tangible business results. Ask me the same question in three months, and I'll be able to share more.

Bruce: I agree with you. The brands we admire aren't ready to announce their AI initiatives yet.

As Marina mentioned in her presentation, many brands are still in the discovery phase and keeping their cards close to their chests.It's clear that some brands are more advanced in their use of AI than others. Certainly, there are businesses and companies that are way ahead because they started around January or February of last year. They asked the right questions: What do we do? Who is responsible for this? Who is driving this? What are the use cases? What can we do? On the other hand, there are many who are just now realizing that it's time to start looking into this.

I anticipate we'll see a lot of announcements from those companies deciding to disclose their use of AI, probably in the next three to five months. There are numerous pilots taking place right now.

Alice: Seeing as we have about five minutes left, if there are no more questions, I'd like to ask one. Following up on David's question, what excites you most about the future of AI? Is there a specific use case or something else that really sparks your excitement in this space?

Marina: That's a great question. Since the first day I began working with AI-generated content, I've remained excited. I knew from the beginning that it's not just about creating content for content's sake, but about what you can build on top of it. It's revolutionizing how we work and simplifying so many things that we, as humans, can do so much more. The opportunities are amazing. It allows us to focus on more creative, strategic, and fun tasks, while AI handles the boring and repetitive work.

I'm also excited by the advances in technology. I'm not one of those people who would post a picture of a cyborg with Arnold Schwarzenegger and say, we're all going to die. That's not me. I see opportunities and admire how companies like OpenAI are implementing different safety measures to ensure we don't end up in a Terminator scenario. The most exciting thing for me is the advancement in technologies and what it can do for us as people.

Bruce: For me, my key excitement for over a year and a half now has been knowledge access and sharing, the ability to summarize, the ability to work smarter. I find the improvements for knowledge workers across all sectors and areas of business tremendously exciting.

We used to discuss the "pain point" of knowledge transfer. When an employee who has learned everything is retiring, moving to another company or division, you have to train someone else with that institutional knowledge. This has been a huge pain point over decades.

I see the liberation of institutional information, policies, procedures, guidelines, regulations, and more. How do you unlock and make this information available across teams, employees, and companies with affiliates or partners? I see this as one of the most exciting developments.

And there we have it.

Alice: Marina, Bruce, thank you so much for your time today. It's been a pleasure having you. Thank you everyone for joining Bruce and Marina. Thanks again.

Marina: Thank you for having us.

Alice: Our pleasure. I hope everyone has a lovely rest of their day.