Sean Barry is a Senior Adviser for Field Marketing at Dell, where he manages the computer giant’s campaigns for large sales organisations and is the marketing lead for Dell’s AI and Cyber propositions. He studied at University College, Dublin, and at Hong Kong University before joining LinkedIn during the tech boom in Ireland. From here, he was appointed head of marketing for LinkedIn’s global accounts in the EMEA region.
KEY TAKEAWAYS:
David Nicholson: Thank you very much for talking to us today Sean. Let’s start with the rationale for your work: what pain points do you address?
Sean Barry: The key word is governance. That’s where I’m focused. I’m looking at how new AI technology can improve marketing efficiency, how it can make it better and faster, achieve cost savings and reduce our reliance on agencies. The UK is typically an innovation catalyst for EMEA, to test new approaches, so we’re trialling some AI applications.
On the marketing side, it’s about brainstorming what’s possible. We’ve created detailed maps of what the marketing team were doing, for example producing blogs: there’s a review process, ideation, publishing. All these activities are mapped out and we ask: how could they be improved? We’re asking: where is the marketing team going to help, post implementation of these technologies?
DN: Tell me more about the blog creation process.
SB: If it’s in B2B, we might do a thought leadership blog. This needs an expert opinion to have value to the reader. But a lot of B2B communication is functional, it’s content that summarises the outcome of the technology you’re selling. If you could input all the features and the brief into an AI , it would give you some ideas, then you’d give a draft to people to review.
We’ve taken out some of the steps and, instead of writing the blog from scratch, you act as a curator. Then after that, you’ll have a draft and compare it with what a copywriter has written. You can conduct a test: run three different articles and observe which is best. We won’t run material that is solely AI-based. It has to be curated. Blogs are relatively simple to create using AI, whereas images are more complex, because how can you tell where they’re from?
DN: What have you found to be the best use of AI?
SB: Scaling the core marketing message at a tactical level. What I’m seeing in marketing is where individual influencers have channels as thought leaders, it’s important that these are in their own voice. The content should be completely written by them. But where you have established core brand guidelines, there is an opportunity to scale that with AI. The closer you get to a brand , there’s a huge amount of human input. The AI can then help to bring that message to many channel formats.
DN: How have you found the tests of AI-generated content versus human?
SB: There’s nothing conclusive yet, we need to do more AB testing. But our initial results show that AI is surprisingly effective, particularly in generating engaging titles and social media content. This is central now, it’s something that’s influencing everything we’re doing. It’s really a question of when, rather than if AI becomes a central part of the marketing toolkit. It’s taking a lot of attention within the organisation, because we recognize that marketing function is not the same as it will be.
To implement this, it needs to be top down for governance and structure, but also bottom up feedback and testing from marketing teams using and experiencing the output of the AI content produced. If it’s only top down instruction, there will be problems.
DN: Why doesn’t it work that way round?
SB: You might have one department using the software, because they’re experienced with it, and then every department might be asked to adopt it, but there are difficulties because of the nuances in technology, so it may not be suitable for everyone. You need regional tests – feedback is essential.
Each company will need to trial several different AI technologies, implementing with caution and governance.
DN: Is there an issue of AI-generated content being discredited as inauthentic?
SB: You’re 100 percent correct that it has to be authentic. The curatorial role can’t go away. It can reduce the administration, but staying true to the message is everything. Most organisations are rightly cautious on how to approach AI. It’s about productivity gain and new ideas rather than a new voice. For example, where AI can create new influencers – AI can in itself be an influencer. Maybe the next question is – where a person can converse with AI, they no longer go to a blog, but straight to ChatGPT or Perplexity. Then the AI will go off and read every source it can find and then pull everything into one answer. “Here are the top five cyber security products for enterprise” in one answer.
As a marketer, we can’t control what AI will come up with, but we can control what information is on our social media and website. So as long as the information is up there, it can be discovered by AI. Both things are happening: human conversation and followers on social media.
When you click on the source link from an AI answer for ‘best solution to x’, for example on ‘Dell’, rather than give you the home page, The AI often goes to product specific pages with technical details on the product or solution. That’s big because it removes some of the established steps in digital journeys a B2B buyer may have followed before. AI has read the product specs, so if you haven’t got a pdf on your website that gives specific detail, it won’t consider that in its recommendation to the buyer asking the AI for recommendations.
DN: So is the point of having information online to appeal to computers now, rather than humans?
SB: Each of these AI programmes are black boxes at the moment, we don’t know how they’re going to change over time. So it’s important to make sure we’re ready. More specific detail is helpful, even if it’s somewhere on a website where humans are looking. Rather than reverse engineering the AI algorithm, ensuring that the most useful and informative content is available on your channels seems like a good direction of focus.
DN: Is this the mystique of AI?
SB: I’m trying to second guess the buyer journey. What a prospective buyer will think and do before buying. The tactics we’ve used before still exist, whereas the AI piece is one that we’re observing, but not something we can control.
Google started as key words, and then moved to edge ranking and weighted content scores. Where Google could deliver a response to key words, it’s the same thing with AI, but with much deeper context.
DN: When Google arrived it was great for small businesses to find new customers, putting them almost on an equal playing field with large corporations. Can we say the same about AI – does it work for the little guy?
SB: Under earlier conditions, large companies would have bigger B2B engines and budgets, making them very powerful against small businesses. Now, if you give AI a brief to solve a specific problem – say “ list companies that solve x problem and tell me why I should buy this” – the hope is that AI is going to give you an equitable answer, where there shouldn’t be a bias between big and small businesses.
I would say the advantages for small businesses are similar to the arrival of Google. AI is similarly accessible, because it’s easy to set up a website, and with AI you can read the whole internet. So the question of volume and quality of your marcomms, depending on AI, helps to weigh that. I don’t see that as a negative. If you’re credibly offering a service, an effective AI should find you.
DN: I’ve noticed that marketers recommend sending out mass press releases to hundreds of websites, giving the impression that a company has high numbers of media mentions. Presumably this is so that AI registers these and accredits them. Is that right?
SB: I don’t know how impactful this is. AI may use frequency of media mentions, but then there’s the question of whether people pay for it. Can AI tell the difference between advertorial versus earned media? It’s a question of how AI calculates this.
If it’s earned media, when something is picked up by a journalist because they’re interested in it, the credibility of that content is higher because it has less bias. As AI becomes more proficient it should focus on quality over volume of mentions. People are using AI accounts to ask these questions.
DN: Sometimes it feels like AI just encourages spam, with indiscriminate amounts of information pushed out, hoping that AI will find it.
SB: In the past, companies sprayed out information and hoped that enough people would see it. Today, buyers are asking questions and coming into the ecosystem. So rather than millions of people seeing an ad, the person who has the problem to solve asks the question – that’s assuming the buyer journey moves on to AI – and finds what they’re looking for – a company list with features and business outcomes.
DN: Where do you see this application of AI heading?
SB: There’s lots of fluidity, but ultimately organisations are looking at ways to improve their productivity and efficiency across all business functions. In a two-to-five year period, there will be a much more structured approach in how organisations market. This will be driven not so much by what companies think, but how the buyer journey changes. It’s up to companies how they adapt to these changing buyer patterns. For example, the level to which they depend on AI and the brief they give to AI.
DN: What other aspects of marketing do you think will change?
SB: I think we will see more automated procurement in the buying process. Can AI automate everything for a small company, or any size of company? Can it be a negotiator, or strategist?
DN: How is Dell helping customers with AI automation – having been a pioneer of advanced technology for many years?
SB: We’ve had an incredible few years. We’re very well positioned to advise customers on their AI journey. We’re working with all the large corporations: we have over 100,000 employees and revenues above $100 billion. The reach of the organisation means we have a very good view of what’s happening in AI and what’s coming in the future. We have a ton of case studies of how Dell is using AI with customers. To give just one example: a beer company is using AI to observe where bottles fall over in its production process. This is dramatically improving its efficiency. There are hundreds if not thousands of examples like this, or using AI at the edge.
The scale and reach of Dell is something that people don’t always recognize. We acquired EMC [Dell paid $74 billion for the company in 2016] which meant we had a data centre sector, a much broader offering. We’re working with Intel and Cambridge University on Project Dawn, building the UK’s fastest supercomputer. This project could improve research and bring profound scientific breakthroughs in the near future. It is incredibly exciting.
