Did you know that 87% of marketers are already using or experimenting with AI? AI-powered tools are transforming everything from personalisation to predictive analytics, making it essential for marketers to stay ahead. In this blog, we break down key insights from our webinar, “The Future of Digital Marketing: Using AI and Analytics to Stay Ahead.”

From redefining strategies and enhancing personalisation to optimising campaigns, AI and analytics are revolutionising the digital marketing landscape. Our recent webinar brought together industry experts Simon Bullmore and Claire du Preez from the LSE Digital Marketing Strategy & Analytics Career Accelerator to discuss the most pressing topics: how AI is transforming marketing, how professionals can harness its potential, and what ethical considerations must be addressed. You can watch the full event below:

Key takeaways from the webinar

The marketing industry is evolving rapidly and staying competitive requires adapting to AI-driven changes. Whether you’re looking to future-proof your career, refine your digital marketing strategy, or understand how AI and analytics can enhance decision-making, these key takeaways will provide valuable insights from our expert panel.

1. Striking the right balance: AI vs. human decision-making

A recurring theme in our discussion was how marketers should decide where to draw the line between AI automation and human oversight. The consensus? AI should be leveraged to handle repetitive, data-heavy tasks, such as audience segmentation, A/B testing, and predictive analytics; while human intuition and creativity remain irreplaceable in strategic decision-making and brand storytelling.

2. Ethical considerations in AI-driven marketing

AI can be a powerful tool, but with great power comes great responsibility. Our panellists emphasised the importance of AI ethics, data hygiene, and responsible AI use. Companies must establish clear policies to ensure that AI-driven marketing efforts remain transparent, unbiased, and aligned with brand values.

3. Career growth in AI-infused digital marketing

AI is not replacing marketers; it’s empowering those who know how to use it. Roles requiring AI proficiency have grown by 237%, and marketers with AI and analytics skills earn up to 45% more in the UK. For professionals seeking to future-proof their careers, mastering AI-driven tools and analytics is no longer optional, it’s essential.

4. The role of AI in hyper-personalisation

Personalisation at scale has become a reality thanks to AI. From real-time customer interactions to predictive behavioural analysis, brands are leveraging AI to create highly customised experiences. AI-powered analytics enable marketers to segment audiences dynamically, ensuring that messaging is tailored to individual preferences and behaviours.

Your questions answered

During the webinar, we received numerous insightful questions from the audience. While we addressed many in the session, we promised to follow up with responses to the remaining questions. Below, we have included every question asked, along with our expert responses. 

---

List of questions
Use the list below to navigate to the question and answer you’re most interested in.

1. How do you decide where the line is in terms of how much you use AI and when to take over manually?
2. A crucial perspective – the ethics of AI in marketing: Where to start with procedures? What matters more – organisational-level policies and procedures or micro-level marketing processes?
3. How do you make sure that the AI tools you’re using are delivering an output that is also reliable?
4. What are the best AI tools for analytics?
5. Data-driven decision-making is not new, marketers have had to do this for decades – what’s different now?
6. In scaling digital marketing services (using AI tools) to multiple geographies and considering different regulatory requirements towards use of AI, are there existing strategies that can address that or is it still something that is in development?
7. Which specific reports have you used AI for?
8. How do you see the role of an AI assistant, often discussed in the context of ChatGPT? Shouldn’t we emphasise more in marketing that this is an assistive technology for marketers rather than a superpower for marketers?
9. How can AI improve email marketing segmentation?
10. How does LSE’s Digital Marketing Strategy & Analytics Career Accelerator address AI learning?
11. What career support does the programme offer?
12. Do you have practical examples of where AI has been able to take away tedious, repetitive tasks? What exactly were these tasks?
13. How should data collection policies and data hygiene practices precede the implementation and use of AI? After all, marketing often deals with a large volume of unstructured and messy data.
14. How can I strategically revamp my skills as a home learner, given the vast array of marketing tools – some of which are enterprise-grade? How should I position my portfolio to appeal to recruiters?
15. How fast is court practice developing with respect to intellectual property/rights on AI-generated content, specifically used in marketing/advertisement campaigns?
16. In yesterday’s AI Masterclass, Claire spoke about giving her AI assistants ‘performance reviews’. How would you go about doing that? Going back to the data and double-checking their work?
17. Can you provide a practical example of using AI for personalisation?
18. AI is a powerful accelerator, but relying on it to “fix everything” without understanding the mechanics can create long-term problems. How do we balance the incredible productivity of AI with the need for solid foundational knowledge?
19. How about tools like HubSpot or SalesForce? Get the data, as it is not included in the list, what is your perspective on this?
20. Does the Career Accelerator detail using these AI tools?
21. ChatGPT 4.5 has been characterised as a somewhat strange “animal” in the sense that it isn’t quite clear what its added value is. The one thing I have heard rather consistently is that it seems to mimic these human, emotional, authentic “moments” we just heard about.
22. It’s all so overwhelming! How does one move on from casual use to starting to use it effectively?
23. What are the top AI tools for email marketing, especially regarding analytics? Do Mailchimp and Klaviyo provide sufficient AI capabilities, or are there better alternatives? Which tools excel in segmentation?

 

---

 

1. How do you decide where the line is in terms of how much you use AI and when to take over manually?

This comes down to understanding both the capabilities and limitations of AI – what is it good for, and what is it not? The best thing to do is to typically treat AI as an aid or intern, and not an all-encompassing solution. A good rule is to let AI handle high-volume, data-heavy or repetitive tasks, while you oversee strategy, creative judgment, critical thinking and final approvals. For instance, AI can draft initial copy, but you need to refine that copy as the AI likely will miss out on nuance, context and brand alignment that your human experience brings to the table. An AI tool can do analysis, but you need to verify that analysis, and interpret it to provide valuable insights that once again take into account your domain expertise, context and nuance. You should look to augment human work with AI rather than replacing it – ​ AI  can amplify your human capabilities.

A 2024 survey found that 48% of marketing leaders say AI most significantly improves customer interactions, yet they still rely on human creativity to craft authentic messaging​.

In practice, many teams adopt a “human-in-the-loop” approach: AI generates options or insights, and marketers curate or modify them. This balance maximises efficiency while avoiding the pitfalls of unchecked AI output​.

“As AI takes the reins on automation and personalisation, it’s crucial to remember that meaningful connections with customers are still forged in the crucible of human empathy and creativity. AI is not a usurper, but a co-pilot, amplifying our capabilities while leaving the wheel firmly in our hands.” 

The key is to use AI where it excels (speed, scale, pattern recognition) and step in when human empathy, context, or ethical judgment is needed.

2. A crucial perspective – the ethics of AI in marketing: Where to start with procedures? What matters more – organisational-level policies and procedures or micro-level marketing processes?

Both organisation-level AI policies and on-the-ground processes are vital, but start at the top. Companies are increasingly creating responsible AI guidelines to ensure ethical use across all departments​.

These high-level policies set the tone on issues like data privacy, transparency, and avoiding bias. For example, marketers should comply with emerging regulations (e.g. the EU’s AI Act or Brazil’s AI Bill) – understanding such laws helps “avoid penalties and ethical missteps”​

However, policies mean little without implementation in daily workflows. It’s crucial to train marketing teams on micro-level practices: obtaining proper consent for data use, ensuring AI-generated content is reviewed for fairness and accuracy, and monitoring campaigns for unintended bias. In 2024, 24% of organisations cited data quality (an ethical and operational concern) as a top challenge, indicating a need for better data governance as a foundation.

Bottom line: Begin with a robust organisational AI ethics framework (covering data handling, disclosure of AI use, etc.), then operationalise it via checklists and marketing process training. This two-tiered approach ensures that big-picture principles translate into ethical day-to-day marketing actions.

3. How do you make sure that the AI tools you’re using are delivering an output that is also reliable?

Rigorous evaluation and oversight are essential whenever AI tools are deployed. In practice, marketers “ground” AI in verified data and double-check its work​.

For example, if an AI analytics tool flags a trend, the team validates it against raw data or alternate sources. Many organisations implement human review stages – treating AI like a junior analyst whose findings need confirmation. It’s wise to set up quality metrics for your AI’s output. If using a content generator, measure accuracy and brand tone compliance; if using AI for predictions, back-test its recommendations against historical outcomes. Indeed, businesses that succeed with AI often keep humans in the loop – 80% of companies stress the importance of human oversight to guide AI and catch errors​.

Another tactic is starting with pilot tests: run the AI on past data where you know the answers to see if it produces reliable results before trusting it with new campaigns. Finally, maintain and update your AI models and data. An outdated model can drift, so retraining on recent data and monitoring performance over time is crucial. In summary, treat AI outputs as hypotheses – verify them through audits, A/B tests, or expert review. This way you gain the efficiency of AI while safeguarding accuracy and trustworthiness.

4. What are the best AI tools for analytics?

Marketers now have many AI-driven analytics platforms to choose from. Google Analytics 4 (GA4) is a popular choice – it has built-in machine learning to deliver insights (like automatic anomaly detection and predictive metrics such as purchase probability)​. GA4’s AI-powered features can spot significant changes in user behaviour without manual analysis​.

Another top tool is Tableau (by Salesforce), which introduced “Tableau AI” to automatically surface trends and correlations from your data​. Tableau’s integration of Einstein AI means it can point out anomalies or key drivers in plain language, helping non-analysts understand their data​

For those deep into multi-channel marketing, Salesforce Marketing Cloud’s Datorama (Marketing Intelligence) is highly regarded – it unifies data from various platforms and uses AI to provide optimisation recommendations and predictive insights​

In terms of campaign optimisation, Albert.ai is an AI platform specifically for marketing campaigns. It analyses performance and even adjusts budget or targeting on the fly to improve ROI​. Alternatively, Julius.ai is a great tool for general analysis, along with ChatGPT directly.

In summary, the “best” tool depends on your needs: GA4 and Adobe Analytics (with Adobe’s Sensei AI) excel for web/app analytics, Tableau and Power BI for flexible BI with AI assistance, and marketing-focused platforms like Albert or IBM Watson Analytics for campaign-specific optimisation. Notably, 54% of marketers are already using AI for reporting and data visualization in some form​showing that these tools are becoming mainstream in analytics workflows.

5. Data-driven decision-making is not new, marketers have had to do this for decades – what’s different now?

It’s true that using data in marketing is not new – even 20 years ago, savvy marketers relied on spreadsheets and databases for decisions. What’s changed today is the sheer volume, velocity, and variety of data, and the AI-driven automation that can harness it. We produce an astonishing amount of data now – 90% of the world’s data was generated in just the last two years​. In 2024, an estimated 147 zettabytes of data will be created over the year​.

This explosive growth means no human team can manually parse all relevant information. That’s where modern AI and big data tools make the difference: they process millions of data points in seconds, providing timely insights. It’s not that data-driven thinking is new, but real-time, AI-assisted decision-making is the game-changer. For example, instead of waiting weeks for a market research report, a marketer can get instant dashboard alerts when KPIs swing, and automation can even trigger responses (like adjusting an ad bid or sending a personalised offer) immediately. The timeliness and scale are key – an AI can detect a tiny shift in customer behaviour overnight and help you pivot a campaign the next day. As one marketing commentary put it, being data-driven decades ago meant looking at last quarter’s results; today it means reacting on the fly with automation and predictive analytics​.

Additionally, AI enables handling unstructured data (social media sentiment, video content insights) that we simply couldn’t leverage effectively before. The takeaway: using data is table stakes (always was), but now marketers have tools to use more data faster and more proactively. Those who combine data with AI-driven speed and scale are leaping ahead. (For instance, companies using AI report automating up to 40% of marketing tasks​, allowing teams to focus on strategy while machines crunch the numbers at scale.)

6. In scaling digital marketing services (using AI tools) to multiple geographies and considering different regulatory requirements towards use of AI, are there existing strategies that can address that or is it still something that is in development?

Deploying AI-powered marketing globally introduces a patchwork of evolving regulations, from data privacy to AI ethics. Right now, there isn’t a one-size strategy fully matured, but companies are developing workarounds. A smart approach is to build a compliance-by-design strategy: implement robust data governance and responsible AI practices that meet the strictest region’s requirements, and use that as a baseline everywhere.

For example, if you comply with Europe’s GDPR and the upcoming AI Act, you’ll be well-positioned for other markets. In practice, firms often maintain separate data environments per region (to respect data residency laws) and add features like consent management and AI transparency notices for regions that require it. We’re seeing strategies like AI risk assessments before launching tools in a new country, and configuration options to disable or tweak certain AI features to meet local laws. Indeed, experts advise a “policy-first approach”, establishing internal guidelines for ethical AI use and vetting any AI tools for compliance and security​.

Major organisations are creating AI ethics committees to keep abreast of legal changes and update marketing practices accordingly. While there’s progress – e.g., regulators in the EU are drafting clear rules (the EU AI Act will categorise AI use cases by risk and impose obligations), and a U.S. federal court just set a precedent on AI training data copyright​ – much is still developing. So, existing strategies revolve around staying agile and informed: investing in legal monitoring, partnering with local compliance experts in each region, and choosing AI vendors that offer compliance support (such as data localisation or documentation for audits). In short, the industry is in a learning phase – leading multinationals are creating their own playbooks (responsible AI frameworks, strict data hygiene, human oversight requirements) to safely scale AI marketing across borders, even as laws continue to catch up​. Expect these strategies to be refined as court decisions and new regulations emerge in the next 1–2 years.

7. Which specific reports have you used AI for?

As marketers, we’re increasingly able to lean on AI to automate and enhance all types of reports. For example, campaign performance reports—using an AI analytics tool to produce a weekly dashboard that highlights anomalies or key drivers across Google, Facebook, and email data or quickly pulling multi-channel data and spitting out a narrative like “Your conversion rate jumped 10% due to X campaign” without you crunching manually. Almost 55% of marketers say they use AI for reporting and visualisation now​.

Another area is SEO and content reports: tools like HubSpot’s content AI or MarketMuse can audit your website and auto-generate reports on content gaps or SEO opportunities, saving hours of tedious analysis.

Personally, we’ve used AI to draft monthly marketing summaries for stakeholders – e.g., feeding raw data to ChatGPT (or BI tools with GPT integration) to get a first draft of the story (“Sales grew x% month on month, driven by [source] or [segmemt]” etc.), which we then fine-tune. In one instance, our team used AI to analyse sentiment from customer feedback and incorporate it into a quarterly report; the AI categorised thousands of comments (positive and negative themes) far faster than a human could.

Another concrete example: using Google Analytics Intelligence in GA4 to ask questions like “What caused the dip in traffic last week?” – the AI might identify a specific referrer drop and essentially write part of your analysis for you. Overall, many tedious reporting tasks (data aggregation, basic trend analysis) can be offloaded to AI, allowing us to focus on interpreting insights and planning actions. This aligns with the trend that over half of marketers are using AI to automate reporting and glean insights faster​.

8. How do you see the role of an AI assistant, often discussed in the context of ChatGPT? Shouldn’t we emphasise more in marketing that this is an assistive technology for marketers rather than a superpower for marketers?

You’re right. It’s important to frame tools like ChatGPT correctly – they are assistive technologies, not magic wands that turn marketers into superheroes without effort. In other words, think of generative AI as your copilot, not an all-knowing autopilot​.

Its role is to amplify your capabilities: speed up drafting, suggest ideas, and crunch data for you – while you remain the decision-maker. Many marketing leaders now stress that ChatGPT should be seen as an intern or assistant on your team. For example, it can quickly generate a first draft of an email or social post, but a human adds creative polish and ensures it aligns with the brand voice. This positioning helps set proper expectations. As one article put it, “AI is not a usurper, but a co-pilot, amplifying our capabilities while leaving the wheel firmly in our hands.”​

In practice, that means we use ChatGPT to assist – to brainstorm copy variations, to summarise data, to translate technical jargon – but we don’t abdicate our judgment. Emphasising the assistive nature also encourages teams to develop AI literacy and not blindly trust outputs. The added value comes when marketers use these tools to enhance productivity (e.g., generating 5 campaign ideas in the time it used to take to draft one), yet still apply their marketing savvy to choose the best ideas. So yes, we should communicate that AI, like ChatGPT, is here to help marketers. It’s a powerful accelerator of tasks, but the marketer remains the strategist. This mindset also helps alleviate fear – the tool isn’t replacing your creativity, it’s augmenting it. By highlighting successes as “AI-assisted” rather than AI-performed, organisations can foster adoption while reinforcing that human insight remains at the core of marketing.

9. How can AI improve email marketing segmentation?

AI can analyse historical user data, engagement patterns, and predictive analytics to create dynamic audience segments, allowing brands to send more relevant, personalised messages. Platforms like Klaviyo, Mailchimp AI, and Salesforce Marketing Cloud offer AI-driven segmentation.

10. How does LSE’s Digital Marketing Strategy & Analytics Career Accelerator address AI learning?

Our programme includes an AI Learning Track, where participants master AI-powered tools, learn how to integrate AI into marketing strategies, and develop practical AI-driven projects under expert guidance, alongside the core programme content. Given the evolving nature of this technology, our facilitators work hard to bring real-world and current experience in leveraging AI in practice during live sessions that help you to apply that knowledge to your learnings. 

While it does not provide deep technical training in AI development, it ensures that learners understand how AI-powered tools fit into modern marketing strategies. By the end of the programme, participants – whether experienced marketers or those newer to digital marketing – will have gained hands-on exposure to AI applications in marketing, equipping them with practical, actionable skills to leverage AI for efficiency, insights, and strategic decision-making.

11. What career support does the programme offer?

The LSE Digital Marketing Career Accelerator is designed not just to impart skills but also to bolster your career outcomes. While it’s not a traditional degree programme with on-campus recruitment fairs, it offers strong career support services. For example, you get access to one-on-one career coaching sessions (the programme provides up to a full year of 1:1 coaching sessions) and group workshops on topics like CV building, interviewing, and current market trends​. Throughout the programme, there’s an assigned “success team” focused on helping you reach your personal career goals​.

A distinctive feature is the Employer Project: a real-world consulting project with an industry partner. In 2024, our learners teamed up to devise a Q4 digital strategy for an e-commerce company as part of a partnership with a marketing agency​.

This not only gives you hands-on experience for your portfolio but also direct exposure to potential hiring managers in the field. Many graduates cite this project and the networking opportunities it provides as key benefits. In lieu of campus recruitment, the programme leverages its industry network – the involvement of industry experts and partner companies can open doors (for instance, you might impress a partner company during the employer project).

Moreover, upon completion, you join LSE’s alumni community, which can facilitate job connections. In summary, while you won’t have a “placement day” like a university campus, you will have career coaching, practical projects with employers, and networking opportunities to support your job search. Many of our alumni note the value of these elements, highlighting the coaching aspects, alongside the employer project as instrumental in landing a role post-course​

12. Do you have practical examples of where AI has been able to take away tedious, repetitive tasks? What exactly were these tasks?

AI shines at taking over the mind-numbing chores that used to eat up marketers’ time. One clear example is data cleaning and integration – previously, assembling marketing data from multiple sources or scrubbing spreadsheets for errors was a dull, lengthy process. Now, AI tools can automatically clean datasets, merge information, and even flag anomalies. This means marketers spend less time managing CSV files and more time on strategy. Another big one is customer inquiry handling: AI chatbots handle repetitive questions (“What are your store hours?”) endlessly, sparing human reps those FAQs. This has a real impact – by 2025, AI chatbots are projected to save businesses about $8 billion in support costs, precisely by automating those routine Q&A interactions​.

Content generation at scale is another tedious task reduced. Need product descriptions for 500 items? An AI writing assistant can draft them in a fraction of the time it would take a person, turning what was once drudgery into a quick review task. In our team, we’ve seen AI cut down the time on weekly report generation – an AI tool now compiles the numbers and even writes a first-pass analysis, whereas an analyst used to spend half a day on it. A McKinsey study found AI could automate up to 40% of marketing tasks​, many of which are exactly those repetitive activities (like scheduling social posts, sorting leads, basic image editing/cropping for creatives, etc.). Email marketing provides a relatable example: AI can automate send-time optimisation and audience segmentation. Rather than manually creating segments and sending schedules every week, the AI learns and does it continuously. Marketers at companies using AI for personalisation have reported significant time savings – one report noted teams saved ~20 hours per week on content and outreach tasks due to AI streamlining these processes​

In summary, tasks like data entry, basic research, initial copywriting, routine customer comms, and simple optimisations have largely been lifted off humans and given to AI. This frees marketers to focus on creative and strategic work, while the “grunt work” gets done in the background.

13. How should data collection policies and data hygiene practices precede the implementation and use of AI? After all, marketing often deals with a large volume of unstructured and messy data.

Clean, well-governed data is the backbone of effective AI in marketing. The saying “garbage in, garbage out” very much applies – if your customer data is messy or collected without clear policies, your AI outcomes will suffer. So, before layering AI, organisations should strengthen data collection policies: ensure you’re gathering the right data (relevant, consented) and storing it in an organised way. This involves establishing data standards (for example, define consistent formats for customer names or a single source of truth for product data) and privacy practices (make sure you have consent for the data your AI will use and that you’re compliant with laws like GDPR). Equally important is data hygiene – de-duplicating records, correcting inaccuracies, and updating old info. Many companies now run periodic “data audits” to sanitise their databases before running AI models on them. The impact of poor data practices can be huge: a 2024 marketing study noted data availability and quality remain one of the biggest barriers to successful data-driven work, with 33% of organisations saying their data is insufficient or low-quality​. In complex AI projects, “data quality is [often] an impeding factor and exposes a fundamental problem” if not addressed upfront​.

Practically, this means marketers might spend time labelling data or working with data engineers to structure datasets before an AI campaign. For example, if you want an AI to personalise offers, first ensure your customer purchase history data is complete and standardised (no mixed formats or missing fields). Also, implement robust data governance: set roles and responsibilities for data upkeep, and maintain documentation so everyone knows the data definitions. It might feel like extra work, but it pays off – companies that manage data will see much smoother AI deployments. In fact, data management was the top cited challenge in AI projects in 2024​, and resolving that by front-loading good data practices will make your AI far more reliable. Think of it this way: before you let an AI drive insights from your marketing data, tune up the engine (the data) so it runs clean. By having strong collection policies (only collect what’s useful and with transparency to users) and hygiene routines (regular cleaning, validation, and security checks), you set your AI up for success and avoid “junk data” leading you astray.

14. How can I strategically revamp my skills as a home learner, given the vast array of marketing tools – some of which are enterprise-grade?  How should I position my portfolio to appeal to recruiters?

With so many AI and analytics tools (some enterprise-only), the key is to focus on broadly accessible skills and demonstrate real projects. As an experienced professional returning after a break, you can start by learning the popular tools that have free versions. For example, get hands-on with Google Analytics 4 and its AI features (it’s free) to show you understand modern web analytics. You could also use free tiers of Mailchimp or HubSpot to practice building automated campaigns with AI-driven elements (HubSpot’s AI content assistant, for instance, is available on free plans). Since enterprise tools like Salesforce or Adobe might be out of reach at home, look for simulation platforms or case studies – many course providers or community blogs share sample datasets and scenarios you can work on to mimic using those tools.

When building your portfolio, the key is to show outcomes and processes. For instance, you might include a self-initiated project: “Predicted e-commerce sales using Python – achieved X% accuracy” or “Implemented an AI-based customer segmentation on a sample dataset using KNIME”. Emphasise your 15-year background by framing your projects around consumer psychology insights augmented by AI (“Combined my consumer behaviour knowledge with an AI clustering algorithm to segment customers more intuitively”).

Recruiters will look for evidence that you can apply AI practically, not just theoretical knowledge. That could mean including before-and-after examples (e.g., a marketing plan you optimised with an AI tool’s input). Also, consider getting certified: certifications in Google Analytics IQ, Azure AI fundamentals, or even a course certificate from LSE’s programme can lend credibility.

In terms of how a recruiter views your portfolio: they’ll appreciate seeing that despite your career break, you’ve proactively upskilled. Showcasing projects completed on your own initiative signals self-drive. It’s also good to highlight any real-world context you can – if you did a volunteer or freelance project using AI, include that. And don’t underestimate the value of explaining the why and how in your portfolio documentation. For example: “I identified that predicting customer churn could be valuable, so I trained a model on open-source data – here’s what I used and the result.” Given that 68% of business leaders struggle to find talent to manage AI solutions​, your blend of psychology, analytics experience, and new AI skills can be very attractive. Just make it easy for them to see those skills in action through concrete portfolio pieces and articulate the impact or insights gained from each project.

15. How fast is court practice developing with respect to intellectual property/ rights on AI-generated content, specifically used in marketing/advertisement campaigns?

The courts and regulators are rapidly trying to catch up with AI’s impact on intellectual property. Over the past year, we’ve seen a flurry of landmark cases and decisions. In the US, it’s been clarified that purely AI-generated works cannot be copyrighted – the U.S. Copyright Office (and a federal judge in 2023) held that if there’s no human author, the content isn’t eligible for copyright protection​. This means if you have an AI create an image or copy for an ad entirely on its own, you likely can’t stop others from using a similar output; there’s no traditional IP protection for that AI-only content. Meanwhile, multiple high-profile lawsuits have emerged regarding the training data used by AI models. For example, Getty Images is suing Stability AI (maker of Stable Diffusion) for allegedly using millions of Getty’s photos in training without permission​. In another case, a class-action lawsuit by artists against AI image generators (like Midjourney/Stable Diffusion) is moving forward, with a judge in August 2024 upholding the artists’ copyright infringement claims – an early victory for creators concerned about their style being mimicked​. Most relevant to marketing content, a recent decision in February 2025 (Thomson Reuters v. ROSS) found that using a copyrighted dataset to train an AI can infringe copyright when that AI competes in the same market​.

In that case, an AI company was ruled against for training on a publisher’s written summaries. This is a precedent-setter: it suggests that if, say, an AI was trained on a news article and spits out something very similar for an ad, it might not be protected by fair use. Courts are basically establishing that feeding copyrighted content into AI doesn’t magically nullify the copyright – context matters (especially if the AI output replaces the original work’s purpose)​

Additionally, regulatory bodies are starting to weigh in: the US Copyright Office released guidelines that if a human simply provides a prompt and the AI does the rest, the output isn’t a human-authored work and can’t be registered​

So where does this leave marketers? Essentially in a cautious position. The “court practice” is evolving monthly. It’s moving fast in the sense that lawsuits dominated 2024 and early rulings like the ROSS case arrived​, but final outcomes (especially at appellate or Supreme Court level) are still pending.

Marketers using AI-generated content need to watch these developments: ensure they have proper rights for any training data used (if they train custom models) and be transparent in usage. Also, be mindful of trademark and personality rights – using AI to generate a celebrity likeness or a brand logo variation could invite legal issues under existing laws, which are being interpreted for AI now. In summary, the legal terrain is actively being mapped. Expect more clarity in the next couple of years as these cases resolve, but for now, the best practice is to err on the side of caution: treat AI outputs as you would source content – vet it for potential IP conflicts (some brands even avoid using AI images in campaigns until there’s more certainty). Keep an eye on cases like the above, because they’re setting the rules that will directly affect how we can use AI in advertising.

16. In the AI Masterclass, Claire spoke about giving her AI assistants ‘performance reviews’. How would you go about doing that? Going back to the data and double-checking their work?

The idea of giving your AI assistants (in this instance, customGPTs) a performance review is essentially about regularly evaluating their output and impact as you would a team member’s work. You shouldn’t just set an AI tool and forget it – you need to periodically assess: Is it doing a good job? Is it improving or causing any issues? In practice, you’d gather some metrics and qualitative observations.

For example, if you’re using an AI copywriting assistant, you might track how content generated by the AI performs (engagement metrics, error rates, amount of editing needed) and compare it over time or against human-generated content. This is akin to a KPI review for the AI. If the AI is a chatbot handling customer queries, you might review transcripts each month and see if it’s satisfying customers – perhaps looking at resolution rates or customer satisfaction scores.

Going back to the data is key: just like a manager would look at an employee’s outputs, you’ll pull reports on your AI’s outputs. Did the AI’s recommendations actually boost conversion? Are there patterns of mistakes (for instance, the AI consistently messes up product names or tone in email drafts)? You’d “coach” the AI by refining prompts or updating its training data based on these findings. Literally, sit down and scrutinise the AI’s work output at regular intervals.

For us, that means scheduling a periodic review (say, monthly) where we audit a sample of AI outputs: maybe 10 email subject lines it suggested and the outcomes of those campaigns. If we find, for example, that the AI’s subject lines with a certain phrasing consistently underperform, that’s feedback to adjust either our usage or the AI’s parameters.

Another facet is checking the AI’s accuracy and compliance. Part of the “performance review” could be running fact-checks on a random subset of AI-generated content to ensure it’s not hallucinating or straying from guidelines. If issues are found, we address them – much like telling an employee where to improve. We can also give positive feedback: identify what the AI did well (“These social posts required minimal editing – great job on tone!”) and use that insight (maybe use similar prompts going forward). It’s a slightly tongue-in-cheek concept, but it underscores an important practice: continuously monitor and fine-tune your AI tools.

In essence, treat your AI assistant as a junior coworker who needs oversight and feedback. By doing so, you maintain high quality and can gradually train the AI (through better prompts or additional data) to get even better. It’s a proactive approach – rather than waiting for a major error, you catch small issues in these “reviews” and course-correct. So yes, it involves going back to the underlying data and outputs, measuring them, and adjusting accordingly, on a regular cadence.

17. Can you provide a practical example of using AI for personalisation?

Starbucks’ AI-powered personalisation is a standout example. Starbucks leverages a predictive analytics engine called Deep Brew to tailor marketing messages to each customer​.

Through the Starbucks mobile app, AI analyses individual purchase history, preferences, and even local weather – then recommends products or sends offers timed to when you’re most likely to want a coffee​

For instance, on a hot afternoon, the app might suggest a cold drink you’ve enjoyed before, while on a rainy morning, it might promote a warm latte. This personalisation strategy increased app engagement and sales, as customers received highly relevant suggestions​

Another great example is Sephora’s Virtual Artist: an AI-driven advisor that provides personalised makeup recommendations and lets users virtually “try on” products. Sephora’s AI analyzes user preferences and feedback, delivering tailored product suggestions – contributing to a 30% increase in online sales​

The common thread: AI ingests lots of customer data and outputs uniquely personalised content or offers, resulting in a better customer experience and improved business outcomes.

18. AI is a powerful accelerator, but relying on it to “fix everything” without understanding the mechanics can create long-term problems. How do we balance the incredible productivity of AI with the need for solid foundational knowledge?

It’s all about augmentation, not automation in isolation. To harness AI’s incredible productivity safely, you need a strong foundation in marketing principles and data literacy within your team. Concretely, this means training your marketers in how the AI tools work – not the deep code, but the logic. For example, if you use an AI to optimise bids, ensure the team understands concepts like how the algorithm makes decisions, what data it’s using, and its limitations. By having that knowledge, they can spot when the AI might be going astray (say, over-prioritising a wrong metric) and intervene. One approach is to establish guardrails: set policies that certain decisions must be reviewed by a human, especially in early implementation. Perhaps AI can auto-generate content, but nothing goes live without human approval. That way you get the speed (AI drafts in seconds) but maintain quality control. 

Another tactic is investing in AI literacy for staff – indeed, many organisations now run workshops on understanding AI outputs. (As a side note, a webinar on “Mastering AI Literacy – What You Don’t Know WILL Hurt You” underscored that blind reliance is risky​.) Supporting data: 80% of companies in a 2024 report said human-in-the-loop machine learning is vital​, meaning they explicitly keep humans involved to guide AI and handle context or ethical considerations. This human oversight ensures that while AI handles the grunt work, the strategy and critical thinking remain human-driven.

We also balance short vs long term by monitoring AI decisions for unintended consequences. For example, if an AI recommends doubling down on one customer segment because it’s easiest to convert, a marketer with foundational knowledge might realise that’s a short-term win but long term could narrow our brand appeal – so they adjust the strategy. It’s similar to how pilots use autopilot but must know how to fly the plane if something goes wrong.

Practically, encourage your team to question AI outputs (“Does this recommendation make sense given what we know about our customers?”) and to use AI as a starting point. Build a culture where AI is the first draft or initial analysis, and human expertise refines and finalises. Over time, as the team’s confidence and understanding of the AI grows, they can let it handle more, but they will always have that fundamental knowledge to fall back on. Finally, mix training with experimentation – allow the team to run small tests where they use AI and measure results vs. a human-only approach. This hands-on experience cements their understanding of the mechanics. In short, to balance the two: never stop learning the “why” behind AI decisions. Use AI’s speed, but always apply human insight before acting on its output. That ensures you get the best of both worlds – efficiency with wisdom.

19. How about tools like Hubspot or SalesForce? Get the data, as it is not included in the list, what is your perspective on this?

HubSpot and Salesforce are indeed major platforms that weren’t on the earlier tool list, likely because they’re all-in-one ecosystems rather than standalone AI tools. Both have been integrating AI features extensively. HubSpot has introduced multiple AI-driven capabilities in 2023, including an AI content assistant (for writing marketing copy) and AI-powered predictive lead scoring and forecasting​.

Under the hood, HubSpot’s AI features work by leveraging the data already in your CRM – web interactions, email engagement, and sales activities. HubSpot centralizes customer data (from form fills, site tracking, etc.), and its AI can then analyse that to do things like suggest the best time to send an email or which leads are likely to convert. It’s essentially pulling data from within the HubSpot CRM and using machine learning models on it. Salesforce, via its Einstein AI, similarly ingests data from across the Salesforce ecosystem – your email opens, ad clicks, sales history, and more (plus external data integrations) – to provide insights and automation. For example, Einstein can automatically segment audiences or recommend content by learning from your past campaign data. Salesforce’s Marketing Cloud (and their acquired tool Datorama) unifies data from ads, websites, CRM, etc., and uses AI to find patterns and optimization opportunities​

Why weren’t these in some lists? Often when people list “AI marketing tools,” they highlight specialised tools or newer entrants. HubSpot and Salesforce might not have been mentioned if the context was AI analytics tools specifically – but make no mistake, they do have strong AI capabilities. Data-wise, these platforms get their data through integration and native tracking: HubSpot has tracking code on websites and forms capturing info, and it merges that with email engagement data in its database. Salesforce can intake data from its CRM entries, e-commerce plugins, social listening integrations, etc., funnelled into its CDP (Customer Data Platform) and Einstein. Both are walled gardens to a degree – they primarily work with data within their system, though you can import external data too. My perspective: If you’re already on HubSpot or Salesforce, their built-in AI can be very powerful and convenient (no need to export data to another AI tool). For instance, HubSpot’s AI can draft you an email using your CRM context in one click​.

Salesforce’s Einstein might highlight that a certain customer segment is at risk of churn by looking at all their touchpoints. The key is that these aren’t “out-of-the-box” AI point solutions – they’re part of larger platforms. So, many “top tools” lists focus on independent AI tools that anyone can adopt. But in reality, many companies will get a lot of AI functionality from HubSpot or Salesforce if they’re users. They were effectively absent from the earlier list because those lists often concentrate on newer, standalone AI offerings, whereas HubSpot and Salesforce are comprehensive suites that include AI among many features. In sum, HubSpot and Salesforce get their data from the integrated customer interactions within their systems (plus any you connect), and they absolutely use AI on that data – even if they weren’t name-checked, they are heavyweight solutions for AI-driven marketing analytics and automation. As a marketer, leveraging the AI inside these platforms can be low-hanging fruit since it operates on data you’re already collecting, albeit typically at an enterprise price point.

20. Does the Career Accelerator detail using these AI tools?

Yes – the LSE Digital Marketing Career Accelerator is structured to provide hands-on exposure to modern marketing tools, including AI tools. The curriculum emphasises the practical application of digital marketing tactics​. In our dedicated AI and Digital Marketing Learning track, you’re introduced to specific generative AI and analytics tools as part of your learning.

For example, during the course you might learn to use tools like Google Analytics 4 or a social media analytics tool with AI insights when covering the analytics module. In the content creation sections, you could experiment with an AI copywriting tool (perhaps using ChatGPT or a tool like Jasper) to understand how to generate and refine marketing copy.

The presence of industry experts on the instruction team also means you’ll get exposure to the tools they use day-to-day. For example, one expert, Jill Quick, could demonstrate using Google Data Studio/Looker Studio with AI integrations for reporting. The programme doesn’t just speak about tools in abstract; it incorporates them.

21. ChatGPT 4.5 has been characterised as a somewhat strange “animal” in the sense that it isn’t quite clear what its added value is. The one thing I have heard rather consistently is that it seems to mimic these human, emotional, authentic “moments” we just heard about it.

GPT-4.5 (the intermediate update OpenAI released) has been noted for its higher “emotional intelligence” in responses and more human-like tone. In other words, its answers feel warmer, more intuitive, and better at mimicking those authentic human “moments” of understanding. OpenAI themselves highlighted that conversations with GPT-4.5 “feel more emotionally nuanced” and attuned to the user’s intent​.

For marketers, this added value translates to AI that can engage customers in a more relatable way. Think of customer service chats or brand social media interactions – an AI that can sense if a user is frustrated and respond with appropriate empathy (rather than a stilted generic reply) is a big step up. GPT-4.5 is basically making AI a better conversationalist. Its role, then, can be as a more effective creative and communicative assistant. For instance, for content teams, GPT-4.5 picks up subtle cues in a creative brief much better​.

If you imply you want a playful tone in an ad copy, GPT-4.5 is more likely to deliver that without explicit instructions, compared to earlier models. It also maintains context over long dialogues with fewer reminders, which helps in brainstorming sessions that mimic a human creative partner. So I see GPT-4.5 being used as an “emotionally aware copywriter” or chatbot that can produce marketing content carrying a semblance of human touch. Marketers have reported that GPT-4.5 gives responses that align more closely with how a human teammate might respond in a brainstorming meeting – including encouragement or understanding of the approach of an idea​.

The consistent feedback that it mimics authentic moments suggests, for example, if a user says “I’m not sure about this product because I feel X,” GPT-4.5 might address that feeling (“I understand how you feel…”) in a way older models wouldn’t. From a strategic view, this can enhance customer experience: imagine AI-driven email campaigns that adjust tone per segment – upbeat for engaged customers, reassuring for hesitant ones. GPT-4.5’s nuance could help power that. However, we should remember it’s still a tool – it’s not genuinely feeling emotions; it’s pattern-matching emotional language. Its role should complement the marketing team by handling personalized touches at scale, but oversight is needed to ensure those “emotional” outputs remain on-brand and appropriate. In summary, GPT-4.5’s key added value is more human-like interaction quality. Its role is likely to be as an even more capable assistant in content creation and customer engagement, perhaps serving as the backbone for more natural AI chatbots, or helping marketers craft messages that resonate emotionally with target audiences. It’s an evolution making AI a bit less robot, a bit more human in flavor – and used wisely, that can strengthen how brands communicate authentically at scale.

22. It’s all so overwhelming! How does one move on from casual use to starting to use it effectively?

Feeling overwhelmed by AI’s possibilities is common – the remedy is to start structured and small. To move from dabbling to real productive use, begin by identifying a specific marketing task or pain point that AI could help with. For example, instead of trying to use AI for “everything,” pick one area: say you often spend too much time A/B testing email subject lines. That could be a focus – you might start using an AI tool to generate and test subject line variations, which is a contained use case. By narrowing the scope, you can learn deeply and see results, which builds confidence.

Next, invest time in learning the tool or technique properly (watch a tutorial or take a mini-course on that particular application). Often, people remain casual users because they haven’t fully learned the features; a bit of front-loaded learning can unlock a lot of value. Set a mini-project or goal: for instance, “This month, I will use an AI scheduling assistant to automate our social media calendar and measure if it frees up 5 hours/week for me.” Having a concrete goal and metric helps you evaluate effectiveness. Many marketers find it helpful to integrate AI into their routine by scheduling it – literally put on your calendar time to use AI for a task, so it becomes a habit. Also, document your experiences.

As you transition to serious use, keep notes on what works or fails. This will turn into your personal best-practice guide and reduce the trial-and-error feeling. Another tip: leverage community knowledge. Join a marketing AI forum or LinkedIn group; seeing how peers apply AI can spark ideas and provide proven workflows. It helps to know you’re not alone and to pick up use cases that have worked for others. A statistic to encourage you: 90% of workers using AI report it saves them time​, so the effort to integrate it effectively does pay off in efficiency.

The trick is to be intentional – don’t try to use AI for its own sake, tie it to a strategy. For example, if your strategy is improving personalisation, maybe start using an AI-driven segmentation tool. As you see positive results (like better engagement from those segments), it will motivate further adoption. Finally, treat moving to effective use as an iterative journey. Start small, evaluate, and then scale up to more areas. Perhaps begin with content generation, then add analytics, then chatbots, etc., rather than all at once. Each success builds momentum. In essence: focus on one use case at a time, align it with a clear goal, measure the impact, and gradually expand. This way, AI becomes a regular part of your toolkit, and the sense of overwhelm gives way to a sense of opportunity.

23. What are the top AI tools for email marketing, especially regarding analytics? Do Mailchimp and Klaviyo provide sufficient AI capabilities, or are there better alternatives? Which tools excel in segmentation?

Email marketing has gotten an AI boost in both content and analytics. For analytics and segmentation, Klaviyo stands out as a platform with strong AI capabilities. Klaviyo offers predictive analytics for e-commerce (it can predict customer lifetime value or the next purchase date) and automatically creates segments based on customer behaviour and likelihood to buy​. In fact, Klaviyo has built-in AI-powered segmentation features that Mailchimp lacks​ – for example, it can dynamically group customers (e.g., “high spenders likely to churn”) and update those segments in real time. Mailchimp, on the other hand, is known for user-friendliness and does have some smart features (like send-time optimization and content suggestions), but it doesn’t yet have the same depth of predictive segmentation as Klaviyo​.

If you’re considering alternatives: HubSpot is a great option if you need an all-in-one with AI – HubSpot’s email marketing tool integrates an AI content writer and can leverage CRM data for personalization (e.g., its AI will auto-tailor email text based on CRM fields, and even record campaign performance automatically)​

For pure-play AI enhancements, tools like Phrasee (now rebranded as Jacquard) specialise in AI-generated copy for emails. Jacquard uses a proprietary ML engine trained on tons of marketing language data, allowing it to generate and optimize subject lines and messaging that resonate with different audience segments​.

It’s enterprise-level (used by big brands for millions of emails) and known to outperform generic A/B tests by finding the right language for each segment. Another tool in segmentation is Optimove – it’s a CRM marketing platform that uses AI to constantly re-segment audiences based on multi-channel behaviour and orchestrate tailored campaigns. Optimove can be heavy-duty, but it excels at calculating customer micro-segments and determining the next best action for each.

If we focus on analytics, Litmus (while primarily an email testing platform) has introduced some AI-driven insights for subject lines and even spam prediction. There’s also Mailmodo, which is newer. It’s more about interactive email, but they’ve been adding AI features (like an AI subject line assistant).

For whether Mailchimp and Klaviyo are “sufficient”: If you’re a small to mid-sized business, Klaviyo’s AI segmentation and reporting is likely more than sufficient – it’s quite advanced and tailored for commerce. Mailchimp is beefing up its AI slowly (they added a creative assistant for design and some smarter targeting), but for now, power users often supplement Mailchimp with external AI tools. For example, one might use Phrasee integrated with Mailchimp to get better subject lines, since Mailchimp’s built-in suggestions are basic.

Segmentino and Zeta Global’s platforms excel in segmentation, but those are more enterprise tools. Also, keep an eye on Salesforce Marketing Cloud—its Einstein AI can analyse email engagement and automatically segment audiences by likelihood to engage and even choose optimal send times and frequencies for each subscriber.

In summary: Klaviyo is top-tier for AI-driven email segmentation and analytics for ecommerce. Mailchimp is catching up but might need help from specialised tools for advanced AI functionality. HubSpot is strong if you want AI within a broader marketing suite. Specialised tools like Jacquard (Phrasee) excel at the content optimisation side of email (subject lines, copy) using AI, often driving higher open rates through language that best fits each segment​

For segmentation excellence, Klaviyo and Optimove are great; for AI content and testing, look at Phrasee/Jacquard or even Seventh Sense (which focuses on send-time optimisation via AI). The best choice depends on your scale and budget, but even with Mailchimp or Klaviyo alone, you can leverage AI – Klaviyo’s auto-segmentation and Mailchimp’s smart send features are built-in conveniences that use AI under the hood.

But it’s changing every day…Keep an eye on your favourite tools’ product roadmap and updates regularly.

Why learning AI and analytics is crucial for marketers

With 87% of marketers already using or experimenting with AI, those who fail to adapt risk being left behind. AI is revolutionising marketing strategies, from data-driven decision-making to hyper-personalized customer experiences. Professionals equipped with AI expertise are securing AI marketing jobs, excelling in data-driven marketing careers, and positioning themselves at the forefront of the future of digital marketing.

The LSE Digital Marketing Strategy & Analytics Career Accelerator equips professionals with the knowledge, tools, and hands-on experience required to lead in an AI-powered marketing world. 

Take the next step in your career

Whether you’re a career advancer, career changer, or career starter, our programme is designed to provide in-demand skills, hands-on learning, and personalised career support to help you thrive in the AI-driven marketing landscape.

Take the first step towards mastering AI-driven marketing: download the brochure or speak to an enrolment advisor today to future-proof your career.