Recently, I completed Andrew Ng’s AI for Everyone course on Coursera. While I didn’t receive a formal certificate as I completed it in audit mode (on a backpacking budget), the goal was to deepen my understanding of artificial intelligence. As AI continues to reshape industries, I wanted to explore its broader social, ethical and business implications, and reflect on how these shifts might influence the way we approach marketing. Here are five key insights that stuck with me and how they intersect with the future of marketing.
1. AI is powerful but it’s narrow, not general.
Key takeaway: Most AI today is narrow artificial intelligence (ANI), meaning that it performs specific tasks such as predicting clicks or recognising speech. We’re still far from artificial general intelligence (AGI), the kind that reasons, understands nuances, and thinks like a human.
Why it matters: AI doesn’t “understand” in the way we think it does; it identifies patterns in data, and uses those patterns to make predictions. This can be incredibly useful, but also limiting. Overestimating AI’s capabilities can lead to inflated expectations and misuse. Understanding AI’s current strengths and constraints help us identify realistic use cases.
Marketing application: AI isn’t going to write your entire marketing strategy, however it can help with narrower scope tasks such as basic copywriting, scoring leads using past data, and helping to optimise ad spend with improved targeting. It’s this understanding of AI’s narrow scope which helps marketers use it where it can add real value, without expecting it to “think” for us.
2. You don’t need big data to get started.
Key takeaway: Many companies decide to delay AI adoption because they think that they require huge datasets. It’s important to realise that this is not necessary, and that you can start small and rely on good quality data rather than quantity.
Why it matters: Waiting until you have the ‘perfect’ dataset can delay innovation. Instead, choose to iterate with small, clean datasets, so that you can reveal insights sooner and most importantly, identify what data is best to collect next. This will ultimately speed up the learning cycles.
Marketing application: You don’t need thousands of clicks to start optimising your email timing or audience segmentations. With as little as a few hundred entries or less, you can begin to identify patterns such as what subjects lines are getting opened on weekends vs weekdays, or how cart abandonment differs by user location. Although, it’s important to ensure this limited data is clean and relevant.
3. Most of today’s marketing tools already use AI and knowing how they work gives you an edge.
Key takeaway: Machine learning is already powering many of the tools we use today, such as search engines, email spam filters, ad platforms, A/B testing tools, and CRM systems.
Why it matters: Marketers use AI on a daily basis, often without realising it. Knowing the basics of how these systems work can give you the confidence to better interpret the results, ask smarter questions, and avoid missteps.
Marketing application: Take Meta or Google Ads for example. Their algorithms predict who’s most likely to convert based on a range of factors such as targeting and user behaviour. If you know that machine learning is at play, you’ll better understand the importance of feeding clean, well-labelled conversion data (not just clicks) into the system.
4. Bias in AI is real and marketers have a role to play.
Key takeaway: AI can amplify existing societal biases if it’s trained on biased data. Whether it’s recruitement tools that favour men or facial recognition failing people of colour, these aren’t just edges cases, but system risks.
Why it matters: Bias isn’t just a technical issue, it’s a human one. Since AI reflects the data it’s trained on, it can reinforce inequality unless we actively intervene. This becomes especially important in marketing where we’re shaping narratives, targeting customers, and influencing perception.
Marketing application: Personalisation is powerful, but it can be exclusionary. A recommendation system trained on bias purchase history may unintentionally sideline diverse customer groups. As such, marketers must scrutinise not just performance, but representation. Are your AI-driven segments reflective of your audience? Is your training data diverse? And are you using multi-disciplinary teams including data scientists, ethicists, and business leads? By embedding fairness and transparency into AI use, we not only improve outcomes but we protect brand trust.
5. Think small and strategic when choosing AI projects.
Key takeaway: Successful AI transformation doesn’t begin with a ‘huge bang’, it starts with small, feasible projects that solve real business problems.
Why it matters: Big, flashy AI projects can often stall AI development within a business due to overreach. In turn, leading to negative sentiment surrounding its use. It’s better to begin with something measurable that aligns with business goals (e.g. increasing conversions or reducing churn), and then scale from there.
Marketing application: Start with a low risk pilot such as using AI to predict the best send time for emails or to cluster customer reviews by sentiment or topic. Track the results, build trust within the business, and then expand. These small wins create a virtuous cycle; creating a better product/service, leads to more users, and then, more data.
I took this course not to become an AI expert but to make a conscious effort to understand how AI fits into what I do and to learn how I can use it more meaningfully as a marketer.
The biggest thing I learned was that AI isn’t a threat to us, at least not yet or in the near future, but it is a powerful tool. However, it’s up to us to use it ethically, creatively, and in a strategic manner.
I was also fascinated by the wide range of roles involved in scaling AI. While a project might start with a data engineer, it can quickly expand to include machine learning specialists, AI product owners, and more. Crucially, success depends on collaboration across department, ensuring the technology is developed alongside subject matter experts who bring essential business context.
I look forward to continuing to expand my knowledge of AI, particularly its use case among marketers.
