Can AI Help Advertisers Deliver Personalisation with a Point?

For the longest time, large-scale personalisation has been touted as the ultimate dream for every brand and marketer - to reach every possible individual consumer in a unique way, tailored to their preferences, purchasing habits, lifestyle and beyond. And now with the rapid development of artificial intelligence tools, this dream feels closer to reality than ever.

However, with the temptations of endless content, creative technologists warn of the possibilities of asset overload, quantity being prioritised over quality, a dilution of strategy and more.

To find out where the industry currently stands on the potential of AI-powered personalisation at scale, LBB’s Ben Conway spoke with AI experts and creative and production leaders. See what they had to say below.

Jeremy Lockhorn

SVP, creative technologies and innovation at the 4A's 

I see three key challenges related to AI-driven content personalisation:

Privacy: True personalisation at scale relies on accurate data. Third-party cookie deprecation is one of many signs pointing to growing consumer awareness of ‘data exhaust’ in an increasingly private digital world. Can AI deliver personalisation at scale without sufficient training data? Probably not.

Personalisation Gone Wrong: Nobody likes endless retargeted ads for a product they’ve already purchased. AI models powering the next generation of personalisation must get targeting right. Misdirected personalisation is far worse than none. 

Personalisation Fatigue: Consumers have come to expect what I call ‘clairvoyant marketing’ that predicts and delivers what they want. Platforms we interact with daily reinforce this expectation (i.e. the auto-complete feature in emails and addresses, iPhone’s Siri suggesting you call a person back after a missed call, etc.) However, if overused, especially given the first two risks, consumers may soon begin to tune out. 

All that said, I’m optimistic that AI-powered personalisation at scale is possible - provided we as an industry address these challenges in an open and transparent way. The 4A’s ‘Perspective on GenAI’ and our related crash course provide a solid foundation. 

James Calvert

Head of generative AI at M&C Saatchi

The dream of personalisation at scale has long captivated marketers, and AI is now bringing this vision closer than ever. Yet, these advances risk overload, strategic dilution, and low-quality content. The principle of ‘garbage in-garbage out’ underscores that the quality of inputs - both creative and strategic - directly affects outputs.

At M&C Saatchi, we understand that effective personalisation taps into human psychology. It’s not merely about recognising a name or your recent purchase, rather it’s about showing every single person the recognition, and creative assets that you are most likely to pay attention to.

Simply put, without a clear hypothesis for why personalisation should work, you risk missing your mark. To forge a genuine consumer connection, you will need to embrace strategic, high-quality personalisation. This calls for the discipline to prevent unchecked variations, and to focus on quality over quantity, so marketers can truly realise the potential value of AI in achieving personalised marketing.

James Galland

Global commerce director at VML

We’re at the precipice of a revolution - not just in AI, but rather a systematic transformation in both the depth and richness of data that brands can receive and the way that data can be leveraged to create better, more tangible consumer experiences.

Whether it's clean rooms, evolving retail media networks or first-party data, brands can understand their consumer base more effectively than ever before. This veritable smorgasbord of insights is a marriage made in heaven for the potential application of AI, and a ‘beacon’ that should guide the pivot from endless irrelevant content to content experiences that resonate.

Fabienne Queyrane

Global growth lead at Publicis Production

With AI unleashing content production capabilities, there’s always a risk of producing more, less good, content. However, it's important to remember that AI is more than just generative AI. We can also use it to better understand audiences and the content that resonates best with them. 

At Publicis Production, we've been using AI image recognition for years to understand the creative performance of assets. We've built this into an engine called PX Perform which allows for real-time creative optimisation of assets. Another interesting application of AI is in ensuring brand consistency, meeting regulatory requirements, and adhering to company guidance. We call this Brand Check and it helps accelerate the delivery of high-quality, brand-compliant assets. 

AI not only facilitates faster content creation but, more importantly, helps us to make better-informed decisions about which content is best for which audience and platform. Our promise to brands is: ‘Make only what you need. Make more of what works. Make every asset work harder’ - and AI is empowering this. In summary, it’s not just about crafting more content but better-performing content, limiting unnecessary asset creation, which also aligns with our sustainability goals at Publicis Production.

Eva Neveau

Global CCO at Designory

Over the last decade we’ve seen that shifting from what’s right for a brand to what’s right for a consumer delivers better results for everyone involved. If AI personalisation truly brings a more honed consumer-centric experience and keeps people from drowning in choices, I’m all for it.

To get AI-enabled personalisation right, brands have to train custom models to perform within the context of their domain. This means their products, features and services, plus the nuances of their specific consumers - no more relying on foundational models that are exclusively trained from public-domain data. These basic models lack the necessary context and guardrails which could lead to implicit biases, generalised messages and a higher degree of inaccuracies or hallucinations. Pulling from an ocean of public-domain data totally misses the point of personalisation. And, we can’t forget about what AI will never know - what it means to be human. Since we’re trying to reach other humans, our expertise and intuition has to be part of the solution.

David DiCamillo

Chief technology officer at Code and Theory

Just because we can make endless content doesn’t mean we should or that it will be effective. Generative AI content fatigue is real and we have to balance what is possible with what is in the best interest of our customers/users. We must ensure that we are adding value or we risk causing a negative experience and more ineffective branded noise.

Personalisation as a concept is now over 15 years old and even the word insinuates a ‘reactive experience’ based on user data. If we are truly going to live into what is possible with generative AI, we need to think about delivering ‘anticipatory’ customer experiences where brands and services are providing benefits to users' lives rather than just reacting with customised content. Think; H&M using geofenced SMS when near its stores to send unique discounts and inspiration based on your order history, and when the user responds an in-store employee readies a dressing room with those deals plus other suggestions – a truly personalised experience and beneficial service to the customer.

Ken Pappanduros

SVP, group creative director at RPA

I think one of the big dangers of limitless personalisation is the dilution of a brand’s core values. What does ‘brand’ even mean when you have the ability to be all things to all people? There’s nothing substantial or permanent in your messaging. Your brand becomes whoever you are talking to in a given moment. It’s the worst kind of inauthenticity. I think people will still want to feel like the brands they believe in stand for something and ultimately say something about who they are as individuals. A brand that changes in every wind can’t possibly offer that. 

Mike Creighton

Director of AI research and development at Instrument

The promise of personalisation. There are two factors in bringing this to fruition: the content you create and the platform(s) it's served on. You have control over the content; you're at the whim of platform algorithms.

So do your best to create the most resonant content by first truly knowing your audience. Then use gen AI as a surrogate for this audience. Have it react to your content and give you critical feedback. This is a way to make content efforts high value. Focus less on only trying to use generative AI to create more content, faster. That'll just add to the noise.

But the algorithmic platform paradigm also has to shift in order to bring forward relevant content. Users will need to volunteer preferences, and platforms will need to use gen AI to ‘listen’ and determine which content will matter most.

Pat Murphy

Founder and CEO at Murphy Cobb & Associates

Navigating the promise of endless content demands a balanced approach. While AI enables personalisation at scale, ensuring quality and relevance is paramount. Strategically deploying AI-enabled content requires clear objectives and continuous optimisation.

Harnessing AI's insights can enhance strategic creativity by identifying trends and audience preferences. However, there's a risk of asset overload and dilution of brand identity if not managed carefully, and everything needs to be in harmony with the original advertising idea.

Achieving this dream necessitates a shift towards dynamic, adaptive content strategies, integrating human expertise with AI's capabilities. I still believe there will be human curation. Industry-wide collaboration to establish standards and best practices is crucial. Emphasising quality over quantity, prioritising audience engagement, and refining measurement metrics will steer us toward personalised content that truly resonates.

Ultimately, success lies in striking a balance between automation and strategic discipline, ensuring each piece of content serves a purpose in fulfilling broader marketing objectives. And don't forget, we are still telling stories that need to connect, albeit at scale and personalised.

Lindsay Hong

CEO and co-founder at SmartAssets 

Executing personalisation to a granular and individual level is challenging, and marketers really need to think about the benefit versus the effort. Firstly, personalisation in a post-GDPR world is not an easy feat. And secondly, it’s difficult to be convinced of the incremental gains of going to the effort of creating hyper-personalised ads for one individual, in comparison to creating high-quality ads tailored for a specific audience. There’s so much poor quality content that fails on the basic principles of: good design ethos, brand naming, correct logos, consistent look and feel, effective translation, adherence to market and cultural norms. You can have a big uplift for an entire brand by just fixing these basic things without having to go down the route of ‘let’s tailor this one specific asset to just one individual’. If you look at the incremental gains of those efforts, marketers would do well to focus on wins at a higher level that are arguably going to be bigger because they're more scalable.

Personalisation drives the need for loads of assets. This is neither sustainable nor cost efficient. The conversation should be moving along from ‘how do I create copious amounts of content at scale quickly and cheaply?’ to  ‘how do I create the best content that's truly consistent with my brand, delivers on my objectives, and gives the audience what they're looking for in a consistent way across channels -  i.e. that's actually going to perform?’. So, rather than content for content sake - just to be present everywhere - we're getting under the skin of what good content looks like and investing in that.

Jim Rudall

GM of EMEA at Intuit Mailchimp

The competition for attention and engagement from customers is at an all-time high - and there is a growing emphasis on the importance of personalisation in cutting through this noise. According to our data, 73% of customers feel more valued when they receive personalised emails.

We see AI as an essential tool for marketers and creatives to leverage insightful data and unlock personalised experiences at scale. This accelerates new methods for marketers to connect to their customers in a more unique and dynamic way - helping teams save precious time and resources without compromising creative and operational control.

AI won’t won’t change the fundamentals of marketing, nor will it replace creatives, but it will provide more space to focus on what makes a brand stand out in the marketplace and reach its customers in an authentic manner. In other words, teams can focus human creativity towards the ‘why’ and ‘what’, whilst AI handles the ‘how’.

Ultimately, personalisation isn't possible without data, but companies must ensure it's handled responsibly and in line with privacy guidelines. Interestingly, our Holiday Report revealed that nearly three in four customers are comfortable with companies using their personal data if they are transparent with how they are using it.

Simon Spyer

CEO, data-driven and performance marketing at Iris

'Content shock' is a reality. The pursuit of endless content has inundated audiences, diluted brand messaging and created consumer apathy. The solution? Focus on niche communities, invest in content they want to participate with and leverage more channels to combat the effects of content shock.

Personalised content has to be meticulously crafted, precisely targeted, and strategically deployed to resonate with audiences effectively. AI can enhance this process by providing insights and automating and optimising distribution. But achieving the dream of hyper-personalisation at scale requires a shift in mindset towards quality, not just quantity. Marketers must focus on honing their creative instincts, leveraging AI as a tool rather than a crutch and ensuring every piece of content serves a strategic purpose and encourages participation with the brand.