Trending Articles
Trending Topics
Trending
When it comes to scaling a team’s creative impact with AI, Constanza Ghelfi is firmly at the forefront. As Chief Product Officer at Making Science, she’s leading the charge with Creative Hub, an advanced AI platform to help marketers create, measure and optimise campaigns faster than ever. For Constanza, this isn’t about using AI as a gimmick, it’s about reshaping every step of marketing campaigns, all without sacrificing authenticity.
Constanza’s approach uses AI to tailor more than ads, with personalisation even extending to landing pages. The creative assets themselves can shift based on real-time data. Consider the targeted ads offered when searching “eco-friendly running shoes”, she told us. These will be completely different to those when searching “fastest marathon trainers”, despite pushing people to the same product. So identical item, totally different story.
“What matters most is that it feels relevant but not invasive,” Constanza explains, and “that’s when people lean in.” AI enables marketers to hit that sweet spot by adapting assets without overwhelming the audience or crossing boundaries. Constanza’s team uses AI to intelligently generate versions in real-time, crafting experiences that feel natural and context-aware.
AI may handle the heavy lifting of this process, but Constanza sees its true value as a collaborator. “It can remix and optimise,” she explains, “but it’s not going to understand brand tone, humour or cultural nuance on its own. It’s a very useful virtual colleague, but it does not substitute the whole team.”
With this balance in mind, we asked about the specific tools making such personalisation happen…
Right now, ad-machina has to be the one I’m most excited about. And I’m not just saying that because it’s ours, it really is changing the way we work. A few years ago, campaigns were like setting a ship to sail. You’d launch with a nice set of creative assets, and then you’d basically watch from the shore, hoping the wind stayed in your favour. If you needed to make changes mid-journey, it was a whole process — brief the creative team, wait for new files, then finally push them live.
Now, that whole cycle is on fast-forward. We can take a single creative framework and spin it into hundreds of tailored versions in seconds. Different formats, visuals, tones, copy, you name it. And it’s not guesswork. Every change is driven by live performance data: which headline is getting the clicks, which colour palette is resonating in Spain versus Germany, even which call-to-action works best at certain times of day.
The beauty is, it’s not replacing us. We’re still in the driver’s seat, making sure every variation feels authentic and appropriate. This way, we maintain control of the brand identity at all times, manually checking for nuances that may or may not fall within the brand guidelines. AI gives us speed and flexibility. We make sure it hits the right nerves. That’s the sweet spot for me.
AI has completely changed what personalisation means. The old version, which was mostly brute individual or group targeting through ads, is starting to feel archaic. Hyper-personalisation, instead, has been in play for years now, allowing us to go further across formats and mediums. This is where the message, the creative, and the landing page, across all channels and even LLMs, can shift based on who you are and your context in that exact moment.
Let me give you an example. Someone searches for “eco-friendly running shoes”. They’re going to be served an ad with a creative that focuses on sustainability; maybe the recycled materials, maybe how the company offsets its carbon footprint. Another person is interested in “fastest marathon trainers”, and they’ll get an entirely different ad, tailored for runners. Same product range, totally different story. And it’s not just ads; once they click, the landing page can reflect that same theme. This goes beyond search as well. These technologies can adapt creative assets across different platforms and every format; whether it’s the copy or videos on social channels, or the imagery on display campaigns online.
It’s not about manually building 200 campaigns or tons of micro-audiences. It’s intelligently generating and adapting those variations across search, social, display, etc. based on live performance signals, using advanced AI systems. And because they’re learning constantly, the personalisation gets sharper over time.
What matters most is that it feels relevant but not invasive. That’s when people lean in. AI makes that level of precision possible, which makes it essential for cutting through the noise.
When AI is used well, it can make a brand feel incredibly in sync with its audience. You get the right content, at the right time, and it’s actually useful. That builds trust, because it feels like the brand is paying attention in a good way. Not stalking you, but understanding you.
The problem is when AI becomes a “black box”, with no humans kept in the loop. And I think that’s one of the biggest risks we’re seeing right now. A lot of embedded AI inside major platforms does amazing work with bidding, targeting, and even creative placement, but it won’t tell you why it’s making those choices. You give it your budget, your goals, and it just… does its thing. That might deliver good numbers, but if something goes out that’s off-brand, tone-deaf, or even unintentionally biased, you might not catch it until it’s public. And by then, the damage is done.
For me, the fix is simple but non-negotiable: keep humans in the loop. We use AI to test and learn at scale, but we always have visibility on why something’s performing. That way, we can pursue what works without losing control of the brand’s voice and identity, or the trust we’ve built with our audience.
Oh, massively. The way we build content and creative now is almost unrecognisable compared to a few years ago. Back then, it was a straight line: you’d plan the campaign, design the assets, launch, and then maybe tweak a few things later if you had time. It was all very fixed.
Now, we think of the creative as something alive. We design in modules — headlines, visuals, calls-to-action — that AI can adapt and recombine for different audiences, formats, and platforms. If something takes off, we can scale it instantly. If it falls flat, we pivot in hours, not weeks. The time and resources required to design and adapt creatives have been slashed significantly.
What’s changed most for me is the feedback loop. AI gives us performance signals in real time; for instance, “This image is performing great with 25–34-year-olds in Italy” or “This headline is underperforming in the US”. That insight lets us evolve the creative constantly, tailoring it towards the exact individual’s interests and intentions in-flight, rather than waiting for the next big campaign cycle. With advanced performance AI tools, we can react in even more specific ways; like making sure an outdoor ad reacts to sunlight conditions, or that a pool doesn’t appear in a creative for a region where it would feel out of place. It’s those kinds of details that make the outputs smarter and more relevant.
But AI doesn’t decide the soul of the creative. That’s still us. It can remix and optimise, but it’s not going to understand brand tone, humour, or cultural nuance on its own. It’s a very useful virtual colleague, but it does not substitute the whole team.
I have a saying: “Don’t use AI — work with AI”. That’s how I see the future of marketing. The tools are only getting smarter, but that doesn’t mean we hand them the keys and walk away. AI is taking over a lot of the repetitive stuff: bid adjustments, asset resizing, endless A/B testing, and, honestly, I’m fine with that. It frees us up to do the thinking. The real work now is in orchestration: knowing how to brief AI so it delivers something useful, knowing how to meaningfully interpret what it spits out, and knowing how to steer it towards the bigger brand vision.
That means marketers will need sharper skills in some areas, such as process design, to optimise workflows and the structure of AI agents. There is also a need for data literacy. If you don’t understand what’s driving the results, you can’t guide the system. Creativity is still critical, but so is the ability to adapt ideas fluidly across formats and platforms using facts and figures.
So yes, AI will do a lot of the heavy lifting. But we’re still the ones setting the direction, making the calls, and bringing the human judgment AI can’t replicate. That’s not going away; if anything, it’s becoming more important, and marketers will need more training around it.
You can’t get personalisation right without getting privacy right first. Otherwise, it’s encroaching, and it will have the opposite effect to what’s intended. For us, it all starts with the data foundation; collecting first-party data in a transparent, compliant way, and making sure any third-party data we use meets the same ethical and legal standards. There can be no shortcuts there.
Once that’s in place, AI can still work wonders without being invasive of consumers’ privacy. You don’t need to know every single thing about someone to be relevant. Aggregated, anonymised intent signals can tell you plenty; for example, if someone’s looking for “family holiday deals” or “luxury spa breaks”. That’s enough to tailor a message without crossing any privacy lines.
And again, we always keep human oversight on every AI output. If something feels wrong for the brand or for the audience, it doesn’t go live. Full stop. And transparency with consumers about how their data is used is key, because trust is fragile. If people feel safe and see value in return, they’re generally happy to share data. The balance is making sure they always feel both safe and valued. AI can help with the latter. The former is on us.
