AI is influencing the entire digital advertising landscape, presenting an array of opportunities and challenges. In this blog, I’ll be giving a quick overview of the value AI can bring to Display & Video advertising today.
For us, AI is no longer just food for thought; it’s an active force reshaping how we approach campaign strategy. There are risks involved, which will be touched upon, but the focal point will centre on how we as marketers can use AI to make us more effective.
A quick definition before we get started
AI is often used as a blanket term to cover a wide range of tools. As marketers, it’s important to be able to distinguish between the specific “types” of AI. Day to day, marketers can typically encounter generative AI such as Large Language Models or LLMs (think ChatGPT) and narrow AI, which powers tools like Google AI in DV360. While generative AI is best suited for creating content and analysing natural language, narrow AI is designed for specific tasks such as automated bidding and budget optimisation. Both come under the AI umbrella, yet provide different services and value to marketers.
How AI has evolved within Display & Video advertising
Just a few years ago, Display & Video advertising was still largely driven by manual workflows. While DSPs offered automation in terms of buying and delivery, much of the strategic work required constant hands on management. You could easily spend hours pulling reports, analysing performance and then spend even longer manually adjusting bids or exclusions based on what you observed. This was a clear pain point for many and still is today.
Google began introducing automated elements to address these pain points with AI features that could adapt and learn from previous campaigns and real-time signals, without the need for human intervention. Automated bid strategies, such as target CPA and max conversions in DV360, began to replace manual bid adjustments by using real-time signals to make more efficient decisions. Lookalike modelling became another significant development as AI could now analyse first-party converters and automatically build segments of new users with similar traits, improving reach and relevance with minimal setup. These innovations marked the turning point and beginning of AI being used to amplify campaign performance.
What the landscape looks like now
Machine learning is helping advertisers move beyond traditional targeting methods. Instead of relying solely on static audience lists or broad demographics, real-time machine learning models can dynamically build and refine audience segments based on signals such as browsing behaviour, search patterns, and app usage. For instance, an ecommerce brand can now identify high-intent users before they even land on the brand’s website, thanks to these predictive models.
AI agents are autonomous systems that perform tasks like creative optimisation and budget reallocation using real-time data and set goals. These agents currently power many DSPs and ad tech tools such as Google’s DV360, which uses AI-driven automated bidding to optimise conversions and viewability with limited manual oversight. While they work within predefined bounds and still require supervision, AI agents are a widely adopted tool within many DSPs.
Natural language processing (NLP) is also changing contextual targeting. Traditional contextual advertising often relied on basic keyword matching, using exact or broad match logic to serve ads on pages containing specific words. But this approach missed the nuance of language and often led to mismatched placements. NLP with human-like semantic understanding can analyse page content in real time. It interprets tone, sentiment, and deeper meaning, going beyond keyword presence to understand true context. This enables advertisers to place ads in environments that truly align with their brand values and messaging. Unlike fixed targeting strategies, contextual NLP models continuously adapt to the live content users are engaging with, helping ensure ads are served at the right moment and in the right environment. The result: more relevant impressions, improved performance, and reduced media wastage.
Large Language Models (LLMs) are a game changer for creative. Rather than relying on human assumptions about which message or format will perform best, LLMs analyse patterns across thousands of impressions to identify what’s working and why. These models can rapidly generate and test multiple ad copy variations, speeding up personalisation and removing creative bottlenecks. Crucially, machine learning also helps detect creative fatigue, keeping campaigns relevant, engaging, and efficient throughout their lifecycle.
With these tools integrated into your strategy, marketers can continuously learn and improve performance.
You are the pilot, and AI is your co-pilot
The machine learning models that have been highlighted throughout rely heavily on historical data and are currently limited by what has already occurred. As a result, they may miss broader brand context, cultural nuance or real-world shifts not captured in the data which could lead to biases from the datasets they’re trained on. This means human oversight is not only helpful, it’s essential. Rather than replacing marketers, these AI tools should be seen as a powerful co-pilot. It can surface insights we might overlook, automate time-consuming tasks and bring objectivity to performance analysis, but it’s your expertise that sets the direction. This is done by defining the brand strategy, understanding the customer beyond the data and making decisions that reflect values, goals and long-term vision.
Creative teams remain essential. While AI tools can excel at scale and speed, they lack the strategic insight, brand nuance and emotional intelligence that creative teams bring. Creative teams are critical for crafting strong concepts, ensuring brand consistency, and bringing storytelling and tone to life in ways that resonate on a deeper level. It’s important to remember that AI can help execute and optimise, but it’s the creative vision that gives campaigns meaning and impact
It is increasingly becoming a useful tool for in-house marketing teams, not as a replacement for expertise, but as a way to enhance it. One of the most practical applications right now is in automating repetitive tasks such as reporting and routine optimisations. By reducing the time spent on manual workflows, marketers can focus more of their time towards higher value activities such as strategic planning, creative development, and experimentation.
However, these streamlined workflows aren’t something you should just set up and forget about. Any outputs still need to be reviewed with brand context in mind. Currently, we believe it is best used as a support system helping marketers work more efficiently, supporting our decision-making rather than replacing it.
What updates can we expect to see added to DSPs soon?
- Improved forecasting with the aid of generative modeling and predictive analytics:
We have already seen some DSPs develop their own forecasting tools. As generative modeling and predictive analytics continues to develop, so too will the usability and value we can derive from forecasts. This could be the ability to run simulations to forecast how different bidding strategies might perform under varying market conditions. Of course this comes with caveats, such as the forecasts only being as good as the data fed into them, however it is an exciting development nonetheless.
- Natural language interfaces for easier reporting:
Soon, you’ll ask questions like “Which creatives performed best last month?” and get immediate answers, without the need for manual reporting and crawling through large datasets
- More accessible and automated experimentation, at scale:
With tools like AutoML, you will become more efficient at setting up and running experiments. Giving you the ability to test new creatives, audiences or bid strategies more frequently will allow you to run experiments at scale and adopt a more test and learn approach to your campaigns.
AI has clear value to us marketers and the opportunities it provides will only improve over the coming years. As marketers we should embrace the opportunities provided by AI technology and learn how to leverage it to improve performance because as the technology matures, we expect the gap to widen between those leveraging AI tools effectively and those who are not.
