Artificial intelligence is changing the way brands approach advertising on Meta, influencing how audiences are identified, how creatives are developed and how campaign performance is evaluated. As AI becomes more deeply integrated into the advertising ecosystem, marketers are navigating new opportunities to improve campaign relevance while addressing persistent challenges around costs, measurement and differentiation.
Meta’s advertising business reflects the growing commercial significance of these developments. The company reported advertising revenue of $63.1 billion in the third quarter of 2026, a 26% year-on-year increase. Its investments in AI have contributed to improvements in recommendations, engagement and advertising monetisation, alongside factors such as higher ad prices and broader advertising demand. (Source: Exchange4Media)
For marketers advertising across Facebook and Instagram, the implications extend beyond the platform’s financial performance. AI is influencing the relationship between audience targeting and creative strategy, changing how campaign results are interpreted and raising new questions about what drives effective advertising. The challenge is to understand where AI adds value and how that value translates into meaningful business outcomes.
Meta Ads and AI Optimisation: How AI Is Changing Campaign Performance for Marketers
Artificial intelligence is becoming central to how Meta Ads identifies audiences, evaluates advertising creatives and delivers campaigns. As Meta continues to develop its AI-powered advertising systems, marketers are seeing changes in how advertisements reach potential customers, how campaign performance is optimised and how results are measured.
The shift is particularly evident in Meta’s ad recommendation systems, which use machine learning to assess the relevance of advertisements to individual users. Technologies such as Andromeda and the Generative Ads Recommendation Model (GEM) are designed to improve how Meta identifies suitable ads and predicts which advertisements are likely to generate a response. Meta’s October 2026 update on GEM also highlights the growing role of AI in improving advertising performance across its platforms. (Source: Meta Engineering)
For marketers, these developments have direct implications for campaign planning and performance. AI is influencing the relationship between audience targeting, creative quality and conversion outcomes. However, while the technology can improve how advertisements are matched with users, advertisers still need to determine whether these improvements translate into lower acquisition costs, incremental sales and stronger returns.
How AI Is Changing Ad Delivery on Meta
Meta’s AI-powered advertising system evaluates a wide range of signals to determine which advertisements are relevant to individual users. Its Andromeda system supports the ad retrieval processMeta’s AI-powered advertising system evaluates a wide range of signals to determine which advertisements are relevant to individual users. Its Andromeda system supports the ad retrieval process, narrowing a large pool of potential advertisements to a smaller set of relevant candidates. Other recommendation models, including GEM, help determine which advertisements are ultimately shown.
This approach allows Meta to consider more complex relationships between user interests, advertising content and predicted outcomes. Instead of campaign performance depending only on the audience parameters selected by an advertiser, the platform’s AI systems play a significant role in identifying people who may respond to an advertisement.
Chandan Kumar of Minimalist describes how this has influenced the brand’s approach to targeting.
“Since Meta’s Andromeda update, we’ve seen a shift from detailed manual targeting to using creatives to help AI understand who the target audience is. This has made audience discovery easier. AI has also reduced the manual effort required to manage budgets across multiple campaigns, as Meta can allocate more budget to better-performing ad sets.”
His experience highlights an important change for advertisers: creative content has become an increasingly significant input into campaign performance. The message, imagery and product proposition can help Meta’s systems identify relevant opportunities to show an advertisement.
For marketers, this makes the relationship between audience understanding and creative development particularly important. AI may help identify potential customers, but the effectiveness of the campaign still depends on whether the advertisement communicates a relevant proposition and gives people a compelling reason to respond., narrowing a large pool of potential advertisements to a smaller set of relevant candidates. Other recommendation models, including GEM, help determine which advertisements are ultimately shown.
This approach allows Meta to consider more complex relationships between user interests, advertising content and predicted outcomes. Instead of campaign performance depending only on the audience parameters selected by an advertiser, the platform’s AI systems play a significant role in identifying people who may respond to an advertisement.
Chandan Kumar of Minimalist describes how this has influenced the brand’s approach to targeting.
“Since Meta’s Andromeda update, we’ve seen a shift from detailed manual targeting to using creatives to help AI understand who the target audience is. This has made audience discovery easier. AI has also reduced the manual effort required to manage budgets across multiple campaigns, as Meta can allocate more budget to better-performing ad sets.”
His experience highlights an important change for advertisers: creative content has become an increasingly significant input into campaign performance. The message, imagery and product proposition can help Meta’s systems identify relevant opportunities to show an advertisement.
For marketers, this makes the relationship between audience understanding and creative development particularly important. AI may help identify potential customers, but the effectiveness of the campaign still depends on whether the advertisement communicates a relevant proposition and gives people a compelling reason to respond.
AI Optimisation and the Cost of Acquiring Customers
AI-driven optimisation is intended to improve the relevance and performance of advertisements. Meta’s recommendation systems use machine learning to predict which ads may be most useful to individual users, while its advertising tools provide features that help advertisers evaluate campaign performance and improve results.
Yet better delivery does not necessarily mean better returns. A campaign may generate more conversions while still becoming less profitable if acquisition costs rise or the customers acquired deliver limited commercial value.
This distinction is particularly important for smaller brands, which often operate with tighter advertising budgets and less room to absorb inefficient spending.
Saloni Pravindra of DEHI explains why campaign efficiency remains a key consideration.
“For a small brand like DEHI, every rupee of ad spend matters. The value of Meta’s AI-driven ad optimisation is in helping match our products with people more likely to buy and directing budgets towards what is working. It reduces the manual effort involved in testing audiences and placements, allowing a lean team like ours to focus more on products and creatives. Ultimately, we judge the effectiveness of Meta’s AI optimisation through conversions and customer acquisition costs.”
Her perspective underlines the need to evaluate AI optimisation through commercial outcomes rather than platform metrics alone.
For marketers, the relevant questions include whether cost per acquisition is improving, whether conversions are generating revenue and whether the customers acquired are likely to make repeat purchases. These measures help establish whether AI-driven improvements in ad delivery are creating meaningful business value.
The challenge is to distinguish between a campaign that performs well within Meta’s reporting environment and one that contributes to sustainable business growth.
Why Creative Quality Matters More in AI-Optimised Meta Ads
AI optimisation is also changing the way marketers approach advertising creatives. Meta’s systems can assess and rank a growing range of advertisements, while its generative AI tools allow advertisers to create or adapt images, videos and other creative assets.
This creates opportunities for marketers to test different messages, product presentations and formats. A brand can explore multiple creative approaches to understand which ones generate stronger responses from its intended audience.
However, the availability of more creative options does not guarantee better campaign performance. The effectiveness of each advertisement still depends on its relevance, clarity and ability to communicate the product’s value.
In an AI-optimised advertising environment, marketers need to consider not only how an advertisement looks but also what it communicates. Different creative executions may appeal to different audiences, and a message that generates engagement may not necessarily lead to purchases.
The implication is that creative testing needs to be connected to the campaign objective. For a conversion-focused campaign, marketers must assess whether creative variations are generating meaningful purchases at an acceptable cost rather than judging performance solely on clicks or engagement.
AI can help advertisers explore creative possibilities, but the strategic direction behind those creatives remains essential.
The Attribution Problem: Measuring the Real Impact of AI Optimisation
Even when AI improves campaign performance, determining its actual contribution to sales remains a challenge. Meta’s reporting helps advertisers understand the conversions attributed to their campaigns, but attributed conversions do not automatically establish that the advertisement generated an additional purchase.
Customers frequently interact with brands through multiple channels before making a purchase. Someone may discover a product through an Instagram advertisement, return through a direct website visit and complete the transaction after receiving an email or WhatsApp message. Evaluating the role of the original advertisement requires a broader view of these interactions.
Chandan Kumar of Minimalist highlights this challenge in Meta’s attribution approach.
“Meta’s first-user-click attribution makes it difficult to assess incremental sales. A customer may already know the brand or may have converted via WhatsApp or email, yet Meta may still attribute that conversion to its ads. This can affect how we evaluate customer acquisition cost (CAC) and overall profitability.”
The distinction matters because campaign attribution influences how marketers interpret performance and allocate budgets. If a platform attributes a purchase to an advertisement without establishing whether the sale was incremental, the reported results may not fully reflect the campaign’s contribution.
For marketers, evaluating AI optimisation therefore requires more than monitoring the figures available in Ads Manager. Platform-level reporting should be considered alongside actual sales, customer acquisition costs and, where appropriate, incrementality testing.
The objective is to understand whether AI-driven ad delivery is helping the business acquire additional customers and generate revenue that would not otherwise have occurred.
What AI Optimisation Means for Meta Advertisers
AI is becoming an increasingly important part of Meta’s advertising infrastructure, influencing which advertisements are considered relevant, how they are ranked and how creative variations can be evaluated. These developments offer marketers new opportunities to improve campaign relevance and performance, but they also require a more considered approach to measuring results.
Three factors remain particularly important.
Audience relevance: Marketers need to understand how creative content, audience signals and campaign objectives work together to reach potential customers.
Creative effectiveness: More creative variations can create more opportunities for testing, but the focus must remain on messages that generate meaningful business outcomes.
Performance measurement: Improvements in conversions or cost per result need to be evaluated alongside revenue, acquisition costs and incremental sales to understand their commercial value.
Ultimately, AI optimisation does not remove the need for marketing expertise. It changes the context in which marketers make decisions. Understanding how Meta’s systems influence ad delivery, developing relevant creatives and interpreting campaign results remain essential to improving performance.
As Meta continues to develop its AI-powered advertising systems, the advantage for marketers will come from understanding how these systems work, using their capabilities effectively and measuring whether the resulting improvements contribute to profitable growth.













