Alison.ai’s Ellad Kushnir Matarasso on unlocking creative performance with AI

My 18 years in online marketing and advertising provided a well-rounded perspective on the challenges advertisers face with creatives and digital marketing. I’ve worked with companies like 888, Perion, McCann, Quality Score, and Taboola, always focused on finding solutions to drive digital campaign performance. A significant challenge I observed was the inefficiency of traditional creative testing. At one leading agency, managing over $100 million in annual media budgets, we constantly performed extensive A/B testing due to rapid creative fatigue. I recall a conversation with my parents, both computer engineers, over a family dinner, pointing out that A/B testing had been around for decades, making me question if it was truly the best practice for creatives.

This revelation precipitated my excitement about Alison.ai’s mission to fundamentally change creative optimisation. The traditional process of producing a large volume of creative content, generating multiple iterations for each concept, and spending significant test budgets to gather data is incredibly time-consuming and costly. When 80% to 90% of creatives typically fail, this approach is inherently wasteful and inefficient. With the rise of AI, particularly the technology Alison provides, user acquisition experts and creative teams can now make genuinely informed, data-driven decisions for creative testing, moving beyond mere best practices or random A/B testing, while saving a lot of time and money.

It’s true, there are many excellent companies in this field. What sets Alison apart is our focused approach on creative insights and our ability to translate raw data into actionable insights for advertisers. Our level of detail and granularity is truly exceptional; we employ over 10 different AI engines to identify virtually every element within each frame of a creative, from colours and text to sound, characters, facial expressions, logos, and concepts. We can identify thousands of elements from a single creative, depending on its complexity.

This deep data is our starting point. We cross-reference this raw creative data with campaign performance data to understand the exact impact individual elements have on performance. More importantly, we show advertisers how to combine these elements to maximise results. It’s not just about knowing that “blue outperforms red,” which you might learn from a simple A/B test. Alison can tell you that a specific combination, such as “blue with a female character, in an outdoor setting, with the logo on the top left, and a ‘Get Now’ call to action on a blue background”, will maximise your creative’s performance. These precise insights tell advertisers exactly what to add, remove, or replace in an existing creative to improve its effectiveness. Furthermore, when developing new ideas, Alison generates a brief, a script, a storyboard, and even a concept video, empowering creative teams to start with a data-driven approach based on actual performance data. This level of detail compels marketers to move beyond generic best practices and embrace an ‘elements-first’ creative strategy. It’s no longer about guessing; it’s about precision-driven creative iteration, where every component is strategically optimized for measurable impact.

We were fortunate to offer a solution at a time when AI became a significant buzzword. Studies show that over 80-90% of marketers are interested in implementing AI solutions, though only 20-30% actually do so, and even then, applying it efficiently into workflows remains a challenge. When we first started, our primary focus was educating marketers that there was a more efficient way to manage creatives, moving beyond the time and money spent on constant asset production and A/B testing. The narrative was about explaining that a different approach existed beyond what competitors were doing.

Today, the conversation has drastically shifted. It’s no longer about whether to use AI, but how to effectively implement it. It’s widely accepted that businesses not integrating AI into their tech stack will be left behind; it’s a matter of staying competitive and augmenting your workforce for greater operational efficiency. We’ve gained many leading global brands, and established strong collaborations with media platforms. The current dialogue revolves around how to best implement AI to make digital campaigns, overall performance, and creative usage more efficient. This transition has been remarkable over the past two and a half years, establishing a new standard for market responsiveness and customer engagement.

My experience across advertiser, agency, and platform roles provided a deep, three-dimensional understanding of the challenges within the digital marketing world, especially concerning creatives. This comprehensive perspective of the ecosystem we operate in has been a significant benefit. I also had prior experience in the creative analytics world, which helped me understand the existing solutions and their benefits.

A key focus for me, beyond fostering strong relationships with advertisers, has been building robust collaborations with media platforms. These platforms have a vested interest in their clients’ success because successful advertisers spend more. Alison.ai fits into this equation by helping advertisers understand how to succeed. While a platform might offer a similar solution, it would be reserved to that specific platform. Most advertisers run campaigns across multiple platforms and need a cross-platform perspective on creative performance. Our ability to connect to all platforms means advertisers get comprehensive insights, allowing them to compare and contrast performance across different campaigns and formats, overcoming the limitations of siloed, local solutions.

The idea that AI could generate creative content, especially video, was once met with scepticism. But the shift has been radical, driven by advertisers’ need for efficiency and cost reduction. While stunning videos can now be generated by machines, this alone isn’t enough for two critical reasons: current technologies aren’t built with advertising in mind (lacking considerations for branding, messaging, etc.), and more importantly, they aren’t connected to data. This means there’s no data-driven basis for the prompts used to generate these videos.

Alison.ai addresses this by connecting the dots. Our current capability involves taking creative data, breaking it down, analysing it, and providing actionable insights, including briefs, scripts, storyboards, and more. The crucial next step, in the very foreseeable future, is to generate the end result in the form of a data-driven, ready-to-use video ad that advertisers can directly upload to their campaigns. This integration of data and generation is a core differentiator.

Another key aspect that distinguishes Alison is our analysis, which is based on both the advertiser’s own creatives and their competitors’. Clients can select competitors within the platform, and we provide a bespoke library of their creatives, along with a gap analysis. This highlights what drives performance for competitors compared to the client’s own creatives, offering inspiration for new testing opportunities. This isn’t about random testing but about data-driven ideas based on what has moved the needle for leading competitors. Ultimately, we aim to provide actionable insights, catering to both data analysts and creative professionals, ensuring everyone knows the next steps for creative optimisation.

Frankly, in most cases, convincing marketers about the value of AI solutions is no longer necessary. The vast majority now understand that implementing AI is essential to stay competitive and avoid being left behind; it’s a matter of “when,” not “if”. Only a handful of more traditional marketers might still have reservations.

Regarding brand safety, it’s paramount for us. Our competitor analysis exclusively uses publicly available information, ensuring no exposure of direct competitors’ proprietary data. Furthermore, our algorithm employs over 10 different AI engines, which means each client effectively has a bespoke list of creative elements based on their own creative content. Even direct competitors might have very different creative elements, leading to distinct insights and element lists. This detailed analysis means I’m not concerned about identical results, even if everyone used Alison.ai, as different audiences, campaigns, and unique brand elements ensure distinct outcomes and adjustments. This uniqueness of brand identity and data-driven adjustments helps maintain differentiation even at scale.

Yes, Alison does have an API. We use it to pull data into the platform from various ad accounts across most major media platforms, including Google, Meta, TikTok, Snap, Unity, IronSource, Moloco, Applovin, Amazon, and LinkedIn. Clients can also pull data from Alison via our API into their internal BI or CRM systems, allowing them to further manipulate and process the data with their own internal datasets.

We do work with agencies, and it’s a common use case for them to use Alison across all their clients. However, our legal agreements place a strong emphasis on data safety and security, limiting how advertisers can use or sell the data. While agencies can leverage Alison’s intelligence for multiple clients at scale, the insights and outcomes for each client will still be significantly different. This is because differences exist in their offerings, campaigns, and even between various brands within the same company. The unique data and context for each client ensure distinct creative insights, maintaining client-specific value.

AI, in general, can assist in various ways, but over time, users increasingly seek AI models and solutions tailored for very specific use cases. For instance, if you want to book a flight, you wouldn’t use a general tool like ChatGPT or Gemini because they are not specifically trained on travel data. Similarly, at Alison, we are building a set of solutions specifically designed for managing and optimising creatives, identifying what drives performance, and pinpointing where fatigue starts to occur. Our models are specifically trained and focused on providing these solutions for marketers and creative professionals. Instead of sifting through vast data or crafting generic prompt, we enable them to make precise, informed decisions for their digital campaigns.

We are already putting this vision into practice. When you enter the Alison platform, you can iterate on a creative or develop a new concept. Alison will automatically generate a brief, a script, a storyboard, and even a concept video for you. This means that a creative team can begin their ideation process with a truly data-driven approach, focusing on elements proven to be impactful based on their actual performance data. The way Alison is built, the agent “hands you over what you need to know, what actions you need to take”. While you can still slice and dice the data if you wish, the agent provides instant solutions for quick fixes or direct guidance. The next significant step in the very foreseeable future is to generate the end result in the form of a data-driven, ready-to-use video ad that advertisers can directly upload to their campaigns and integrate into their creative stack.

A substantial challenge for advertisers is identifying, in a timely manner, when a creative begins to show signs of performance fatigue. This is difficult within standard ad accounts because a single creative asset might be used across numerous ads and campaigns, making it hard to gain a unified view of its overall impact. Alison.ai overcomes this by providing a deeper hierarchy, down to the creative and even the individual creative element level, which is simply not visible in standard ad accounts.

As a result, we can alert clients via platform notifications and email when a creative is starting to show signs of performance fatigue. Advertisers frequently tell me it takes them over a month to realise a creative is underperforming. This delay leads to significant wasted spend, which in turn impacts market share and campaign ROI Alison.ai’s timely alerts empower advertisers to take fast action, whether by shifting budgets to better-performing creatives or by optimising the fatigued creative based on our recommendations. You don’t have to discard a creative; you can adjust it to perform better, ensure every dollar invested in creative delivers maximum return, and sustains brand visibility.

When advertisers run thousands of creatives, often due to extensive A/B testing because they lack precise insights, Alison.ai offers a more efficient alternative. We show you exactly what adjustments are needed for a creative, eliminating the need to create five iterations for testing. This helps to reduce production costs and the necessity of managing an excessive number of creatives in campaigns, allowing focus on those that genuinely impact performance. While our immediate focus is on data-driven generation of ready-to-use video ads, the logical progression, and what we are actively laying the groundwork for, is the full automation of creative adjustments, ensuring continuous, adaptive optimization at scale.

I absolutely agree with the prevailing view that “creative is the new targeting”. Looking back over the past 20 years, we’ve moved from manual search and display advertising to audiences and lookalikes, then to social media, and now we face limitations introduced by changes like iOS 14.5 and reduced third-party data. This means that advertisers’ ability to impact performance from the media side is now quite limited. Research indicates that the main reason for a campaign’s success is 89% creative, with only 5% attributable to media and 6% to brand. Therefore, creative remains the primary tool and ammunition for influencing campaign performance. Alison.ai helps advertisers bridge the gap between media and creative teams by equipping them with the capability to automatically generate a brief, a script, a storyboard, and a concept video. This enables creative teams to embark on their ideation process with a genuinely data-driven approach, focusing on elements that have proven impactful based on their actual performance data, precisely pinpointing the direction for new concepts. By providing a deep understanding of the impact each individual element within creatives has on overall performance, and, more importantly, how to combine these elements to maximise results, Alison moves beyond basic A/B testing. This approach dramatically reduces the need for the wasteful, time-consuming, and inefficient process of producing and testing countless iterations.