Creative execution and production in marketing and advertising is still largely manual and fragmented. As a result, brands and agencies have a gap to address: creative teams and output haven't kept pace with content demands.
Artificial intelligence can help. Celtra uses computer vision and machine learning to accelerate the creation of content and creative campaigns.
We spoke with Celtra's founder and Chief Product Officer, Matevz Klanjsek, to learn more about this AI-powered solution.
In a single sentence or statement, describe Celtra.
Celtra is the creative automation company.
How does Celtra use artificial intelligence in its products?
Celtra uses computer vision to detect subjects/elements in creative assets:
- To automate treatment of assets (e.g. product shots), such as auto-cropping, positioning within design, auto-sizing, etc.
- To automate building creative taxonomies around assets and mapping it to brand-specific taxonomies.
Celtra also uses machine learning:
- To automate auto-generation of multiple ad layouts (eg. multiple ad sizes, variations of layouts, cross-channel creation) by learning from user adjustments to originally auto-generated layouts.
- To predict ad performance and provide creative guidelines for new campaigns based on historical on-market performance data by mapping it to brand-specific creative taxonomies.
We also use early experimentation with generative design (GAN):
- To auto-generate new creative completely from scratch using historical user-generated creatives in Celtra as a training set.
What are the primary marketing use cases for Celtra?
Celtra can be used for any type of high-volume digital advertising use case.
For example, an ecommerce company running campaigns to drive awareness, traffic, sales, and reconversions. A multinational CPG conglomerate that needs to activate global campaign toolkits in local markets who then use Celtra software to automatically scale brand-approved templates across local media plans. Another example use case is a grocery chain that needs to run volumes of localized and personalized weekly offers in the form of digital ads in the cities and regions it operates in.
What makes Celtra smarter than traditional approaches and products?
While workflow around media and ecommerce has consolidated and automated, creative execution and production is still largely manual and fragmented. As a result, enterprise advertisers, media businesses, and agencies all have a growing creative gap in their business. Creative data, creative teams, and creative output has not kept pace with today’s content demands.
The pain Celtra cures is the business reality of being dull and boring to consumers, employees, and partners. Businesses lose if they cannot connect with or communicate with consumers, or when they cannot cultivate creativity or nurture creative talent. Content, creativity, and learning are central to success. Businesses that are the most nimble and creative will thrive—those that are slow and boring will be disrupted or die.
Manual production, outsourcing to agencies, and hiring more people in-house to keep up with creative production is expensive and slow. With Celtra, brands can launch more content and campaigns faster and save money while doing so.
Are there any minimum requirements for marketers to get value out of Celtra? (e.g. data, list size, etc.)
We currently see large enterprise brands get the most out of our platform, if they’re running at least 500+ unique creatives on digital at any given time, then they’re likely to benefit from scaling and production with automation.
Who are Celtra's ideal customers in terms of company size and industries?
Large enterprise brands with at least $100MM in revenue and 500+ employees. On the brand side, we currently focus on retail, apparel, CPG, internet companies, and quick-serve restaurants. We also serve agencies and have a bespoke product for publishers.
What do you see as the limitations of AI as it exists today?
1. AI is good at solving well-defined functional or workflow related problems, but less good at solving creative problems that require imagination and out-of-the-box thinking. Advertising and marketing challenges are increasingly of creative nature and AI is struggling with that. For example, AI can help drive better performance within defined creative boundaries and with existing creative material, but not yet able to invent something new that would have a more significant impact.
2. When solving more complex problems with AI, the development is still relatively lengthy and expensive. Solving those types of problems with human labor is mostly still cheaper and faster, and still better and makes more economic sense than true AI.
3. Designers and developers of AI solutions are overwhelmingly white and male (80%-90%) resulting in strong built-in biases and effectively disqualifying AI as a solution to many very relevant and important marketing challenges. In many cases where such AI is still utilized, the results can be poor, outcomes negative and even damaging to brands.
What do you see as the future potential of AI in marketing?
Detection of gender, racial and cultural bias and stereotypes and providing creative guidelines based on created work or raw assets/focusing on non-obvious and hard to detect aspects of content, such as character roles, outfits, age and environment or screen time, direction of gaze and facial expressions.
Automating ad accessibility/using AI to automate validation and auto-correction of color contrast ratios for text, auto-generating text transcript overlays for video or audio-only content, auto-generation of alt texts for all visual content, etc.
Any other thoughts on AI in marketing, or advice for marketers who are just getting started with AI?
Ask yourself: how can AI help you go beyond executional tasks in marketing to play a role where it can drive actionable insights and guidance in the stories you tell.
Paul Roetzer is founder and CEO of Marketing AI Institute. He is the author of Marketing Artificial Intelligence (Matt Holt Books, 2022) The Marketing Performance Blueprint (Wiley, 2014) and The Marketing Agency Blueprint (Wiley, 2012); and creator of the Marketing AI Conference (MAICON).