The Center for Data Innovation spoke with Hikari Senju, Founder and CEO at Omneky. Omneky is a San Francisco-based startup that uses AI to generate advertisements for social platforms. The platform collects and analyzes data from a user’s past ad campaigns, then generates ads positioned for success using deep learning algorithms and OpenAI’s DALLE-2 and GPT-3 models.
Becca Trate: What problem prompted the founding of Omneky?
Hikari Senju: I saw how hard A/B testing was at scale and how much money was being spent on inefficient ads after selling my first company and becoming Head of Growth at the acquiring company. This inspired me to build a platform that, by combining the ability of AI to quantify design and to generate creative content, would generate AI-powered content to scale A/B testing. I launched the Omneky platform in 2020 and have since grown the company into a platform that helps businesses grow by generating and managing content at scale.
We believe that great ideas come from all places and we want to create the tools to power the next generation of entrepreneurs all over the world. By lowering the cost of advertising effectively online, we are democratizing growth and empowering small businesses to grow. We’re lowering the barrier to entry for advertising, allowing businesses to successfully connect with customers online no matter the size of their budget. Our mission is to empower human creativity and democratize growth through AI.
Trate: What is AI-generated art and advertising?
Senju: Text-to-image generation has been in vogue recently with OpenAI’s DALLE and Stability AI. These are foundation models that generate images based on text prompts. AI generated art is the process of creating art by engineering different text prompts into these models, until you get a novel piece of artwork.
Omneky utilizes these foundation models by combining the generative capabilities of AI with real-time marketing data to generate personalized ads at high volume. With AI, it is possible to generate an ad targeting every person individually, based on their design and messaging preferences.
Trate: How are AI and machine learning used to generate successful advertising campaigns?
Senju: Omneky’s proprietary software layers on computer vision to provide correlative relationships between ad features and key performance indicators (KPIs). Our platform gathers data on optimal character counts, color schemes, sentiments, and more, which helps our customers detect important trends. Then, leveraging the capabilities of machine learning, our platform predicts the impact that an advertisement will have on a customer’s KPIs, offering actionable insights that help them make informed advertising decisions in a fraction of the time. AI and machine learning provide customers with the data they need in real time to run effective advertising campaigns targeted to their specific consumer demographics.
Trate: What roles do AI and automation play in generating advertising copy?
Senju: AI and automation have become increasingly powerful tools in the development of photo and video assets as well. We’re now entering an era where machines are better at writing and creating images than most humans. Our AI platform integrates real time marketing data with foundation models as well as APIs like DALL-E 2, Stability AI’s Stable Diffusion and GPT-3 to generate personalized ad copy and ad media. However, no ad is launched unless it’s explicitly approved by humans. This way, we empower humans to harness the power of generative AI while maintaining control of their brands.
Trate: As the industry shifts from traditional, static ads to content-based advertising on social media, how do you expect AI-generated advertising to change?
Senju: There is a saying that content is king, distribution is temporary. Advertising will always change based on what distribution channels capture the most attention. Currently, there is a lot of excitement around streaming ads on Netflix, Disney, Fubo, and Tubi. In the future, Omneky plans to utilize AI to generate compelling video advertisements targeted these new media buying opportunities. Though text-to-video generation is not great yet, we expect to get there within a year.
We’re also seeing an increasing fragmentation of the media environment and know AI-based advertising and insights will help marketers make sense of what channels and content are driving ROI.
Beyond advertising, we’re in a Cambrian explosion of open-sourced generative AI. Companies like Hugging Face and Stability have taken Generative AI out of the pandora’s box to the general public, and it is no longer just a plaything for large technology companies. I’m seeing top researchers at Google and Facebook depart corporate labs to join startups and the open-sourced movement since there is now greater energy outside these companies than within. I expect that the majority of advertising and marketing content will be generated by an AI in the next five years and that there will be stories created by AI that will compete in terms of cultural impact with the greatest masterpieces in cinema and literature.