Kostas Hatalis: Designing multimodal AI for marketing

The CEO and co-founder of GoCharlie.ai is working with SRI to reimagine how digital content is created.

Kostas Hatalis, CEO and co-founder of GoCharlie.ai, is collaborating with SRI’s Center for Vision Technologies to help marketing professionals augment their workflows. Kostas and the GoCharlie.ai team will leverage model customization and multimodal model extensions from SRI and are working closely with the institute’s experts to further develop their Large Language Model (LLM) into a revolutionary multimodal AI model that will transform the field of marketing.

Here, he discusses his unshakable passion for AI and the family connection that sparked his idea:

My interest in AI was really ignited in high school when I read a book called “The Singularity is Near” by Raymond Kurzweil, which deeply inspired me. This was back in the late 2000s, right after the “AI winter” of the 90s, when the field seemed to move at a snail’s pace. As I approached grad school, it felt like we were still moving slowly, especially in neural networks; some of my university professors told me that machine learning was essentially a dead field and that I shouldn’t do my PhD in it. But I believed AI was the future. I got my PhD in electrical engineering with a specialty in deep learning, but I never intended to become a professor or work in a lab.

I graduated in 2019, and my co-founder Despina Christou and I launched our generative AI company GoCharlie.ai in 2022. We’d seen the success of GPT-3 and wanted to build improved products, particularly in the field of marketing. My sister is a marketer, and I’d watched her struggle and be overwhelmed trying to deal with all the parts that go into creating personalized social media content for multiple clients. We realized that text is just one component of writing blogs and posts; images, video, and audio were the future.

That’s when we connected with SRI. We met with Todd Stavish at SRI International’s venture’s group and discussed extending our model into something more powerful that could provide not just marketing copy, but these other components as well. Now we’re working with a great team, including Ryan Lewis within the group and Karan Sikka and Ajay Divakaran from SRI’s Center for Vision Technologies to create a multimodal – or multimedia – model that we believe will transform the field of marketing. It feels like a match made in heaven.

My grand vision for this company is to take the parts of a job that can be boring and repetitive and automate them so creative workers can be free to do the things they love and that makes them happy. My sister, for example, has a passion for photography. Using our large learning model, Charlie, she could spend most of her time taking the photos she loves, then let AI use those photos to create personalized marketing content. This technology has the potential to be transformative in copywriting and content creation for fields ranging from HR and finance to healthcare and education.

Today AI is creating basic content, and I anticipate that tomorrow there will be hyper-personalization, where individuals will be able to create any content they want, be it cartoons, videos, movies, music, poems, or books. Even bigger than that will be real automation, where AI will make autonomous decisions and be able to run complex tasks, like creating and running an email campaign. It will be truly disruptive in every field, but I look forward to a future where many of the work stresses our parents and grandparents endured will be a thing of the past, and creative workers in many fields will be free to focus on the work they feel passionate about.

Read more from SRI