When AI hit mainstream at the end of 2022, we, and so many other businesses, suddenly had to move very quickly.
At the time we were a platform for easily customizing illustrations online, and were building a wider platform for end-to-end design, from initial branded suggestions to publishing on your platforms.
However, with OpenAI speeding things up significantly, we needed to rethink our timelines and how we could start integrating LLMs into our roadmap.
I interviewed our CEO, Francois Arbour, to get some quick thoughts on this acceleration, and how he thinks about AI when leading designstripe.
When did you first notice AI picking up, and do you remember the moment you realised it would be a bigger deal this time around?
The question is interesting because this has happened before. We thought we had an AI revolution in, I would say, 2014, 2015, and it was kind of a flop back then. Everyone thought adding .ai to their domain names was going to grow valuations by millions, and for a while, it did. People had .ai domain names, and investors were keen. But we lacked the essential elements, like the GPU accelerated computing power we have today, and LLMs were just starting out. They weren't good and were too expensive, so everyone reverted to traditional machine learning.
When we started designstripe, AI was part of our plan from day one. The idea was to build a virtual agency. Our initial pitch deck shows this, promising illustrations, designs, and everything else needed - essentially functioning as a virtual agency. What’s surprising is how quickly this happened. We expected a four to five-year timeline, but it unfolded in just two. The speed and quality of advancements were unprecedented. Early AI tools like DALL-E or Midjourney weren't practical yet, similar to the current state of AI for 3D models - interesting but not quite usable. But AI’s progress, both in terms of speed and quality, outpaced our expectations by two to three years, surprising even the creators of these tools.
What was your decision-making and thought process like, given we were an illustration customisation platform at the time? How did you make the decision to pivot?
The turning point for me was when we started fine-tuning our own models, enhancing our vector illustrations with AI. The results were getting good. Then platforms like Midjourney and Stable Diffusion entered the market, signaling a major shift. The demand for visual assets grew, and people wanted immediate results. This was challenging for traditional, template-based models. AI’s capabilities in vector illustration weren’t perfect but were improving.
About two years ago, I realized AI's potential was real. And then ChatGPT, especially GPT-4, was a revelation. As a coder, the speed of its advancement was hard to believe; we were entering a new era with AI. The decision for designstripe was clear, as for many startups: embrace AI. The improvement in production cost and speed was undeniable. It felt like transitioning from dial-up to broadband internet.
AI has become a utility, necessary like water. There are many providers, each with unique strengths, whether in coding, copy generation, or science and math. At designstripe, our pivot wasn’t just towards AI but evolving our virtual agency model. We focus on understanding a business's needs and producing content that aligns with their brand. AI enables us, but our pivot is towards offering comprehensive services, acting as a virtual agency.
Our journey with AI is about responding to market needs, not following trends. By adopting the virtual agency model, we accelerated our plans, combining what would have been separate stages into a unified, AI-enhanced service. Our mission remains to provide beautiful marketing assets.
How do you keep up in such a fast-moving industry, and how would you recommend other business owners view AI?
Staying ahead in this fast-evolving industry means being hands-on. I've spent thousands of hours coding with APIs as CEO, understanding these new technologies. It made me an enthusiastic advocate for AI, constantly encouraging our team to integrate AI into their work. This approach isn’t about following trends; it's about efficiency and competitiveness.
So every day the team asks themselves, how did I use AI today? And it's not about just because it's a trend or whatever. It's because I know that it makes everything much, much faster. So I want my team to be on it and to really understand the power of these tools. And I realized by doing that, that it's pretty intimidating for a lot of people because it, first of all, it's kind of scary to replace some of the things you were doing before and maybe you were proud of doing. Like, let's say you're a great copywriter and you have a certain way to do copy. All of a sudden, you have this new tool that you have to work with; it's not replacing you as a copywriter completely, but it's definitely replacing your tools that you were using before. Maybe you were using Grammarly or Hemingway or some of these tools, or maybe it was just from your head, but now to compete in this market, you have to go faster and you have to use these tools.
So now you have to relearn how to do it. And that's something that we've been pushing pretty hard at designstripe to really make sure that people understand that this is the new way and you know you have to embrace it and make it part of your daily routine.