Revolutionizing Photo Editing with AI: An Interview with Tatevik Maytesyan

In this exclusive interview, we sit down with a seasoned member of Picsart, a leading photo editing platform, to delve into their role within the company and the transformative power of AI in the field of visual content.

With Tatevik Maytesyan, we discuss the technologies driving Picsart's success, the impact of AI on the photo and video production industry, their competitive edge, and their vision for the future of the industry. Join us as we explore the fascinating world of AI-driven photo editing with a Picsart expert.

— What do you do as a member of Picsart, which projects do you lead?

— June will mark my 4th anniversary at Picsart, which is really exciting. Over this time, my role has evolved from guiding international strategy and operations across global markets to leading product strategy and operations for content. More recently, I’ve led projects at the intersection of generative AI and content for our B2C and B2B products. For example, last year I worked on the integration of generative AI tech into our internal content design process, which led to a seven times increase in our monthly productivity.

— Picsart applies AI technologies widely. Could you tell us about the precise technologies you use in the app and how do they improve the user experience?

— Our computer vision and machine learning technologies are a growing list, but key applications include background and object removal, upscaling, enhancement, filters, effects and font generation. We’re also using generative AI for image, text, background and object generation, avatars and more.

Recently, we have launched a B2B SaaS product Adpilot, which uses a combination of the tech mentioned above. It’s designed for marketers and small business owners, allowing them to create a variety of ads for different channels and testing in seconds.

— AI has become viral lately. How does Picsart treat this buzz and do you feel like AI has influenced or will influence the future of photo and video production?

— Unlike other platforms, we’ve been using AI for years, well before it was buzzy. Picsart was founded by an AI professor and his students, so we’ve always believed in the transformative power of the technology. To date, we’ve already seen the value of stock photography dramatically change, the speed of execution and productivity for design accelerate, and creative roles and expectations evolve.

Now, what I foresee coming is a major change in the advertising industry. While there’s been significant changes in communication channels with TV and internet, AI is changing how the processes work. GPT models are challenging the brainstorming and conceptualization process, while photo and video generative AI are setting new benchmarks for the quality and uniqueness of the design, and the speed to market. The impact of machine learning on visual content is not just for the future, it is happening now.

— Is Picsart planning to compete with heavier AI models, such as Midjourney, and what is your forecast for the future of the photo processing industry?

— Yes. Looking ahead, text-to-video will be the next big innovation for the industry. We recently became one of the first to release open-source code for this.

— What do you think are your biggest advantages if compared with your competitors?

— Number one is the engagement we see among our community – it is a major differentiator. Second is that, unlike many single feature tools, Picsart is an all-in-one editor that offers deep tech and high quality results at a very low cost. Third, we’re operating at an enterprise scale, allowing any business or developer—be that in printing, marketing, real estate, e-commerce—access to our leading tech. This is widening the reach of our creative software.

— What are the main challenges you are facing while implementing AI in Picsart and how do you tackle them?

— One of the biggest challenges is the balance between speed to market and quality. On one side, we still have the startup mentality and rush to innovate as fast as we can. But, as we’ve grown into a platform with 150 million monthly active users, we need to be mindful of the quality of our features and how they perform at scale.

— Which technologies and vectors of the later Picsart development do you believe are the most prospective for the future of the industry?

— In general, generative AI has been great for image and text generation. But, when it comes to video, there is still a lot of exciting change to come. I believe that Picsart’s text-to-video model is igniting this, with our AI Research team (PAIR) recently releasing their work publicly - you can check it out here.

— How does Picsart work with users’ data and how do you guarantee its safety and confidentiality?

— We take user security very seriously and have strict protocols in place to protect data. These include secure coding practices, regular penetration and stress testing, and more.

— What were your most impressive projects at Picsart and how do you feel they have changed the industry of photography processing?

— I would have to say the international expansion of Picsart. While many tools remain either US- or China-focused, my team and I worked hard to ensure not only language localization, but the localization of the whole user experience – from the login pages to pricing, local content and collaborations. Making products like Picsart accessible and relevant for hundreds of millions of users across the world has been one of the proudest moments of my career.

— How does the process of development of new functions look like in Picsart, and how do you prioritize the features to be implemented in the app?

— We follow a product development process that’s been created by our Product Operations team, and continue to experiment with a cadence between 3 or 6 months. This briefly looks like a list of high level goals cascading down to our Business Unit owners to decide who will take ownership of what, followed by their teams meeting together to create a list of initiatives through brainstorming and the review of current data and user insights. After a voting process, we align across teams, get sign off from senior management, and begin developing.

— Which tendencies do you forecast to appear in the photo and video industry in the not so distant future, and how does Picsart plan to respond to them in its evolution?

— The first one is AI driven and guided (assisted) editing. Up until recently, a lot of the content has been taken by cameras and phones, then manually edited in various editors. Now, people can create any photo they like with a prompt and edit it in only a few taps. And in the near future, I believe automated editing tools will become even more sophisticated, offering users quick solutions and recommendations – right after the photo and video was taken.

Another big change to come is video editing. With new text-to-video tools, and the increasing availability of powerful hardware and advanced algorithms, real-time video processing capabilities will become more accessible. This could lead to features like live video editing, applying visual effects on the fly and streaming high-quality content with minimal latency.

— How does Picsart work with other companies and partners in order to create innovative products and to improve user experience?

— We’re helping companies fulfill the growing demand for powerful visual products and experiences. With our API and SDK offerings, we’ve helped partners like PopSockets, eProductivity Software (ePS), Mediaclip and more with product customization on their platforms. We’ve also offered some incredible integrations over the years with platforms like Shopify, Zapier, Windows and Google Drive.

— What piece of advice would you give to young specialists who aspire to work in the area of technologies and AI?

I believe that curiosity is the fuel for creativity. Therefore, one piece of advice I’d offer is: continue to learn and experiment every day, and share and collaborate with peers and partners. This is how great products are being built.