Whether you’re an Android or iOS user, most people have heard of PicsArt. The platform launched over a decade ago and has become one of the go-to services for all things image and video editing, with over 150 million monthly active users.

However, the journey has not been easy for the company. Despite being an early player in the smartphone publishing space, the company has faced significant competition from players like Canva and Adobe, who play a game of cat and mouse for some time now, creating their own similar products. When I spoke with Artavazd Mehrabyan, the company’s CTO, at the recent WCIT conference in Armenia, he spoke quite clearly about the challenges, saying that it is difficult to be or at least stay different for long. in this market.

“A lot of things that PicsArt had before have been copied into competitors. PicsArt was the first all-in-one editing service on mobile. There were no other players before 2011. We started with this approach and it was copied, among other things,” Mehrabyan said. He pointed out that the same thing is happening with AI, where competitors, including consumer photo services, offer very similar capabilities.

For example, PicsArt offers object generation, allowing users to use advanced AI to create the required photo assets. The same capability has also been integrated into other products in the category, creating a sort of overlap.

However, instead of looking to stand out by adding more tools to its existing set of more than two dozen AI capabilities, the company is looking to leave its mark on users by improving the quality of what it offers . Specifically, Mehrabyan said, the focus is on how they produce and adapt features to help clients achieve their goal — whether they want to remove a specific object from a vacation image or generate ads visually attractive, accompanied by images and copies.

Train high-quality creative AI

In the early days, when AI didn’t exist, Mehrabyan said most of PicsArt’s research and technology efforts were aimed at making mobile editing seamless.

“It was very difficult to get all these offline editing features working on the device. Then the next challenge was scaling our ecosystem and infrastructure to support a growing user base. This led us to a hybrid infrastructure. We started with multi-cloud and one data center, which, until now, remains the best solution because it is more cost-effective, high-performance and very flexible,” explained Mehrabyan.

With this technology stack in place, the company launched its first AI feature in 2016, running a set of small models offline on users’ devices. This gradually evolved into a large-scale AI effort, with the company transforming into an AI-focused organization and leveraging its infrastructure and back-end services to serve models and APIs larger in order to benefit from more enhanced features such as background removal/replacement. More recently, with the wave of generative AI taking shape, PicsArt has begun training its own creative AI models from scratch.

In the creative field, it is very easy to lose a user. A small mistake here or there (resulting in poor quality results) and there’s a good chance the person won’t come back again. To avoid this, PicsArt focuses extremely on the data side. It selectively uses data from its own network – marked by users as public and free to modify – to train AI models.

“We have a special “free edition” license. If you post publicly and mark your image – from a stock photo in any category to a sticker or background – as free to modify, this allows another user of the Service to reuse it or to work on it. So, in essence, the user contributes this image to the community and to PicsArt itself,” Mehrabyan said.

The license has been in place since the service’s debut and has given PicsArt a massive stockpile of user-generated content for AI training. However, as the CTO pointed out, not all of this is high quality and ready for immediate use. Data must go through multiple layers of cleaning and processing, either manual or AI-driven, to be transformed into a secure, trainable dataset.

“Ultimately, we have a fairly large dataset that is proprietary to PicsArt. We don’t need additional data,” he said.

However, having a large volume of high-quality data was only part of the puzzle.

The real challenge for PicsArt, as Mehrabyan described it, was creating the “data flywheel.” A self-reinforcing cycle covering not only data accessibility, but also aspects such as how to annotate the data, how to use it and eventually how to leverage it as part of a continuous learning process to self-reinforce improve over time.

Establishing a feedback loop to achieve this was a long and complex process, he said.

“We built our own annotation technology. We have developed all associated infrastructure and ecosystem technologies in-house, including those for identifying and classifying images, labeling them and adding different types of labels to them,” Mehrabyan said. “Then we created a team to help refine the pipeline and provide feedback over time. This has mostly been very automatic, AI-driven, with human feedback in between, so we can continually improve.

Feedback Loop Leads to Contextual Intelligence

While the human feedback loop has played a vital role in improving PicsArt products – improving the quality of the results they generate – it is also moving the company towards what Mehrabyan calls “contextual intelligence” or the ability of the platform to understand user needs. and deliver exactly what they want.

This feature is particularly important to the platform’s growing base of professional users who are looking to work directly from their smartphone. Whether it’s generating graphics or a full-fledged ad for a social media campaign. The platform is still primarily used by people looking to edit personal content, but the company says its research shows many want to use it, particularly for marketing use cases.

“Contextual intelligence not only tracks your history or what you were doing to help you be more productive in your journey, but also predicts your next intention. It’s both reactive and proactive,” he explained.

This way, whenever an individual uses the platform to create something for their work, they won’t have to define the language and tone of the brand. The product would already have context in place and use it to generate the required content. Mehrabyan said the company also plans to launch a brand kit feature that would allow users to tailor this context to their needs and further improve the quality of builds.

Creative AI agents on the way

Ultimately, Mehrabyan says contextual intelligence will lead PicsArt toward an agent-based ecosystem. This is where users will have a sort of co-pilot – with all the relevant knowledge about their work and design preferences – to help them with their tasks.

“This co-pilot would understand your intent and historical context to provide you with interactive support and guide you to be even more productive. We see this use case as integrated into the entire PicsArt ecosystem, from a user perspective,” he said.

Beyond that, he also expects AI agents to help PicsArt users perform certain tasks in bulk. For example, if a user needs to apply the same design or design logic to multiple resources, they can use an agent to automate the workflow on their behalf.

In this way, the company hopes to be a key driver in the creative industry, leapfrogging its competitors and enabling users to grow their creativity, and eventually their businesses, without much effort.

Mehrabyan noted that AI will bring major change, but that users – from businesses to designers and marketers – must try to understand how it affects them and take advantage of the changes to do more than is currently possible.

“From the current point of view, this will have a negative impact. But if you take a different perspective, such as the future, you will see that these people will leverage AI to learn much more. They will no longer be restricted specialists. They will cover larger areas, deeper and faster with AI,” he noted.

According to Future Markets Insights, the global AI image editor market is expected to grow from $80.3 million in 2024 to $217.9 million by 2034, growing at a CAGR of 10.5%. Meanwhile, AI-based generation, which has become an essential part of most image editing tools/services, including PicsArt, is expected to grow 38% from $8.7 billion in 2024 to $60.8 billion in 2030.