This Startup Founder Wants to Democratize Access To AI

This article first appeared in the March/April issue of Inc. Arabia's digital magazine. To read the full magazine, click here. 

Born and raised in Germany, Hassan Sawaf is a Syrian-German engineer and entrepreneur with 31 years of experience in AI. “In essence, my journey is about the democratization of AI,” he tells us. For him, education and access to technology and data are key to enabling equitable access to AI for everyone.  

A former lecturer and senior researcher, Sawaf sold his machine translation startup AIXPLAIN AG to US-based AppTek in 2004 and relocated to the US, where he headed up AI divisions at eBay, Amazon Web Services, and Facebook. “During my tenure at these large corporations, I saw that many of the people who needed AI spoke one language and the people who were building AI spoke a different language. That made me think: Why don’t we consider this a translation problem and build an AI agent that translates business problems into AI solutions? And that’s what drove me to start aiXplain,” he says.  

In the mid-2020s, Sawaf founded aiXplain to create a marketplace for equitable access to AI. Through aiXplain, users can build their own solutions for business problems, benchmark them, and deploy them, paying as they go. “Our vision is to build a go-to AI platform that transforms ideas into real-world impacting solutions. We want to lead with simplicity and continue with affordability,” he tells us.  

aiXplain works like a marketplace for AI business solutions, with ready-made solutions available for immediate use, and custom-made solutions built upon request. aiXplain’s agent, Bel Esprit, French for “Beautiful Soul,” guides the team behind aiXplain as well as its customers in building solutions. Today, the platform has 43,000 models and employs 1,400 scientists and engineers across 5 continents.  

For Sawaf, engaging engineers from around the world is a key component of democratization, allowing scientists and engineers who are not employed by Big Tech to benefit from AI. At Bel Esprit’s prompting, these scientists get involved to help resolve challenges that the agent cannot. Unlike other AI-powered solutions which offer a one-size-fits-all solution to users, aiXplain allows users to build their own solutions.  

One of its strongest features is its ability to deal with different languages and dialects, says Sawaf, who has seen a significant uptake of his startup in the Middle East. Almost half of aiXplain’s customers are based in the Middle East, he tells us, with the rest of its customers coming from the US and Europe. 

“The rate of adoption I’m seeing in the Middle East is actually very refreshing. People in the Middle East connect to systems that can deal with English, Arabic, and French, instead of being only English-centric, like many of the solutions that are coming out of the US are,” he says. He stresses the need to have AI that can handle Arabic and its multiple dialects.  

One use case, for example, could be as simple as recording, transcribing, summarizing, and assigning action items to a conversation. This service, offered by multiple platforms, may seem redundant. aiXplain’s advantage is being able to handle languages and dialects of common languages, which many AI transcription tools cannot.  

Because AI can understand, process, compare, and automate requests, it becomes an important resource when addressing exceptional situations, like natural disasters, when companies cannot deploy sufficient resources to address customer needs. “Sometimes, the only way to solve problems is through automation, because it’s difficult to find trained resources to address challenges,” Sawaf tells us.  

Sawaf is no stranger to the anxiety that comes with technology replacing humans. He explains that, while his machine translation startup did displace translators in the beginning, it also created more use cases where human translators were needed. “When I started at eBay, my first project was to expand it into new countries. This required making listings and queries available in the same languages within 200 milliseconds, and there was no way that a human could do that.”  

This new use case, he tells us, created 15,000 jobs for human translators, which were needed to improve indexing and readability.  

Similarly, he describes AI as a sidekick that will allow writers, doctors, artists, economists, and other professionals to do their jobs more efficiently by tapping into a wealth of tools and knowledge that they would not have otherwise been able to access.   

“We should be excited about AI,” he says. “There is still a lot to be done. Right now, we are just scratching the surface.”