How Deel Combines AI and Human Insight for Better Work, with Abhijit Mehta, Senior Director of Product

Deel’s mission has always been to simplify the most complex aspects of global employment. Today, AI plays a central role in delivering on that mission. From automating document processing to cutting visa timelines from 30 days to 5, we’re building AI solutions that deliver measurable results: thousands of hours saved each month, faster compliance across 150+ countries, and employees empowered to focus on meaningful, high-value work.

At the forefront of this work is Abhijit Mehta, Senior Director of Product for Deel’s AI and Knowledge department. Abhijit’s team focuses on three pillars: building strong AI platforms, delivering customer-facing products, and enabling every Deel team to use and build with AI.

We sat down with Abhijit to get his insights on what drives successful adoption, why a solid knowledge base is fundamental, and the key AI principles every leader should follow.

Meet the expert: Abhijit Mehta

This might sound surprising, coming from someone who builds AI products, but I don’t believe that AI should be everywhere. It’s a means to an end, and doesn’t need to be implemented where it’s not needed. That’s why I focus on building products that deliver real outcomes and add real business value. With this focus, we’ve saved thousands of hours each month, and increased AI adoption within our teams by making processes that are useful and accessible.

My department, AI and Knowledge, has three focus areas: building strong AI platforms so we can scale new products quickly, creating customer-facing AI-powered experiences, and enabling every employee at Deel to build with AI without having to understand the technical complexities. These three pillars allow us to move fast and deliver on two of our most important goals: improving customer experience and driving operational efficiency.

Cutting processing times from months to days

One of the biggest time drains in global HR is document processing. Every stage of the worker lifecycle—hiring, payroll, offboarding—comes with a flood of tax forms, IDs, licenses, and contracts. For years, our teams had to manually open each document, pull out the relevant information, and put it into our systems. It was repetitive, error-prone, and costly.

We built our document AI to change that. Today, the system processes documents automatically, saving us more than 1,000 hours every month. Our people no longer have to do that manual, repetitive work.

The impact is even bigger in visa processing. A certain type of U.S. high-skilled visa used to take 30 days with heavy paralegal involvement. With AI-powered petition drafting, it now takes just 5 days. But let me be clear: we don’t leave compliance to AI alone. Human experts remain in the loop to review and validate every case. AI does the repetitive drafting work so specialists can focus on work that requires the human touch.

The business benefits are obvious, but the human impact is even greater—workers face less stress and uncertainty at one of the most critical moments in their careers.

Achieving speed and compliance with knowledge systems

When you operate across 150+ countries, the complexity is staggering. Payroll, compliance, and HR processes look different everywhere, and you can’t solve that just by hiring more people or giving teams more time. What you need is a system that makes the right knowledge accessible to everyone, instantly.

That’s why one of our first big investments was building a knowledge base. Knowing how to do things in a compliant way across so many geographies is an extremely difficult problem. But by putting all that information into a single, living system, we made it easy for teams to find what they need. This knowledge base became the foundation for our AI.

I’ve always believed that for great AI products, you need even better knowledge systems. The quality of AI outputs depends on the quality of the information they draw from. By investing heavily here, we’ve enabled compliant decision-making at scale, while making information accessible to everyone across Deel.

Empowering people and building trust

Introducing AI into the workplace can make people anxious. But at Deel, what we saw was excitement. Teams that had been stuck doing repetitive tasks like document processing welcomed the change, because nobody enjoyed that work in the first place.

The feedback has been consistent: AI helps people achieve outcomes faster, and that makes them look better in their roles. It gives them a productivity boost that not only benefits Deel and our customers but also helps individuals in their careers by turning them into top performers. That’s been both very gratifying for the teams building AI at Deel, and also a significant driver for adoption.

Of course, trust is critical. We’ve made conscious decisions to put privacy and compliance first, even when it made our work harder. For example, when we use OpenAI models, we enforce a zero data retention policy so no personal information is ever stored. All of our AI actions run on Deel’s certified platform and permissions framework—if someone doesn’t have the right to access certain data, neither does AI. We even built our own auditing and traceability systems to ensure accountability. It hasn’t always been easy, but it’s the right thing to do.

Scaling to every vertical with AI agents

We’ve now reached a point where AI can do more than just automate single tasks. With Deel AI Workforce, our customers will be able to add AI agents to their teams just like they would a human colleague. These agents handle repetitive, structured work—almost like hiring an intern who never gets tired.

This is still early, but the potential is enormous. Industry leaders like IBM have saved 12,000 hours in a single department in one quarter by automating HR processes, and the UK government has shown that generative AI can save civil servants 26 minutes per day, nearly two weeks per year per person. Early adopters of AI agents will see similar results.

For Deel, this is just the beginning. We’re continuing to give our AI agents more skills, and soon every vertical we offer—payroll, employee engagement, compliance—will be powered by AI.

Lessons for leaders

From our own journey, we’ve learned two principles that every leader should remember when building an AI strategy. First, invest in knowledge systems early. AI is only as good as the information it can access, and without a solid foundation, outputs will falter. Second, move fast. With AI, real learning doesn’t happen in the lab. It happens in the field, with live data and real users. Waiting too long means wasting time you could spend learning and improving.

At Deel, our approach has always been about balance: using AI to free people from repetitive work, while keeping human oversight central to compliance and decision-making. The result is more than just operational efficiency. It also means empowerment for our teams, our customers, and the workers they support.

Reimagine what your people can achieve when AI takes care of the rest, and get more information on Deel AI Workforce.