Scaling immigration at Amazon scale: Inside the AI-powered 'Go to Work' playbook
Amazon operates one of the largest cross-border workforces on the planet. With more than 1.5 million employees globally and corporate teams spread across dozens of jurisdictions, the company routinely moves engineers, operations leaders and specialists between countries — sometimes for multi-year assignments, sometimes for a single week of business travel. Every one of those moves has an immigration footprint: visas, work permits, compliance obligations, country-specific rules.
For Amazon's Global Mobility and in-house legal teams, the challenge wasn't the rare, complex relocation. It was the sheer volume of day-to-day movement across hundreds of immigration regimes, each with its own forms, deadlines and risk profile. At Amazon's scale, a manual, consult-a-lawyer-for-every-trip model simply doesn't work.
Amazon's Global Mobility leadership, led by its Global Mobility Strategy head Salman Cheema, mapped where immigration friction actually landed. The answer surprised people. The most painful cases weren't the long-term international transfers — those were already handled by dedicated counsel. The real friction was in short-term visas and work permits for international business travel, where a software engineer needed to attend a three-day summit in Singapore or a product leader had to fly to Brazil for a launch review.
The company's original approach — publishing general immigration guidance and asking employees to self-serve — had produced exactly the outcome you would expect. Employees were confused, especially in the more complex jurisdictions. Legal and mobility teams were pulled into one-off questions that looked different on the surface but were answerable from the same underlying rule set. The time cost was enormous and the employee experience was poor.
Amazon partnered with Deloitte to build a tool called Go to Work — an AI-powered, employee-facing immigration assistant designed specifically for Amazon's global workforce. Instead of a static FAQ or a ticket queue, the tool gives employees personalised, jurisdiction-aware guidance on whether they can travel, what documentation they need, and how long the process will take.
Under the hood, Go to Work combines several AI capabilities:
The result is a self-service experience for routine cases and a warm hand-off to human immigration counsel when the facts cross into complexity — an AI-first, lawyer-in-the-loop model rather than a lawyer-first, AI-adjacent one.
For Amazon's employees, the change is immediate: a question that used to require a back-and-forth email thread with a mobility specialist is now answered in minutes, in the right language, with the right documents attached. Employees get clear, country-specific immigration guidance before they travel, which materially reduces the risk of arriving at a border without the right paperwork.
For Amazon's legal and mobility functions, the leverage is structural. Internal reports describe the team now supporting international movement at enterprise scale — hundreds or even thousands of workers at a time into new markets — without the linear cost growth that would be required if every case were handled manually. Immigration counsel time is concentrated on genuinely complex or novel matters rather than on answering the same short-term-travel questions repeatedly.
Go to Work sits alongside other Amazon immigration programs — including the Welcome Door program, launched in 2022 and since expanded to Germany, Poland and Australia, and the company's public commitment to hire 5,000 refugees in Europe. Together they paint the picture of a corporate legal and mobility operation treating immigration not as a cost center but as an AI-enabled capability that can scale with the business.
