G42, the Abu Dhabi-based technology holding group chaired by Sheikh Tahnoun bin Zayed Al Nahyan, announced today that it has begun recruiting AI agents into enterprise roles across the organisation, opening an application process for agents capable of operating within approved sovereign infrastructure and delivering measurable enterprise value.
Submissions will undergo technical validation, performance testing, reliability checks, and user-experience assessment, and successful agents will enter a probationary phase before scaled deployment, with human oversight and final accountability remaining central to all decision-making. Developers whose agents make the cut will be compensated through a value-linked model tied to enterprise outcomes, something closer to a revenue-share than a traditional SaaS contract.
The framing is deliberately provocative (AI agents "applying for jobs" is, at minimum, a creative repackaging of enterprise software procurement), but G42 has been putting its money where the press release is. In February 2025, the company led a $55 million Series A in Applied AI, the Abu Dhabi-based company behind Opus, a platform that enables enterprises to design AI-native workflows for regulated industries, and which has mapped approximately 1.5 million of an estimated 2.7 million business processes globally.
Applied AI charges $1 to $1.50 per "AI man-hour" compared to $36 or more for human labour, effectively positioning the product as a labour cost rather than a software expense, and the company raised a pre-Series B from Mubadala and Arbor Ventures in January 2026, less than 12 months later, having been on track for $20 million in revenue in 2025.
Its founder Arya Bolurfrushan told the FWDstart Podcast that Opus is designed to keep humans as moral decision-makers rather than obsolete bystanders, a philosophy that rhymes neatly with G42's insistence that human oversight remains central to everything it deploys.
The emphasis on "approved sovereign infrastructure" is not throwaway language either. G42 recently unveiled its plan to protect imported US AI chips from outside interference, a prerequisite for the UAE to receive up to 500,000 Nvidia semiconductors annually through 2027, deploying a "Common Operating Picture" platform that provides continuous visibility into chip location, access, and end-use, alongside a Regulated Technology Environment with personnel screening and continuous monitoring.
US undersecretary of state Jacob Helberg called the safeguards "incredibly positive," saying they set a standard for Pax Silica, the US-led coalition securing AI supply chains, and when G42 says its AI agents must operate within approved sovereign infrastructure, it is speaking to that audience as much as to Abu Dhabi.
Globally, the pattern is already well established. In July 2025, Goldman Sachs announced that it had "hired" Devin, an autonomous AI software engineer built by Cognition Labs, to work alongside its 12,000 human developers, with CIO Marco Argenti describing Devin as a "new employee" and saying the bank planned to deploy it in the hundreds, potentially scaling into the thousands. Cognition, which is backed by Peter Thiel's Founders Fund and valued at roughly $10 billion, went so far as to publish a year-end performance review for Devin in January, complete with strengths, weaknesses, and areas for improvement.
Salesforce has perhaps gone furthest in codifying the whole concept, describing its Agentforce platform as a "digital labor" force that works alongside human employees, with a consumption-based pricing model at $0.10 per agent action and a command centre dashboard for humans to monitor and optimise their AI agents in production.
Whether G42's announcement represents a genuine shift in how the company deploys AI internally or a well-timed piece of employer branding is, for now, difficult to distinguish from the outside, and in all likelihood both things are true simultaneously.
