You have 1 article left to read this month before you need to register a free LeadDev.com account.
Estimated reading time: 6 minutes
The rise of AI-assisted coding and coding agents is turning the offshoring business on its head.
It didn’t take long: workplaces are starting to slow recruitment of offshore developers in favor of using AI to do the job.
“I think of agent shore as the new offshore. It’s happening, and it’s cheaper than traditional offshore work,” said Lexi Reese, CEO of AI observability company Lanai. By agent shore, Reese is referring to an emerging practice of companies using AI-first processes and agents to fulfill their engineering needs, instead of an offshore team. She says it has come up consistently in conversations with customers.
“The core question has shifted from where to hire to whether certain categories of work, or tasks within existing workflows, need to be staffed at all,” she added.
But this approach isn’t one-size-fits-all. Some companies are searching for talent to fulfill specific needs as they build out AI products and services.
The reevaluation of offshoring underlines an idea that’s been simmering in the tech industry: increasingly, tech leaders are referring to AI agents as “digital teammates,” “digital labor,” and the like. Software development is one area where this idea is quickly being put to the test, and offshore talent may be the first to feel the effects.
Your inbox, upgraded.
Receive weekly engineering insights to level up your leadership approach.
AI talent remains in-demand offshore
Lauren Thomas, an economist at the global employment platform Deel, tracks the intricacies of offshoring trends, predominantly in the tech space. This year, she found that the firm has seen a decrease in general offshore developer and engineer roles in the tech industry, but an increase in developer and engineer roles for jobs specifically geared toward developing AI.
“In terms of the roles where these US and UK companies are looking across the border, they’re a little less likely to be looking for software engineers than they were before, but they’re a little more likely to be looking for AI,” she told LeadDev, emphasizing that the percentage of AI-specific cross-border roles is still a very small percentage.
This is true even as the overall number of software jobs is growing. “There’s still more every year, and it’s still growing,” she said. “In the first half of 2025 there was a big increase upwards, but the share [that is being offshored] seems to be falling or remaining constant.”
She attributes the slight increase in tech companies seeking offshore workers for AI roles to the fierce talent wars gripping the industry as of late. Everybody has been trying to recruit the same engineers, and especially for the smaller startups and companies that are not traditional tech companies, offshore was a viable pool to turn to.
“I think there was a push to sort of find AI engineers where they could find them, and maybe in less competitive markets,” she said.
This is similar to how Sam Clemens, cofounder and CEO at Reprise, a platform for creating sales demos, is thinking about it. He tapped an offshore team based in Ukraine and Poland to build an AI-native infrastructure layer for his platform, thinking they were the best team for the job because they had already done this type of AI implementation for others.
“If you bring on an offshore team, in some cases, you can bring on a team that’s already done it once, twice, or three times, so they’ve already made some mistakes and they’ve learned,” he told LeadDev.
This freed up his team to work on the frontend and more customer-facing issues, and overall, allowed each segment of his team to play to its strengths. The back-end team could deliver that specific technical expertise his US-based engineers didn’t have, and since they’re not customer facing, it doesn’t matter how well they speak English was not a requirement, he said.
Clemens sees it as an advantage for startups specifically because while late-stage tech companies typically have processes baked in, startups have more freedom to be flexible in dividing up the work and finding talent wherever they can.
“For earlier stage companies,” he said. “There’s probably less of that [baked in] factor. It’s more about, where can you bring the most expertise to bear the most quickly?”
More like this
Agents > offshore
Conversely, there are circumstances where offshoring is completely losing out to the gains made possible by AI.
Patrick Husting, a longtime developer and entrepreneur, told LeadDev that he’s hired offshore developers for the last 15 years to supplement his work or complete tasks outside his main areas of expertise and interest, like back-end API work. This all changed this year with the performance increases in coding IDEs and agents.
“It completely gets rid of offshore,” he said. “If you have a person that’s good with code from end-to-end, you can orchestrate an entire team using the AI tools that are available today.”
When an offshore developer he’s worked with for ten years recently asked if he had any more projects coming down the pipeline, Husting said he had to turn him away. He’s now tackling projects completely on his own, relying on AI coding tools for work he previously would’ve needed help with or simply wouldn’t have gotten around to before.
For example, he wanted to rewrite the ride-tracking feature in his app for equestrian lovers into React Native. He used Google Antigravity IDE to do the whole thing seamlessly. He said the developer agent got everything correct almost on the first try, which should have taken him weeks. When Copilot was first announced, he wasn’t at all convinced by the demo, he said. Now, after experiencing what AI coding tools can do today, he’s all-in.
“If it can help me convert eight years of code, hundreds of screens, tens of thousands of lines of code to this new framework, that’s huge,” he said. “If not, I would still put it off and run on my old framework for years. It’s a horse app; it’s not a big money maker. It’d be really tough to pay somebody offshore to do that migration.”
Hustings believes this can be a boon for startups, allowing them to bring a minimum viable product quicker, as long as you have a talented engineering manager who “can make these tools sing.”
“For people that are like me who are just kind of software crazy,” he said. “It allows us to compete like we’ve never been able to compete.”
London • June 2 & 3, 2026
Rands, Nicole Forsgren & Matej Pfajfar confirmed
AI as labor, not tools
The term “agent shoring” is starting to bubble up in the occasional LinkedIn post, but Reese said the business leaders she speaks with use all kinds of phrases: “AI-first delivery,” “automation-led engineering,” or “reducing the junior layer.”
In working with companies on their AI strategies, she’s found that the ones finding the most success are those that are thinking of AI not as a tool or assistant, but as labor.
“Companies get out of pilot purgatory when they stop treating AI as a tool and start treating it as a workforce decision,” said Reese.
This is a rising perspective in the tech industry, with some starting to refer to AI agents as “teammates” or “coworkers,” or simply thinking about AI in terms of allocating who – or what – accomplishes what work. Companies, such as Salesforce, for example, are even positioning their AI agent solutions to customers in this way, calling them “digital teammates.”
It’s also a fine line the industry is navigating as fears around AI taking jobs from human developers swirl. When Husting had to turn away his longtime offshore developer, telling him he didn’t have any work for him, the implications were immediately palpable to him – and the developer.
“I didn’t say [to the developer] ‘you’ve been replaced with AI,’” Husting said. “But he did. He responded back to me and goes, ‘I think I’ve been replaced by AI.’”
