Should the AI revolution pause your offshoring, insourcing, or Global Capability Center plans? Almost certainly not — and yet a growing number of executives are doing exactly that. Across industries, organizations are shelving resource transformation programs on the assumption that AI will eliminate these positions anyway. The rationale sounds plausible. The economics say otherwise.

Up until 18 months ago I was an AI skeptic. Not that I saw any issue with the technology — I've just been through far too many technology hype cycles, including previous AI hype cycles, that left me cynical. After examining the technology, the tools, and their implications, I'm now firmly in the camp that AI will be highly transformative and will have a major impact on the technology job market. I just don't think it will happen nearly as fast as many are predicting.

The reason is Institutional Inertia — the inability of most companies to get out of their own way long enough to take advantage of a new technology. Middle managers are incentivized to do the same thing year in and year out. If a manager runs a $5M function and finds a way to do it for $3M, they get the same pat on the back as before — but with all the downside risk of having changed something. That dynamic doesn't disappear because AI exists. RPA automation has been available for over 20 years, and large companies still have dozens of automation opportunities sitting untouched. Institutional Inertia is why.

The Breakeven Math Still Works

The breakeven point for offshoring, insourcing contractors into a GCC, or creating a new center is typically 18 months or less. Consider the numbers at just 300 FTEs:

  • Offshoring 300 US resources to India: $24M–$45M in gross annual savings ($80K–$150K per FTE)
  • Insourcing 300 India-based vendor contractors: minimum $6M/year, averaging $9M or more
  • Building a new 300-person GCC from scratch: capital investment of approximately $2M (office buildout at $45–55/sq ft plus legal and entity costs)

Even in a worst-case scenario — where you spend $4M in setup and knowledge transfer costs — you're saving a minimum of $9M in the first 18 months of operation. That's a positive return even under the most pessimistic assumptions.

What Are We Actually Hedging Against?

AI's impact on the job market will come in two waves. The first is automation — think RPA on steroids. Like traditional RPA, you may free up capacity equivalent to 150 FTEs, but in practice you've freed 1/3 of each of 450 people. You still need the other two-thirds. With effort, you can consolidate and eliminate 25–40% of the affected positions — but that's a multi-year process.

The second wave is the elimination of specific IT roles throughout the software development lifecycle — junior developers, business analysts, and QA personnel in particular. I estimate 60–80% of those roles will become unnecessary, but on a 3–5 year horizon, not 18 months.

And there's a wild card the pessimists consistently underestimate: organizations have an insatiable appetite for capacity. More often than not, increased productivity from AI gets absorbed by new work program rather than headcount reduction. Projects that couldn't justify ROI under the current operating model suddenly become easy approvals when AI improves productivity. The backlog of that kind of work — particularly in legacy modernization — is enormous.

The Lease Hedge

For organizations renewing or signing GCC leases (typically 5 years with two optional 5-year renewals), I do recommend negotiating an option to return space at renewal or mid-term as a hedge against AI accelerating faster than expected. That said, office rent in India runs just $1,200–$2,000 per FTE per year — it is not a meaningful factor in the overall savings equation.

The Bottom Line

Labor is the largest component of running an IT or BPO department. Offshoring, insourcing, and building or expanding a Global Capability Center are among the most powerful levers available for optimizing it. The economics grow more compelling as headcount scales — I used 300 FTEs deliberately to show how substantial the returns are even at a modest scale.

If your organization is stalling these programs due to AI speculation, you're leaving $9M to $45M per year per 300 FTEs on the table while you wait for a disruption that — by most honest assessments — is still 3–5 years from materializing at scale.

The question isn't whether AI will eventually reshape the workforce. It will. The question is whether you can afford to wait.