IBEX: The AI Re-Rating Just Met a Margin Reality Check

IBEX just posted the best year in its history, and on September 10 it told investors the growth continues into fiscal 2027. That alone would not be news for a company still trading at roughly 11 times trailing earnings and about six times EBITDA. What makes it interesting is the timing: the stock has run up about 30% over the last four months — a re-rating built almost entirely on an artificial-intelligence story — right as the numbers underneath the quarter started to bend in the wrong direction. That gap between the AI price and the operating proof is the whole question here.

A boring business with an unusually good year

IBEX runs the customer-service operations other companies outsource. When your bank, phone carrier, or retailer's call center answers, an IBEX agent is frequently on the other end — the company employs more than 36,000 people and serves more than 140 enterprise clients. It is a "business process outsourcer," a sector the market has long treated with suspicion, on the theory that AI will eventually replace the human agents it sells by the seat.

Against that backdrop, the numbers just reported are hard to argue with. Fiscal 2026 revenue rose 15.4% to $644.1 million, a record. Adjusted earnings per share climbed 28% to $3.52. Free cash flow hit a record $31.2 million, and the company ended the year with roughly $31 million in net cash and debt-light books — at ~30% returns on invested capital. The sixth straight quarter of double-digit revenue growth closed out the year.

That is a profitable, cash-generating, growing company trading at a multiple most 15%-growers would not touch. The bear case against the stock is not the balance sheet; it is the core fear that the seat-based model is structurally shrink-wrapped.

The quarter that should have supported the rally didn't

Here is the snag. The AI narrative is what pulled the stock up 30% in four months, but the quarter that delivered it went sideways on the profit line. Fourth-quarter revenue grew 12% to $164.3 million, yet adjusted EBITDA was essentially flat at $20.2 million, and the adjusted EBITDA margin slipped from 13.9% a year earlier to 12.3%. Management blamed ramp-up costs on new client wins, the cost of shifting work between delivery centers, higher fuel prices, and about $2 million in one-time charges.

Then the forward guidance decelerates. IBEX guided fiscal 2027 revenue to $700–715 million, growth of 9–11% — a clear step down from the 15% it just printed. Adjusted EBITDA is guided to $90–94 million, growth of 9–14%. Management frames this as margin expansion returning after near-term investment, and the stock held its gains on the print. But the quarter a reader was given to believe in the re-rating showed a margin going down, not up.

The AI story, and what would make it real

The re-rating rests on the AI pivot. In May, IBEX announced a partnership with Sierra, a conversational-AI platform whose other clients include companies like GAP, Rocket Mortgage, SoFi, and Wayfair. On top of that it runs its own AI push, Wave iX. Management says it has deployed AI agents with a double-digit number of clients across five verticals, and describes AI as a way to increase revenue and support human agents rather than replace them.

There is some tangible evidence already, not just a pitch. On the call the company cited a couple of AI-driven wins: BJ's Wholesale, taken on specifically through the AI solution, and Philippine Airlines, where an AI deployment automated part of the customer journey. One hallmark of the AI era that has not fully hit this business yet is showing up in paid contracts at scale; IBEX's claims about margins rest on future contribution.

The honest way to read this is that the market has already started paying for the AI upside — that is what the 30% move was — while the reported economics in the quarter still show a base business whose margin is under near-term pressure and whose growth is deliberately slowing. In the persona's terms, the cheap-multiple purchase has become a proof-of-AI purchase: the stock is no longer priced like a broken BPO, it is priced like one that is about to re-accelerate on AI revenue.

What that means for a decision now

For a beginner, the discipline is to separate the business from the stock. The business just had a genuinely strong year, and at roughly 11 times trailing earnings and six times EBITDA with a clean balance sheet, it is not priced like a company in financial distress. The AI thesis gives it a reason to deserve better than the sector's typical discount. If you had bought the cheap core before the re-rating, the case has gotten better, not worse.

But the marginal re-rating has already happened. Buying today means paying up on the assumption that the AI contribution shows up and that the decelerating 2027 guidance actually marks a profit-margin trough, not a trend. The falsifiable test sits in the next two quarters: does adjusted EBITDA growth re-accelerate against that guided range, and does AI show up as paid, growing contract value rather than as pipeline color? The margin move in the quarter just reported argues for patience — a two-quarter wait costs little for a business this size, and it buys the one thing the 30% run can no longer be argued about: proof.