IBEX's $700M–$715M FY27 Guide Beat Consensus. The Sierra AI Math Is Still Unproven

IBEX closed its fiscal year with record results Thursday and then told investors the year ahead looks bigger still: it guided fiscal 2027 revenue to $700 million–$715 million, a 9%–11% gain that landed above what analysts had penciled in. The named driver is a partnership with Sierra, a conversational-AI company, that management says is scaling through client deployments. On a headline level, this reads as a momentum story with a tailwind.

The detail that should stop a careful reader sits in the quarter behind that guide. IBEX's fiscal Q4 adjusted-EBITDA margin fell to 12.3% from 13.9% a year earlier, adjusted EPS slipped to $0.85 from $0.87, and GAAP earnings dropped 11%. Management attributed the compression to training on new client wins, shifting work between delivery centers, and higher fuel costs. The guidance waved forward, in other words, while the last confirmable quarter showed profitability pausing.

A stock that grows double digits and trades at roughly 11 times trailing earnings and about 6.2 times EBITDA is cheap in absolute terms — and that cheapness is real against a comparison set, not just a low number. The cleanest peer is TaskUs, the other growth-oriented BPO, at around 5.8 times EBITDA but with slower growth. Concentrix, the biggest pure-play U.S. peer, is loss-making and trades far cheaper for reasons that make it a poor benchmark. On top of the multiple, IBEX ended the year with roughly $31 million in net cash and returns on invested capital near 30% — a company that earns its own growth and prices at a fraction of its compounding pace.

So the valuation and growth factors are strong, which is precisely what makes the next question the load-bearing one: is the AI revenue real, or is the market being asked to underwrite an ambition it cannot yet verify?

The Sierra deal, announced in early May, has no disclosed financial terms — no revenue-share percentage, no minimums that would let an investor size its contribution. What management does say is specific: AI agents are handling 20%–40% of enterprise call volume in deployments so far, resolution rates on early wins like Philippine Airlines and BJ's Wholesale are running above 20% and 40% respectively, and customer-satisfaction scores are matching or beating human agents. The company frames the AI layer as net additive rather than cannibalizing — AI captures volume faster than hiring human agents ever could and drives new wins, including case studies where Sierra introduced IBEX to the client.

That is a coherent story. It is not yet a financial fact. The confirmable evidence in fiscal Q4 cuts the other way on profit, and the forward guide is a forecast that management must now prove quarter by quarter. There is a genuine, unresolved risk that AI agents someday replace the human-seat volume that is IBEX's existing revenue base rather than simply growing around it — management says the handoff is designed the other way, but the economics of the partnership itself remain opaque.

The market's posture reflects that gap. IBEX shares had rallied roughly 30% over the four months into the print, yet the aggregate signal from AInvest still labels the stock Hold with a weak fundamental grade, a cross-check that lines up with the evidence: good growth, cheap valuation, dependable balance sheet — and a profitability trend that just inflected down.

For an investor, this reads as a show-me compounder, not a finished story. The cheapness and the growth argue for owning it in a value-growth sleeve, but the specific trigger that upgrades the case is margin re-expansion, which management has guided to resume in fiscal Q1 2027. Watch the adjusted-EBITDA margin, not the revenue topline — the next three months tell you whether the AI layer widens economics for the contact-center model or simply reallocates its cost structure. Until a margin beat lands, the honest grade is a strong business priced below its strength, with the AI math the part you cannot yet take on faith.