Concentrix and TP are squeezed by the same forces, and on revenue TP has had the worse year. The market punished Concentrix anyway: it borrowed to buy its largest labor-heavy business just as generative AI arrived, cut guidance twice, and described its AI progress in measures that change every quarter. The gap is credibility and debt, and closing it is a board decision more than a product one.
Most coverage of the Concentrix sell-off reads it as the market pricing AI replacing contact center agents. If that were the whole story, TP would have fallen too. It faces the same offshoring, the same automation and the same price pressure, and on revenue it has had the worse year: like-for-like revenue fell 1.7 percent in the first half of 2026 and net profit fell. Its stock still rose about 11 percent while Concentrix lost about 40. This report puts the two companies side by side through their filings to find what actually separates them. The answer is debt, two guidance cuts and AI metrics that changed every quarter, set against a rival that replaced its leadership with an AI operator and whose anchor shareholder bought more stock. If you buy outsourced customer experience, hold either stock, or sit on a board deciding how to answer the AI question, the difference matters more than the headline.
Seven findings from Concentrix's fiscal 2026 filings through the quarter ended 31 August 2026 and TP's results through June, each sourced and labeled by evidence type.
Constant-currency growth went from plus 1.9 percent in the first quarter of fiscal 2026 to plus 0.6 percent in the second and minus 0.5 percent in the third. The fourth quarter is guided to a decline of 3 to 5 percent, and full-year revenue guidance has been cut twice since January.
TP's like-for-like revenue fell 1.7 percent in the first half of 2026, worse than Concentrix over the same months. Its stock rose about 11 percent this year, and about 50 percent from its March low, after it named a former McKinsey customer-care AI leader as chief executive.
Revenue rose 35 percent in fiscal 2024 on the Webhelp purchase, but pro forma constant-currency growth was 2.7 percent. Goodwill impairments of $1,523 million and $1,050 million followed within nine months, $2.6 billion in all, both attributed by the company to its own share price.
On the third-quarter call, offshoring cost about 3 percent of revenue, and two large technology clients are ending support for some customer groups faster than planned. AI shows up as clients picking cheaper automated options and as price pressure the company says it won't chase.
Within a year Concentrix described its AI business as $60 million of annualized revenue, a software line pacing toward $120 million, $1.3 billion running through its iX platform, and $3 billion of AI-influenced revenue. None of these is defined in a filing.
Groupe Bruxelles Lambert, paid in Concentrix stock for Webhelp, sold enough by April 2026 to lose its board nomination right. Saham Group, the vehicle of TP's chairman, spent about EUR 355 million in September raising its TP stake to about 12.7 percent.
Concentrix shipped AI products early. What the record shows is a debt-funded bet on labor revenue made four months after ChatGPT's launch, then forecasts the company kept missing. TP's board answered the same doubts by changing the chief executive. Concentrix's board hasn't.
At 31 August 2026 Concentrix carried net debt of about $4.13 billion against an equity value of about $1.53 billion at the 5 October price. Debt is about 73 percent of enterprise value, so a 10 percent fall in what the business is expected to be worth cuts the equity by more than a third. TP carries about 0.8 of net debt per unit of equity value, and Concentrix about 2.7. Measured against earnings the two balance sheets are close, so debt doesn't explain why the stocks diverged. It explains why Concentrix falls so much harder when expectations drop. The stock trades at roughly 2.3 times adjusted earnings guidance with a dividend yield near 5.9 percent, which says the market doesn't believe those earnings will reach shareholders before the debt is paid.
The AI threat itself runs through price before volume. HFS Research projects that per-outcome pricing shrinks the BPO market by about $61 billion against the conventional path, and Klarna's reversal in 2025 showed that what AI removed first was the outsourcer standing between a client and its own agents. By Concentrix's own numbers, offshoring and two clients' decisions are bigger drags this year than AI.
Probabilities are the author's judgment and the least reliable part of the report. The earliest visible signs are the useful part, because anyone can watch for them in public filings.
The fourth-quarter decline lands inside guidance, the offshoring drag fades as movable work runs out, and the hyperscale losses finish as planned. Revenue is close to flat by the fourth quarter of fiscal 2027 and debt falls by more than $550 million.
First sign: a narrower decline in the March 2027 release
Declines of 3 percent or more continue through fiscal 2027 as clients cut prices on renewal and move simple work to software vendors. The dividend is cut to speed up debt repayment and pressure for a leadership change or a sale rises.
First sign: fiscal 2027 guidance calling for a decline
The board starts a strategic review, sells a line such as PK's engineering services, agrees a take-private, or names a new chief executive. The stock moves on deal terms instead of operating results, and buyers face a change of owner mid-contract.
First sign: an activist 13D, a review, or an 8-K
New business outruns the losses and the iX software line reaches a size that matters. Constant-currency growth turns positive in the second half of fiscal 2027, and the call in this report is overturned.
First sign: software revenue defined in a filing
What would overturn the report's call: a fiscal 2027 quarter in which Concentrix returns to constant-currency growth without an acquisition. TP's forgiveness is fragile too. Its Future Forward plan targets 4 to 6 percent like-for-like growth a year to 2028 while it is currently shrinking, so if it misses its own 0 to 2 percent guidance for 2026, the reset fails its first test.
This isn't a vendor summary. Every sentence is labeled by what stands behind it: verified fact, vendor claim, third-party estimate, my assessment, hypothesis, or scenario. Sources are numbered and clickable. Forward-looking sections use scenarios with observable tripwires, not forecasts. It's the same method behind every market assessment I write.
Twenty-two pages, built from public sources with no client brief and no interviews. Read it in the browser or take the PDF.
Each report here answers a real question, directed and researched against public sources and evaluated against a stated assumption, then delivered as Word and PDF. If you're weighing a platform, sizing a category, or defending a number to a board, tell me the decision behind it and I'll tell you honestly whether a report is the right tool.
Commission an assessment