The Attribution Dam Has Broken: AI Displacement Is No Longer Subtext
TexTak's [white-collar-displacement] forecast sits at 70% — but that number requires a confession before an argument. We've been tracking a forecast about public attribution behavior, not about displacement itself, and today's news forces us to sharpen that distinction considerably. Coinbase, Freshworks, and Cognizant have now done something that, until recently, companies were carefully avoiding: they've named AI explicitly as the cause of workforce reduction, not just a background condition. The question is whether what we're seeing constitutes resolution of our forecast or merely its precondition.
Let's start with the definitional problem we've been sitting on too long. Our [white-collar-displacement] forecast has read as 'First major layoff wave explicitly attributed to AI automation' — and a knowledgeable reader could fairly argue that IBM's 2023 announcement about pausing hiring for AI-replaceable roles, or BT's workforce reduction statements, already met a loose version of that threshold. We didn't retire the forecast then, and we should explain why, because the distinction matters.
Our working definition requires something more specific than an executive mentioning AI in the vicinity of a headcount announcement. We're watching for a company to state, on the record, that a defined set of roles are being eliminated because AI systems have demonstrably assumed those functions — with the attribution clear enough that it could appear in a regulatory filing or earnings disclosure. IBM in 2023 signaled intent to pause hiring. What we're seeing in May 2026 is different in character: Freshworks CEO publicly acknowledged over 50% of new code is now written by AI and simultaneously announced an 11% workforce reduction in a quarter when revenue grew 16%. That's not restructuring dressed up with AI language. That's an explicit productivity-displacement statement. Coinbase's Brian Armstrong wrote to employees that 'this is a new way of working, and we need to leverage AI across every facet of our jobs' while cutting 14% of staff. The causal chain is being narrated, not implied.
So why are we holding at 70% rather than calling this resolved? Two reasons, one methodological and one substantive. Methodologically: our resolution criteria still require us to define minimum affected headcount for a single announced reduction explicitly attributed to AI. The Cognizant 'Project Leap' figure of 12,000-15,000 is the closest candidate — but it remains unconfirmed as of this writing, and the attribution language isn't yet public. If Cognizant confirms both the headcount and the AI-attribution framing, that's plausibly our resolution event at scale. Substantively: the Gartner finding that 80% of enterprises cut staff after AI deployment but 'were just as likely to see negative outcomes or marginal gains as meaningful ROI' is the sharpest evidence that what's happening isn't a clean productivity story. Companies may be cutting staff in anticipation of AI productivity gains that haven't materialized yet — which means the attribution is real but the underlying thesis (AI is actually replacing human work at scale) is murkier than the announcements suggest.
Here's what would move us: Cognizant publicly confirming Project Leap with explicit AI-attribution language covering 10,000+ roles would push us above 80% and likely trigger a resolution call. Conversely, if Q2 earnings across this cohort show AI cost write-offs or deployment reversals — consistent with the Gartner ROI distribution — we'd reframe the forecast as 'companies attributing displacement to AI before AI has actually delivered the productivity' and reassess what we're actually measuring. We're watching Cognizant's next earnings call and any SEC filings that characterize the restructuring rationale. Those are the documents that matter.