For years, the corporate artificial intelligence (AI) story ran on soft metrics: employees trained, chatbots launched, pilots underway. This earnings season, companies gave investors something more concrete.
monday.com told investors this week that annual recurring revenue from its AI products doubled between Q1 and Q2 2026 and now accounts for 17% of net new ARR, Co-CEO Eran Zinman said on the company’s earnings call. Zinman called it evidence that customers are “actively choosing AI features and are willing to pay more when they see value.” That is a materially different signal than employees simply using an AI tool. Customers are paying a separate, measurable price for it.
Airbnb offered the mirror image on cost. CEO Brian Chesky told investors on the company’s call that customer support costs per booking fell about 16% year over year, driven by an AI assistant resolving nearly 45% of guest issues without a human agent, PYMNTS reported. AI also cut concept-to-launch time by up to 60%, letting Airbnb ship nearly 80% more features in the first half of 2026 than a year earlier.
Boards Have Stopped Accepting Adoption Metrics as Proof
monday.com is showing AI as a revenue line customers pay into directly. Airbnb is showing it as a cost reduction inside an existing line item. Both are more specific and harder to dispute than the adoption metrics that dominated AI reporting through 2024 and 2025.
That gap between adoption and proof shows up across finance departments broadly. PYMNTS Intelligence has tracked enterprise Gen AI sentiment monthly since March 2024, drawing on more than 1,000 observations from companies with at least $1 billion in revenue. The share of executives reporting a favorable, positive result from generative AI has climbed to 96%, up significantly from a year earlier, PYMNTS reported, even as the technology remains early by most executives’ own admission.
That same confidence is showing up in how CFOs are putting generative AI to work. 87% of mid-market CFOs see generative AI as important to accelerating the close and flagging anomalies, PYMNTS reported separately, with the same share using it for debt modeling and stress tests. Friction has eased as programs mature: reported errors dropped from 80% in July 2025 to 35% by December, PYMNTS Intelligence found in a survey of 60 CFOs at $1 billion-plus companies. Integration challenges fell from 70% to 45% over the same period.
Bill for Running AI Is Becoming Its Own Line Item
None of that comes free. Worldwide spending on AI platforms and models is on pace to hit $64.25 billion in 2026, up 63.4% from $39.3 billion in 2025, and 45% of CFOs still direct AI budgets toward general productivity gains rather than the strategic outcomes boards want, according to Gartner data reported July 20. Total global AI spending, including infrastructure, is projected to reach $2.52 trillion in 2026, Gartner separately forecast. That is the bill behind every AI feature monday.com sells or Airbnb deploys.
Finance teams are building tools to track that bill. Ramp launched AI Token Spend Management on July 16, after token spend across its customers rose 20.7 times since June 2025, PYMNTS reported. CloudZero launched a similar tool from a different angle, tying AI spending to the customers and features that generated it rather than tracking token counts alone.
That mismatch between soft metrics and hard financial ones is what kept boards skeptical. Analysts at Wedbush found many enterprises ran AI pilots without any framework for measuring success, PYMNTS reported, leaving them unable to justify the spend even after significant investment. monday.com’s ARR disclosure and Airbnb’s per-booking cost figure are two of the clearest examples yet of companies meeting that challenge with a number instead of an anecdote.
For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.
