AI Meet P&L

The AI, Applied Wave 3 Benchmarks suggest that AI spending is accelerating faster than most finance teams can model it. A myriad of data points are combining to paint a challenging picture for the B2B CFO. This chapter explores the benchmark data on AI costs and ROI.

It is a challenge to describe this cost increase prediction as a forecast — it is more of a budget event already in motion. In our experience, engineering budgets were not built for usage-based costs that scale at this speed.

B2B technical decision makers report AI initiatives at a median 15% of total IT spend in Wave 3 — and the Runner tier is 2× that, at 30%. For finance leaders modeling AI budgets, spend intensity is high and rising, and it is the most AI-mature organizations pulling the average up fastest. How finance and R&D teams navigate this P&L uncertainty will remain a focus of this series.

Across 18 months of accelerating AI adoption and use across both R&D and GTM teams, the share of surveyed companies that can name a revenue or cost-savings P&L line tied to their AI program has not moved — Wave 1: 45%, Wave 2: 46%, Wave 3: 45% — and remains a minority today. While attribution is genuinely hard when AI is embedded across functions, this hard ROI measurement gap is the one CFOs will likely close in on as AI spend becomes a material line item in the P&L.

AI has a P&L story that CFOs are likely to understand: reduced headcount costs. The 29% of technical decision makers using AI to justify entry-level technical role reductions is the only clean P&L story some finance teams are hearing — a story heard from only 16% of GTM decision makers. However, headcount reduction as the primary AI ROI narrative is an incomplete business case, as cost-increase approaches like upskilling and hiring new AI-specific talent are cited more frequently as talent responses to AI by both Tech and GTM decision makers.

AI leading to headcount reduction appears frequently in the media, but it’s not showing up yet in our benchmarks data. Upskilling, AI-skilled hires, AI data costs, and the gap between cost increases and commercial ROI gains is the dominant data set so far in this benchmark series.