Return on AI: Weekly Digest — July 13, 2026
This week splits across three areas worth your attention. The strongest financial evidence runs through government and enterprise deployments: HMRC's tax recovery tools, McKinsey's EBITDA data on firms that moved past pilots, Albertsons' back-end results, and AT&T's cost reduction through model routing all point to where operational AI is producing measurable returns. The legal sector gets sustained coverage, with document review timelines collapsing, headcount holding flat against rising workloads, and new pressure on Big Law's pricing model. Separately, the infrastructure story has its own constraints: transformer shortages are extending data center build timelines by years, which has direct implications for any capital plan built around near-term AI capacity.
Key Takeaway of the Week
McKinsey's data shows AI leaders have already grown EBITDA by 20% while most organizations remain stuck in pilot mode, and the operational evidence supports why: document review cycles that took two weeks now close in 48 hours, legal teams are absorbing greater workloads without adding headcount, and AT&T reports model routing cuts enterprise AI costs by 90%. The consistent pattern across industries is that firms treating AI as a production system rather than an experiment are compounding real margin gains while their competitors run tests. The ROI gap between leaders and laggards is no longer theoretical, and the longer pilots run without deployment decisions, the wider that gap becomes.
Sponsored by Jetson Collective