From pilot to daily habit: how enterprise AI adoption is actually scaling in 2026

Two years ago, 32 people at law firm Ropes & Gray were sending a few hundred prompts a month to an AI platform called Harvey. Today, nearly 2,200 employees at the same firm generate more than 282,000 prompts a month, and each person is using it roughly three times more than they were a year prior. That figure, reported by the Wall Street Journal, is not an outlier. It is increasingly the benchmark that enterprise operations leaders are being asked to match.

The mechanism behind that growth is less about technology procurement and more about organizational design. The Wall Street Journal profiled how Ropes & Gray, Citigroup, and Mars have all engineered similar jumps through internal 'AI champion' networks: regular employees designated as peer advisors who help colleagues connect AI tools to actual recurring work. At Ropes & Gray, Howard Glazer, who co-leads the firm's private equity transactions practice and now also serves as Head of Practice AI, oversees a network of more than 60 such champions across the firm.

Where the real numbers are

Citigroup's figures are equally striking. The bank has crossed 80% employee AI tool adoption and logged 42 million AI interactions since its platform launched, according to the Wall Street Journal. More than 10,000 of Citigroup's engineers now use AI tools daily. Those are operational metrics, not pilot KPIs, and they signal that the bank has moved well past the evaluation phase.

The broader market context reinforces how fast this window is moving. Forbes Advisor's compiled AI statistics, verified through July 2026, reflect an enterprise environment where AI tool deployment has accelerated sharply across sectors, with financial services and professional services leading adoption curves. The data suggests the gap between early-scale organizations and those still in experimentation mode is not closing on its own.

According to workplace analytics research cited by ai365.blog, peer-supported adoption achieves roughly double the sustained usage of top-down mandates. That finding matters operationally because it reframes the rollout question: the bottleneck is rarely the platform, it is the pathway from license to habit.

Workflow prompts, not curiosity prompts

The distinction that separates high-volume adopters from stalled ones comes down to prompt type. As ai365.blog noted in its analysis of the Wall Street Journal's reporting, the prompts that drive measurable productivity gains are not one-off exploratory questions. They are repeatable 'workflow prompts' built around a specific recurring task, a contract review checklist, a client memo template, a daily risk summary, run the same way on a predictable schedule.

This reframes how IT and operations leaders should think about AI tool ROI. License utilization rates and seat counts are lagging indicators. The leading indicator is whether employees have identified and codified two or three workflow prompts they run every week. Organizations that get to that point at scale are the ones generating six-figure monthly interaction volumes.

Mars, Inc. is among the non-financial-sector examples the Wall Street Journal identified as deliberately building toward this. Consumer goods organizations face a different set of recurring tasks than law firms or banks, but the structural approach, peer champions, shared prompt libraries, task-specific rather than general-purpose guidance, transfers across industries.

What this means for your team

• Audit current usage before expanding seats: if existing licensed users are still sending one-off queries rather than repeatable workflow prompts, adding licenses will not move the needle. Identify who is already running structured, recurring prompts and build outward from them.
• Designate champions by function, not by seniority or IT proximity. The Wall Street Journal's reporting across Ropes & Gray, Citigroup, and Mars consistently points to peer credibility, a colleague who does the same work, as the trust driver that converts skeptics.
• Build and circulate prompt templates for your five highest-frequency team tasks. The Ropes & Gray trajectory from hundreds to 282,000 monthly prompts was enabled by making successful habits copyable, not by requiring each employee to experiment independently.
• Set interaction volume and per-user frequency as dashboard metrics alongside seat utilization. Organizations measuring only adoption rates (who has access) miss the operational signal of engagement rates (who is using it daily for real work).