9 in 10seats in weekly use at Day 60
12firm-wide Skills built by their own team
3.0 → 7.3actions per prompt, baseline to Day 60
1,100Claude Cowork sessions in the final 30 days
A professional services firm, 60-Day Claude Rollout, 2026.
- Engagement
- The 60-Day Claude Rollout
The Problem
- Claude bought firm-wide, with usage concentrated in a handful of people and no shared structure underneath it. No agreed way to start a piece of work, no shared skills, nothing that made one person's good workflow available to the next one.
- Client work that is document-heavy by nature, handled by hand from end to end.
- Leadership needed adoption to be something they could report on rather than something they hoped was happening.
The Solution
Foundation training for all staff, non-skippable, then role tracks split by how the work differs across the firm's practice areas.
Custom builds delivered into the firm's own environment, built alongside the people who would own them, with a stakeholder in each practice iterating on their workflow with us between sessions rather than receiving it finished.
Weekly office hours across the engagement, open to anyone, which is where most of the second-order questions got answered.
A power-user build session and a library handoff, so the workflows built during the engagement stayed available to everyone afterward.
Governance and workspace structure agreed up front, then three-axis measurement across adoption, enablement and ROI, with a baseline taken before any training happened.
The Outcome
- Weekly active members as a share of assigned seats went from just under 70% at baseline to above 90% within four weeks, and closed the engagement at nine in ten seats in weekly use on the Day-60 console pull.
- Near-total coverage in the 30 days to Day 60, a literal count from the console.
- The average number of actions Claude took per prompt more than doubled across the engagement, from three to over seven. People stopped treating it as a search box and started handing it multi-step work: research, then a document, then the email.
- Twelve named Skills live and invoked at Day 60, against none on day one. We built eight of them; the firm built six, one by extending ours. At least thirteen people across the firm have now built a Skill of their own.
- Claude Cowork went from a single session per 30 days to 1,100, reaching 52% of the firm. Conversations rose nearly eightfold over the same window, from 154 to 1,200.
- On the firm's own before-and-after assessment, nobody who completed both went backward on self-rated proficiency or confidence. Most moved up. The respondents skew toward the people who engaged, so read the movement as a ceiling rather than a firm-wide average.
- The strongest signal came from the client. A team inside the firm built a Skill of their own partway through, kept iterating on it, and had it in daily use by the end. Someone else took a Skill built in a single session and extended it well past what we handed over.
“I view this pilot as a success. All the stats you just showed are not lost on me. We went from we're not doing anything to we now have a team motivated.”
Chief Executive Officer
“What we did accomplish was we got our team with licenses generally using it to make them smarter day-to-day. And that, I think, is evident in the data you presented. And I think we got a subset of that group that gelled together as a cohort that are moving things forward on their own.”
“It asked me to create it myself using your skill as reference. Just booted up the artifact and it's awesome. This would be very useful because of the pre-call prep needed for most external calls.”
Want results like these?
Book a 30-minute call. Nicole will assess your team and recommend the right engagement.