CustomersHarrow Lane LLP
Seven in ten asks, answered from the firm’s own deals.
Harrow Lane’s commercial group had a good playbook and a better memory, held mostly by four senior associates. In eight weeks they made that memory searchable, and cut a markup turn from two days to one morning.
Firm
Harrow Lane LLP, London
Size
240 lawyers, 31 in the pilot group
Practice
Commercial and technology
Archive
4,100 agreements, three years
System
iManage Work, UK residency
Pilot
Eight weeks, March to May 2026
The problem
The answer was always in a deal someone else had done.
Every incoming markup began with the same question in the team channel: has anyone given a 2× cap to a logistics client? Usually someone had. Finding the deal, the turn and the reason took an associate most of a morning, and the reasons rarely made it into the playbook.
When two of those associates moved to the finance group in 2025, the team felt it within a month.
Measured on 14 live turns, weeks 3 to 8. Time from receipt to partner sign-off.
Eight weeks
No change to how the group worked: same Word, same email, same sign-off. Vellcite sat in the middle.
Week 1
Connect
Three years of the group’s matters indexed inside Harrow Lane’s own tenancy. Ethical walls checked against 40 restricted matters before anyone signed in.
Week 2
First live markup
A 58-page supply agreement from a logistics customer. 37 changes; Vellcite answered 22 from precedent. Two associates checked every citation by hand.
Week 4
Trust, measured
71% of substantive asks answered from the firm’s own deals. Partners overrode 9% of suggestions, mostly where a client relationship had changed.
Week 8
Decision
The group moved to the Practice plan, and the firm’s finance group asked to be next.
71%
of substantive asks answered from precedent by week four
5h
median time to a signed-off reply, down from two days
9%
of suggestions overruled by a partner, each with a reason recorded
“I stopped reading whole drafts. I read the issues list, I click on the two citations I don’t believe, and I’m usually wrong to doubt them.”

What surprised them
The overrides were the most useful part.
Every time a partner rejected a suggestion, Vellcite asked why. After eight weeks the group had 61 written reasons, most of them things the playbook had never said: a client relationship that changed, a regulator that started asking a new question.
Those reasons now sit next to the precedent they overrule.
“The partners argued with it, and the arguments became the new playbook.”
