Book Your Call
One call. Clear direction. Faster execution.
How an advisory firm put safe AI rules in place in 4 weeks.
Staff were already pasting client data into free AI tools, and leadership was close to banning AI altogether. Skillfi wrote a clear AI use policy, reviewed every tool's terms, set up approvals and trained every team, so the firm could use AI safely instead. It's how our legal & compliance service works: practical rules that protect you and let your people work.
AI was in use. The rules were not.
The firm handles sensitive financial data for hundreds of business clients, under contracts with strict confidentiality terms. An internal survey showed 63% of staff were using free AI tools at work, often to summarise client documents and draft reports.
Nobody knew which tools kept or trained on what was typed into them. Clients started asking about AI in contract renewals and proposals, and the firm's insurer wanted to know its controls. Some partners wanted to ban AI completely, while others worried the firm would fall behind competitors already using it well.
Our clients trust us with their most sensitive numbers. We had no idea where that data was going once someone pasted it into a chatbot.
Managing Partner, client
Don't ban AI. Make the safe way the easy way.
We built rules people could follow in seconds, approved tools that were safe for client work and trained everyone on what changed. Hands-on training on the approved tools came through our Claude workshops.
Clear rules for every kind of data
A short AI use policy with green, amber and red data levels, showing exactly what can go into which tools, plus rules for checking AI output before it reaches a client.
Every tool's terms reviewed
All nine tools checked for data retention, training on inputs, security and data location. Three business-grade tools approved and six blocked.
A simple path for new ideas
An AI review panel and a one-page request form, so new tools and uses are approved in days instead of being used in secret.
From no rules to safe AI
A board-approved policy in four weeks, then rollout and training across all four offices. Here's how the engagement unfolded.
- Week 19 tools found
AI use & risk assessment
Staff survey, tool inventory and a review of client contracts and insurance terms that mention data or AI.
- Week 23 approved
Vendor terms review
Every tool's privacy, security and data use terms reviewed, with three business-grade tools recommended.
- Week 31 clear policy
Policy & approval process
The AI use policy, data levels, review panel and client disclosure statement drafted with the partners.
- Week 4Board approved
Board approval
Policy, tool list and rollout plan presented to the board and approved in one meeting.
- Week 5–8100% trained
Rollout & training
Approved tools switched on, unapproved ones blocked, and every person trained and signed off on the policy.
- OngoingQuarterly review
Review & improve
New tool requests reviewed by the panel, and the policy updated every quarter as AI rules and tools change.
Rules people follow in seconds
Four parts that make safe AI use simple for staff, clear for clients and visible for leadership.
Green, amber, red
- Public info: any approved tool
- Internal: business tools only
- Client data: approved only
Safe tools, switched on
- 3 business-grade tools
- No training on firm data
- Risky tools blocked
A human checks it
- Review before client use
- No AI-only advice
- Client disclosure statement
Approved in days
- One-page request form
- AI review panel
- Incident reporting
Three months after kickoff
Staff using unapproved AI tools
While use of the approved AI tools grew to 81% of staff.
We nearly banned AI. Instead we have clear rules, safe tools and a team that uses them well. When clients ask how we handle AI, we now have an answer that wins their confidence.
Managing Partner, clientIs your team using AI without clear rules?
Our legal & compliance team will write a practical AI use policy, review your tools' terms and set up simple approvals, so your people can use AI safely. Want security controls checked too? Add a fractional CISO.
- An AI use & risk assessment
- A clear AI use policy
- AI vendor terms review
- Approvals, training & sign-off
30 minutes. No obligation.
Related Case Studies