The problem
Most financial advice firms have a client base that looks something like this: a small number of highly engaged, high-value clients who love what you do, a large middle group who are satisfied but not particularly active, and a long tail of smaller clients who take disproportionate time relative to the revenue they generate.
The problem is that most firms do not have a clear picture of which clients are in which group. Without that clarity, you cannot make good decisions about service levels, pricing, capacity, or growth. You end up treating every client the same regardless of value, which means your best clients are probably underserved and your smallest clients are probably over-served.
Proper client segmentation used to require a data analyst, a CRM project, and several days of work. With the right AI workflow, one adviser can do it in an afternoon.
The system
Step 1: Export your client data
From your back-office system (Intelliflo, Curo, or similar), export a client list with the following fields for each client:
- Annual fee or recurring revenue
- Assets under management (if applicable)
- Date of last review meeting
- Number of review meetings in last 12 months
- Products held
- Age
- Time as a client
You do not need every field. Revenue, AUM, and engagement (number of meetings) will get you most of the way there. Anonymise or use client reference numbers rather than names when working in AI tools.
Step 2: Build your segmentation criteria (Claude)
Paste a sample of your data (10 to 15 clients, anonymised) into Claude along with a description of your firm and ask it to help you design a segmentation model:
"I run a UK financial advice firm and I want to segment my client base to help with service level decisions and capacity planning. Here is a sample of my client data (anonymised):
[paste sample data]
Based on this data and typical IFA practice management principles, suggest a segmentation framework with three to four tiers. For each tier, define: what the criteria should be, what service level is appropriate, what the approximate capacity per adviser should be, and what actions I should consider for clients who are not in the right tier."
Review and adjust the criteria to match your firm's philosophy and commercial reality.
Step 3: Analyse your full client base (Claude)
Paste your full anonymised client list into Claude with the agreed segmentation criteria:
"Using the segmentation framework we agreed, classify each of these clients into their tier. Then provide a summary showing: how many clients are in each tier, total revenue and AUM by tier, and any clients who appear to be misclassified relative to their current service level (for example, small clients getting premium service, or large clients who have not had a review in 18 months)."
Step 4: Build an action plan (Notion AI)
Paste Claude's analysis into a Notion page. Use Notion AI to help you turn the findings into a practical 90-day action plan:
"Based on this client segmentation analysis, create a 90-day action plan for improving our client service model. Include specific actions for each tier and any clients flagged for review."
The results
Before: No formal segmentation, decisions made on instinct, inconsistent service levels across the client base.
After: A clear four-tier model with specific service criteria, a list of clients to review or re-segment, and a 90-day plan to address the gaps.
The financial impact is often significant. One firm that ran this exercise identified eight clients who were paying for a premium service level but receiving only basic contact. Upgrading their service experience immediately improved retention risk. They also identified twelve clients where the economics no longer worked at the current fee level, leading to constructive conversations about pricing or offboarding.
Good segmentation is not about valuing some clients less. It is about being honest with yourself about where you can genuinely add value.