To request access, please have the admin user of the account contact support@precisefp.com.

PreciseFP holds a rich financial dataset on every household you serve — income, assets, liabilities, insurance, retirement plans, employment. Analyze is about asking questions of that data as a whole, rather than opening records one at a time.

This is the pillar that tells you who your clients actually are.

If you haven’t connected yet, start with How to activate PreciseFP MCP.

How many accounts do I have?

What is the average net worth of my accounts?

What percentage of my clients have a mortgage?

How many of my households have life insurance?

What’s the average age of my client base?

Which of my clients are within five years of retirement?

Completeness is its own useful question.

Which accounts have the least profile data?

What’s the average data quality score across my clients?

Which households are missing income information?

Every account carries a quality score — a percentage reflecting how many profile fields are populated. A newly created account with contact details only sits around 24%. A fully loaded single-client profile rises to roughly 67%. It’s the fastest signal for how much of a record is actually filled in.

Which of my highest-net-worth clients haven’t had an engagement sent in the last six months?

Show me prospects in the Discover stage with no income data.

These are the questions that are genuinely hard to answer in the web app, and they’re where Analyze earns its place.

If some of your accounts have full profiles and others are contact-only shells, an average across all accounts is misleading — the empty ones drag it down.

Ask for the metric over accounts that have data:

What is the average net worth of accounts that have net worth data, and how many accounts is that based on?

A well-behaved answer tells you both the figure and the sample size. If it doesn’t, ask. “66% of my households with data have a mortgage” is a real answer; “66% of my households have a mortgage” may not be.

Analyze is where the connector’s architecture actually matters, because reading financial data is a two-step operation.

The schema-first model. There is no way to read a financial field by name in one call. The field’s identity has to be resolved against your live PreciseFP dataset schema first, which returns opaque UUID keys for the panel, group, and element. Only then can values be read by reference to those keys.

In practice:

This is why “what’s the average net worth?” involves more work than it looks like — and why it follows your tenant’s actual schema, including any customization you’ve made, rather than a fixed assumption about field names.

How the profile is structured. Profiles are hierarchical: panels → groups → elements. Panels are the top-level sections (Client Details, Employment & Income, Assets, Liabilities, Insurance, Retirement Plans, Expenses & Taxes, Cash Flow & Net Worth). Groups are subsections. Elements are individual fields.

Two properties of that structure shape what you get back:

So “how many of my clients have a mortgage?” is a question about rows in a MULTIPLE group on an ACCOUNT panel, while “what’s the average age of my clients?” reads a field on an INDIVIDUAL panel keyed per member.

Calculated fields. A large share of fields are CALCULATED — PreciseFP derives them from other inputs. Employment Income and Total Income are summed from salary and bonus. Monthly insurance premiums are divided from annual premiums entered elsewhere. The entire Net Worth panel is All Assets minus All Liabilities.

This is good news for Analyze: net worth is computed for you as soon as assets and liabilities are loaded. It matters more for Act, where trying to write a calculated field is wasted effort.

Coded values. Dropdown fields store internal codes rather than display labels — marital status “Domestic Partner” is stored as partner, a fixed mortgage is fixed, a “Universal Life” policy is universal. Worth knowing if an answer comes back with a code you don’t recognize.

Read limits. accounts-profiles handles up to 100 accounts per call, and accounts-activities the same. For a book larger than that, the work runs across multiple calls.

Why is my average net worth lower than I expected? Almost always the completeness gap — shells with no financial data are being counted. Ask for the metric over accounts with data, and for the count it’s based on.

It says Net Worth is empty, but I’ve loaded assets and liabilities. Net Worth is calculated by PreciseFP from those inputs. Give it a moment and re-read the account; if it’s still empty, check that the underlying asset and liability rows saved correctly.

Can I export the analysis? Your AI tool can produce a table or a file from the answer. The connector itself doesn’t export — it returns data to the tool you’re working in.

Is analyzing my client data safe? Reads are scoped to you and your organization, use your existing PreciseFP permissions, and never return SSNs. Do bear in mind that whatever you ask about is then part of that conversation in your AI tool — treat those chats as you would any document holding client data.