PreciseFP MCP connects PreciseFP directly to your Claude or ChatGPT account. Once connected, you can ask questions about your clients and prospects in plain language, analyze your whole book of business and take action, all without clicking through screens or building advanced searches.

MCP stands for Model Context Protocol. It is a secure way for an AI assistant to work with the data in another application, in this case PreciseFP.

Beta: PreciseFP MCP is currently in beta, so features and wording may change. We would love your feedback at support@precisefp.com.

PreciseFP MCP is available to current PreciseFP customers. If you have not yet requested beta access, email support@precisefp.com and we will get you started.

• A PreciseFP login

• Beta access to PreciseFP MCP

• A Claude or ChatGPT account

[Confirm any plan or subscription requirements for Claude and ChatGPT before publishing.]

1. Request beta access. Email support@precisefp.com and ask for MCP access. The team will turn it on for your account from the back end.

2. In Claude or ChatGPT, go to the plugins area, search for PreciseFP and add it as a connector.

3. Sign in with your PreciseFP login and approve access when prompted.

4. Start a new chat and ask a simple question to confirm the connection works.

Example: Show me the contact details for [client name].

Tip: Everything shown in the PreciseFP MCP webinar works the same way in Claude and ChatGPT. You simply use whichever one you already have.

Use plain language to look up details about a client or prospect.

Show me [client name]'s contact details and family members.

Has [client name] opened or finished the engagement I sent?

Searches like these used to require advanced search filters with AND and OR operators. Now you can describe what you want.

Show me engagements that have stalled.

Show me all my clients who have at least two children.

Show me clients with no life insurance.

Show me clients with no financial goals on file.

Missing goals are one of the most common reasons a client's data quality score is low, and that score relates to Know Your Client requirements. A list like this shows you where the gaps are.

Good to know: When you ask about clients, your assistant may leave out prospects and archived accounts and will usually tell you so. If you want them included, say so in your request.

After your assistant gives you a list, you can ask it to act on that list. A phrase like "send them" refers to the clients it just showed you.

Can you send them an insurance engagement?

Can you send them a goals engagement?

Tip: Review the list and the engagement before you send. Engagements go to the email addresses on file, so confirm they are correct first.

If your contacts live in a spreadsheet, you can ask your assistant to create them in PreciseFP.

1. Prepare your file with one row per person and a header row. Helpful columns include name, email, phone, date of birth, address, marital status, occupation, income and notes.

2. Attach the spreadsheet to your chat.

3. Ask your assistant to import it.

Import the contacts from this spreadsheet into PreciseFP and include the notes as comments on each record.

4. Check the results.

Show me the prospects I just imported and their comments.

Tip: Try a small list of a few contacts first so you can see how your columns are mapped. Include a notes column if you want that context to come along with each record.

• Duplicates are caught. If a contact is already in PreciseFP, your assistant will tell you instead of adding a second record.

• Larger imports take a little longer. A four-contact import took about a minute and a half in the webinar.

• Bringing in a list from your CRM? If your CRM can export a spreadsheet, you can give that file to your assistant to import. We recommend trying it with a small list first.

Once your prospects are imported, you can ask your assistant to use their comments to suggest what to ask first.

Review the four prospects we just imported, including their comments. Give me the three most useful discovery questions for each person.

Your assistant can suggest questions based on what is in your PreciseFP account.

Based on what's in my PreciseFP account, what other useful questions should I be asking you? Give me ten, grouped by ask, analyze and act.

• Actions take seconds to a few minutes. Your assistant shows a timer so you can see how long a task is taking.

• Your assistant may warn you before repeating an action, for example if you already sent the same engagement to the same people recently.

• Stalled engagement counts rely on the dates an engagement was sent and opened. The connection does not report the last edit date, so an older opened date alone cannot confirm that a client is inactive.

• Keep a person in the loop. Review lists, questions and engagements before you send anything to clients, and make sure the results match your firm's approach.

• Integrations. You cannot yet ask your assistant to use a PreciseFP integration, such as exporting records to Wealthbox or Redtail. Continue to do that from PreciseFP as you do today.

If there is something you want PreciseFP MCP to do, tell us. Feature requests during the beta help shape what we build next.

• Be specific. Name the client, the status or the time period you care about.

• Build on the last answer. Follow-up requests like "now show only the ones in Arizona" work on the list you just received.

• Ask for the details behind an answer. If something looks off, ask your assistant which records it used.

• Review before you act. Check lists and engagements before sending anything to clients.

• Start small. Try a few records first, then scale up once you are comfortable.

Client data accessed this way follows your existing Claude or ChatGPT contract. Whatever your contract says about how your data is used or protected continues to apply when you use PreciseFP MCP.

If your broker-dealer allows you to use Claude or ChatGPT, using them with PreciseFP is not expected to be a problem. We recommend checking with your compliance team and your firm's AI and data policies before you connect.

Confirm that your beta access has been approved, then check that you have added the PreciseFP connector in your Claude or ChatGPT settings. If it still does not appear, contact support@precisefp.com.

Sign out of PreciseFP, sign back in, then try connecting again. Make sure you are using your PreciseFP login.

Try rephrasing with more detail, then ask your assistant which records it looked at. If the result still looks wrong, compare it with what you see in PreciseFP.

Check that your file has a single header row with clear column names. You can also ask your assistant which columns it matched to which PreciseFP fields.

Email support@precisefp.com. Tell us what you are trying to do and what questions you want to ask. Your ideas help shape what we build next.