Back to the Library
Library · Guide

The Ideal Taxpayer Profile (ITP) Lead Finder

Describe the kind of client you want in plain English. An AI agent turns it into a sharp profile, goes and finds real prospects who match, and logs them to your CRM (or Notion). You review and reach out.

No spreadsheet, no code, no list to bring. You describe who your best future client looks like. The agent does the looking.

Want it to just run? This workflow ships as a skill: lead-finder. Invoke /lead-finder [your ITP] in Claude Code or a Claude.ai Project and it runs every step below with its own web and CRM tools. This page is the explainer and the per-CRM setup behind it.

New idea, borrowed from sales: the best firms do not market to "everyone." They define an Ideal Customer Profile and aim everything at it. For us it is an Ideal Taxpayer Profile, the specific kind of taxpayer or business you are best at serving and most want more of. Get the ITP right and the sourcing, the message, and the close all get easier.

Read GUARDRAILS.md first. This runs on public information about prospects, not client data, so the §7216 exposure is low. The rules that do apply (public-data-only, privacy for individuals, CAN-SPAM, your state board's advertising rules, and you-are-the-sender) are in The rules below.


How it works (three steps, all in plain language)

Step 1, describe your ideal taxpayer

Answer a short set of plain-English questions (the worksheet in itp-and-sourcing.md). For example:

You do not need to know the perfect wording. The agent sharpens it for you.

Step 2, the agent builds your ITP and sources matching prospects

The agent turns your description into a structured profile (who they are, the signal that means they need you, where to find them, and who is NOT a fit), then searches public sources for real organizations that match. For each prospect it returns:

It never invents a prospect, a fact, or a contact. If it cannot verify something, it says so.

Step 3, it logs them to your CRM or Notion

The agent writes each verified prospect into your pipeline as a new lead, with the fit, the signal, the source, and a status of "to review." It connects to your CRM through an MCP link (the secure bridge between your AI tool and your CRM). The workflow file has one-time setup and a logging prompt for Notion, HubSpot, and Airtable, plus a CSV fallback for practice-management tools (TaxDome, Canopy, Karbon, Keap) that do not offer an agent connection yet.

Then you do the part only you can do: review the list, decide who is worth your time, and reach out in your own voice.


The rules

This is a safe first use of AI because the sourcing runs on public information. Two kinds of rules apply: how you handle the prospect data, and how you solicit once you have the list. Each is documented with primary-source cites in REGULATORY-FOUNDATION.md.

Handling the prospect data

Soliciting once you have the list (the part most people under-think)


What's in this folder

File What it is
README.md This overview
itp-library.md Ten ready-made Ideal Taxpayer Profiles to adapt instead of starting from blank
source-playbook.md Where to look, by prospect type: the best public sources for each kind of client
itp-and-sourcing.md The worksheet, the three agent prompts (build the ITP, source the leads, log them), and per-CRM logging setup for Notion, HubSpot, and Airtable (plus a CSV fallback)

Built from a real CPA outreach pipeline, redesigned so you start from who you want, not from a list you have to find first.

The AI Lab for Accountants · An educational resource, not legal or tax advice.