consultance.ai
Finance and data

Finance AI Model Map

Most finance teams picked one model and run everything through it. That is the mistake. This is the routing map: which model wins which finance job and why. Diligence is Claude, distribution is GPT, design is Gemini. The breakdown, the gaps, and a starter prompt for each lane, so two of your three jobs stop feeling mediocre.

Rent it forever, or own it once.

Most finance teams picked one model and run everything through it

Step 1

Paste this setup prompt — Claude installs it for you

Easy mode · paste this into Claude

Claude installs it for you, step by step.

Never used Claude before? It is free to start. Open it in a new tab, copy the prompt, paste it in. It asks one question, then walks you through everything.

  1. Step 1
    Open claude.ai ↗

    Sign up free. No card. Takes 30 seconds.

  2. Step 2

    One click. Lands on your clipboard.

  3. Step 3
    Paste + send

    Claude asks what you need + guides you the rest of the way.

Open claude.ai ↗
Tune the prompt for your level (optional)
Preview the prompt (you do not need to read it)
I want to SET UP the three-model finance stack the right way so each job runs on the model that wins it, not all of it through one chat window. Walk me through it step by step. Treat me like a deal-team lead or fund operator who has never deliberately routed work across models, and define every term once.

This is NOT a coding install. There is nothing to compile. It is three accounts and a routing habit. The whole point: diligence on Claude, distribution on GPT, design on Gemini.

## Step 1 — Stand up the three accounts
Walk me through ONE at a time, waiting for me to confirm each:
1. **Claude (the underwriter).** Pin **Opus 4.8**. Create a private **Project** for diligence — a Project is a private workspace with its own knowledge other chats cannot see. Your data room stays in your tenant. This is where rent rolls, T12s, LBO math, QoE, and IC memos run.
2. **OpenAI GPT (the operator).** Use it for sourcing, outreach, CRM, and LP updates. Connect the integrations/actions to the CRM and outreach tools you already use, so it runs the follow-ups, not just drafts them.
3. **Google Gemini (the designer).** Use it for charts, maps, teasers, and deck visuals. Keep the writing on Claude.

## Step 2 — Route by job, not by habit
Before you open a chat, name the lane:
- Numbers, documents, a deal decision → **Claude**.
- Reaching people, CRM, updates at scale → **GPT**.
- Visuals, a deck, a chart → **Gemini** (then Claude for the words).
If you are not sure, ask "is this a judgment call on numbers?" If yes, it is Claude's.

## Step 3 — Keep one source of truth
Whatever GPT or Gemini produces, the figures tie back to the Claude-built diligence model. The underwriting model is the number of record. Marketing and outreach never invent a figure the diligence did not produce.

## Step 4 — Privacy
For client, manager, or deal data, use each tool's private/enterprise workspace (e.g. a Claude Project, an enterprise tenant) so nothing leaves your control. Never paste a live data room into a public window.

## Rules for walking me through this
- One step at a time. Define each term once: model, Project, integration/action, connector, pin, lane, tenant.
- Do NOT tell me a step is "not possible." If an integration is missing, tell me what to check or the manual fallback.
- This is a routing habit plus three accounts. The upgrade — wiring the right model under each task inside your deal pipeline so the routing is automatic — is a separate build, not part of this setup.

First message: ask me "which lane hurts most right now — diligence, sourcing, or deck design? — and which of the three models do you already have?" Then start with the lane I name.
where it breaks

Before you connect live data

  • • Run dummy data first. Real client data is not a test bed.
  • • API keys never go in a public repo. Use env vars and a secrets manager.
  • • Add logging, access control, monitoring, and a rollback path before launch.
  • • Read the license. Forking a repo without checking is how lawsuits start.
license note

Credit the original author

Guide authored by consultance.ai. Claude, OpenAI GPT, and Google Gemini are referenced as the model that wins each lane; model labels are the lane, not a version lock, pin the current frontier of each. Run it in your own accounts; nothing is hosted by us. This is not investment advice.

the newsletter

AI news worth opening.

The AI tools, launches, and shifts that actually matter, in plain English. New library drops the moment they land.

100% freeNo paywall, everUnsubscribe anytime

Want this running in your business, not just your laptop? We build it and hand you the keys.