Bonus Section

How to Build a Custom GPT

Turn every prompt in the Advisor OS library into a standalone Custom GPT that your entire practice can use — no technical skills required.

What is a Custom GPT?

A Custom GPT is a purpose-built version of ChatGPT that has been pre-configured with specific instructions, knowledge files, and conversation starters. Instead of pasting the same prompt every time you start a new chat, a Custom GPT remembers everything permanently.

Think of it this way: the Prompt Library gives you the blueprints. A Custom GPT is the finished machine — built once, used forever. Your team opens it, clicks a conversation starter, pastes their input, and gets a structured, compliant, on-brand output every time.

Below you will find a step-by-step guide for building any Custom GPT, followed by a complete blueprint for every prompt in the Advisor OS library. Each blueprint tells you exactly what to name it, what to paste, what files to upload, and what conversation starters to add.

8 Steps to Build Any Custom GPT

01

Open ChatGPT and Navigate to GPTs

Go to chat.openai.com → click your profile icon (bottom-left) → select 'My GPTs' → click '+ Create a GPT'. You need a ChatGPT Plus or Team subscription to create Custom GPTs.

Tip: If you are on a Team plan, you can share GPTs with your entire practice. This is the recommended setup for advisory firms.
02

Name Your GPT

Give it a clear, functional name that your team will recognise instantly. Use the naming convention: [Practice Name] + [Function]. For example: 'attooh! Pre-Meeting Prep Engine' or 'attooh! Compliance Scanner'.

Tip: Avoid generic names like 'My Helper'. Specific names help your team find and trust the right GPT for each task.
03

Write the Description

In the Description field, write one sentence explaining what this GPT does. This appears in the GPT gallery and helps your team understand its purpose at a glance.

Tip: Example: 'Prepares structured pre-meeting briefs for client review meetings using practice-specific context and compliance awareness.'
04

Paste Your Instructions (The Prompt)

This is where the magic happens. In the 'Instructions' field, paste the full prompt from the Buy Back Friday™ Prompt Library. This becomes the GPT's permanent system prompt — it governs every response the GPT produces.

Tip: Always include your Advisor OS Master Prompt as the foundation, then add the specific task prompt below it. This gives the GPT both your practice identity and the task-specific instructions.
05

Add Conversation Starters

These are pre-written prompts that appear as clickable buttons when someone opens the GPT. Add 3–4 starters that reflect the most common use cases. For example, a Pre-Meeting Prep GPT might have: 'Prepare a brief for a retirement review meeting' or 'Generate a pre-meeting brief for a new client onboarding'.

Tip: Good conversation starters reduce friction. Your admin team can click and go instead of figuring out what to type.
06

Configure Knowledge Files (Optional)

You can upload reference documents that the GPT can search when generating responses. Upload your Tone Blueprint, practice SOPs, product comparison sheets, fee schedules, or compliance checklists. The GPT will use these as context without you needing to paste them every time.

Tip: Keep files under 20MB each. PDF and DOCX work best. Update them quarterly to keep the GPT current.
07

Set Capabilities

Toggle on/off: Web Browsing (for market research GPTs), Code Interpreter (for data analysis), and DALL·E (usually off for advisory GPTs). For most advisory prompts, you only need the default text generation capability.

Tip: Less is more. Turning off unnecessary capabilities keeps the GPT focused and reduces unexpected outputs.
08

Test, Refine, and Deploy

Use the Preview panel on the right to test your GPT with real scenarios. Run 3–5 test cases before sharing. Check that the tone matches your blueprint, the output structure is correct, and compliance flags are working. When satisfied, click 'Save' and choose your sharing setting.

Tip: Share as 'Anyone with the link' for your team, or 'Only me' while testing. Never share client-specific GPTs publicly.

GPT Blueprints — Your Complete Library

14 ready-to-build Custom GPTs. Click any card to see the full build specification.

Pro Tips for Managing Your GPT Library

01

Start with three GPTs

Do not build all 14 at once. Start with the Master Prompt, Pre-Meeting Prep, and Post-Meeting Multi-Deliverable. These three cover 80% of daily advisory work.

02

Use a naming convention

Prefix every GPT with your practice name. This makes them instantly recognisable in the GPT sidebar and prevents confusion if team members have personal GPTs too.

03

Update knowledge files quarterly

Set a calendar reminder to refresh your Tone Blueprint, fee schedules, and compliance checklists. Outdated knowledge files produce outdated outputs.

04

Pin your most-used GPTs

In ChatGPT, you can pin GPTs to the top of your sidebar. Pin your top 3 so they are always one click away.

05

Test before you share

Run at least 3 real scenarios through each GPT before sharing with your team. Check tone, structure, and compliance flags. A poorly configured GPT erodes trust faster than no GPT at all.

06

Create a GPT for your admin team

Your admin team has different needs. Build a dedicated 'Admin Task Engine' that combines the Task Delegation Engine and SOP Generator. This gives them a single entry point for all operational work.