The AI Consultant · Fable 5 — the honest cheat sheet for Agostino

Fable 5 is a scalpel, not a hammer

The hype says "build a tool with Fable this weekend." The real research says: use it for the one big thing, keep everything else on Opus. Here's the honest version — on one page.

🔥 Free-ish only to July 7. From July 8: metered $10/M in · $50/M out — 2× Opus. It drains your weekly pool ~2× faster even now.

The verdict from the people using it

"Horrible, but super capable"

A disposable-genius builder. Point it at a bounded outcome, let it rip, take the artifact.— from advisors actively using it, corroborated by independent deep research

It's faster and often better than Opus — but it does whatever it wants: tries things, reverts them, ignores the rules you set. Opus is slower but consistent and rule-abiding. So the skill isn't "use Fable" — it's knowing the few jobs worth pointing it at.

Long & autonomous
its real edge — multi-hour, one-shot builds, not quick tasks
2×
faster token drain than Opus — be selective
July 7
last day it's included; metered after
Opus 4.8
still the right tool for 90% of his week

When to reach for which

Fable vs Opus — the split

✅ Fable 5 — one-shot, disposable-genius

  • Long, autonomous builds you'll keep but not hand-maintain
  • Cloning a paid app to kill a subscription (one member saved £64/mo)
  • Heavy one-burst estate/tax/financial analysis & long policy/legal-doc review
  • A red-team / security pass before you ship something to a client

🛑 Keep it on Opus 4.8 (his $20)

  • Day-to-day case prep, client emails, objection handling
  • The Board of Advisors, the $1,000-prompts run
  • Anything that must obey a strict process or be reused with clients
  • Simple, single-turn tasks — Fable adds nothing but drain

The test before you spend a Fable token

Use Fable only if all three are true

1

Bounded outcome

You can say "done = X works" — not an open-ended system you'll babysit.

2

Lasting leverage from one burst

A single autonomous run gives value you keep — a tool, a killed subscription, a reusable demo.

3

You don't need to control how

You won't hand-maintain the code line by line. If any of these is false → Opus.

If he wants to use it well

The one or two things worth Fable — for him

All bounded, one-burst, "build once and keep it" — the only shape that earns the tokens.

📊

Full Case & Estate Analyzer

One big build: model estate value with vs. without cover, tax exposure, and the meeting script — a tool he reuses every client.

Bounded · reusable · his best fit
💸

Clone an app he pays for

Point Fable at a paid TTS / transcription / meeting-recorder → "build me a local copy." Kill the monthly bill for good.

Perpetual ROI from one burst
📚

Long financial / legal review

Fable's proven strength: senior-grade analysis over long policy docs, illustrations, and legal redlines in one pass.

Genuinely beyond Opus here
🛡️

Security pass before shipping

Let it red-team anything he's about to put in front of a client. Its rule-breaking becomes a feature.

Cheap insurance
Guardrail — because it breaks things as it goes: every Fable build gets a regression test suite. Tell it to write the tests and run them on every change ("regression is the magic word"). Never ship a Fable build to a client without it.

The plan question — the honest answer

He probably doesn't need to upgrade

Pro
$20 /mo

His current plan runs Opus 4.8 — the right tool for ~90% of his week. For day-to-day, this is genuinely enough.

Max 5×
$100 /mo

Only worth it if he commits to one big Fable build before July 7. Otherwise it's headroom he won't use.

Max 20×
$200 /mo

Not for him. That's for people running heavy back-to-back agentic builds.

The honest take: no need to rush into $100. If he has one ambitious build he'll keep (the case analyzer, or cloning a paid app), upgrade for the window, do it, then drop back. If he doesn't — stay on $20, run everything on Opus, and revisit when he has a real target. From July 8 it's pay-per-use anyway.

Run this first · ~5 min · in Opus 4.8

Find his personal $1,000 prompts

A fast, tangible win that runs on his current plan — best if he feeds his exported ChatGPT history first.

The instant-value prompt
Act as a senior AI strategist who understands high-value sales work. Using everything
you know about me and my business as a life-insurance / HNW estate-protection advisor
(read my attached ChatGPT export and infer the rest), find my personal "$1,000 prompts"
— prompts that, run regularly, would each be worth $1,000+ because they win/save a deal,
remove hours of work, or sharpen a client conversation.

1. Map where the money and time sit in my week.
2. Give me 8-12 candidate prompts (exact text + the job each does + why it's worth $1,000+).
3. Rank by value-per-effort; name the top 3 to start this week.
4. Write the #1 fully, tailored to me — ask me at most 2 short questions if you truly need them.

Be specific to insurance sales — no generic filler. Prompts a top advisor would guard.
How this connects: this is the honest one-page briefing — the full step-by-step playbook (Board of Advisors, Command Center, the Foundation setup, the build engine) lives on the main guide.
→ Open the full playbook