The Next Input — Issue #094

In partnership with

⚡ The Briefing — 60 sec

• Three years since ChatGPT’s launch
  Hard to believe it’s been three years since the iconic “I’m sorry, but as a model trained by OpenAI…” era. Simpler times. Slower models.

• HP to cut up to 6,000 jobs to ‘focus on AI’
  Would be nice if, just once, the headline was “6,000 executives replaced by AI.” But no… it’s never that story.

• Aussie teenager charged over erotic roleplay with ChatGPT
  Erotica is one thing, but the model saying “heckin” like a 2015 Tumblr dog? That’s the real crime.

🛠️ The Playbook — The Tone Filter: Stop AI From Talking Like a Cornball

Mission Force AI outputs into a consistent brand voice—clean, human, direct, no cringe.
Difficulty Medium
Build time 1.5 hours
ROI Eliminates 90% of tone problems before they ever reach a client or audience.

0) Why This Matters

Models default to whatever tone they think is “safe,” which often lands somewhere between corporate oatmeal and weird internet baby-talk.
If your brand has an actual voice—straightforward, modern, human—you need a filter that catches cringe before it escapes into the wild.

1) Architecture

2) Workflow

1. Create a “Tone Rules” table in Airtable with:

   • allowed tone markers

   • banned words

   • example good vs bad phrasing

2. User drops any draft (email, LinkedIn post, newsletter, product copy) into Notion or a folder.

3. Make/Zapier detects the new file and sends it to Claude.

4. Claude rewrites the text according to your brand rules.

5. GPT-5-mini runs a second pass to:

   • highlight any lingering cringe

   • flag unnatural phrases

   • catch sudden tone shifts

6. Final output is delivered back to Slack or email for use.

3) Example Prompts

Rewrite Prompt (Claude 4.5 Sonnet)

Rewrite the text in a clean, modern, confident tone.
Rules:
- No cutesy, nerdy, or overly corporate phrasing.
- No slang that feels forced or outdated.
- Keep sentences tight and punchy.
- Maintain the writer’s intent, not the model’s habits.
Banned terms: heckin, uwu, delightful, leveraging synergies, etc.
Output only the rewritten text.

Tone Validation Prompt (GPT-5-mini)

Scan the text for tone issues:
- cringe phrasing
- inconsistent voice
- forced slang
- corporate filler language
Suggest minimal corrections to restore a clean, human tone.

4) Guardrails

• Don’t remove necessary nuance in legal or technical content.

• Keep rewrites structurally identical—this is tone, not meaning.

• Always surface banned terms clearly.

• Never add slang the user didn’t imply.

5) Pilot Rollout — 90 minutes

1. Build your Airtable “Tone Rules” table.

2. Add 10 “bad → good” examples to train the engine.

3. Connect a Notion or folder trigger to Make.

4. Run five internal drafts through the pipeline.

5. Compare rewritten versions with originals and adjust banned terms.

6. Turn on auto-processing for daily use.

6) Metrics

• Percentage of drafts requiring manual cleanup afterward

• Number of banned terms caught per week

• Reduction in inconsistent voice between authors

• Time saved per draft

• Team satisfaction score after rollout

Pro Tip: Update the banned-terms list weekly. If a phrase annoys you once, it’ll annoy you twice.

🎯 The Arsenal — Tools & Prompts

• Writer.com Styleguide · Build tone rules + automated checks · https://writer.com

• Notion AI Workspace · Easy drop zone for drafts + workflow triggers · https://www.notion.so

• Airtable · Store voice rules, banned terms, and examples · https://airtable.com

• Make.com · Automate rewrite flow end-to-end · https://make.com

Copy-paste prompt block:

You are enforcing tone consistency.
Rewrite text to be:
- clean
- modern
- confident
- zero cringe
No forced slang. No childish phrasing. No corporate fluff.
Maintain meaning; upgrade tone.

💡 Free Office Hours

Want help implementing anything? Book a free 15-minute Office Hours slot—no sales pitch, just workflows solved.

→ https://calendly.com/aaron-cylentis/the-next-input-office-hours

Start learning AI in 2025

Everyone talks about AI, but no one has the time to learn it. So, we found the easiest way to learn AI in as little time as possible: The Rundown AI.

It's a free AI newsletter that keeps you up-to-date on the latest AI news, and teaches you how to apply it in just 5 minutes a day.

Plus, complete the quiz after signing up and they’ll recommend the best AI tools, guides, and courses – tailored to your needs.

Sign up to start learning.

🕹️ Game Over

Set the tone once—let the system keep it tight.

— Aaron Automating the boring. Amplifying the brilliant.

Subscribe: https://cylentisai.beehiiv.com/subscribe