🎮 The Next Input — Issue #157

In partnership with

⚡ The Briefing — 60 sec

• ‘Copilot is for entertainment purposes only,’ according to Microsoft’s terms of service Your enterprise usage of Copilot? It’s a laughing matter if you believe Microsoft’s lawyers. The gap between how these tools are marketed and what they’re legally willing to stand behind is… worth noting.

• AI bot party in Manchester shows future of socialising I genuinely laughed at this. But also… people hanging out with bots at a party is one of those “this is dumb” moments that quietly turns into “oh this is just normal now” faster than expected.

• Cognitive debt: Brain scans reveal impact of AI overuse If you read this newsletter this may not apply to you, but still — outsourcing all thinking to AI might come with a cost. Being resistant to AI is risky, but so is switching your brain off entirely.

🛠️ The Playbook — The AI Usage Discipline Engine

Mission
Use AI to amplify thinking and execution without turning your team into passive operators or legally exposed button-clickers.

Difficulty
Intermediate

Build time
3–5 hours

ROI
Stronger outputs, sharper teams, and fewer situations where you blindly trust tools that won’t back you when it matters.

0) Why This Matters

Three angles, one core tension.

First, the legal reality check. If a major enterprise AI tool is effectively saying “don’t rely on this,” that should immediately reframe how seriously you treat its outputs.

Second, the cultural shift. AI isn’t just a productivity tool — it’s starting to show up in social environments, which sounds ridiculous until it suddenly isn’t.

Third, the human side. There’s early evidence suggesting overreliance on AI could reduce cognitive engagement — what some are calling “cognitive debt.”

So the play is not:

• use AI everywhere

• trust everything it says

• automate thinking entirely

The play is:

• use AI deliberately

• keep humans doing the thinking where it matters

• treat outputs as drafts, not truth

• build systems that require engagement, not just acceptance

1) Architecture

2) Workflow

1. Identify where AI is currently being used in day-to-day work.

2. Classify tasks into assist, draft, and decision-making categories.

3. Restrict AI to assist and draft roles for most workflows.

4. Require human reasoning or validation before final decisions.

5. Add lightweight checks to challenge AI outputs, not just accept them.

6. Track where AI improves outcomes versus where it replaces thinking.

3) Example Prompts

Challenge Prompt

You are reviewing an AI-generated output.

Your job is to challenge it.

Check:
- what assumptions are being made
- what could be wrong
- what is missing
- whether the conclusion actually follows

Return 3 critical points.

Second Opinion Prompt

You are providing a second opinion.

Given the original AI response:
- identify weaknesses
- suggest alternative interpretations
- highlight uncertainty

Keep it concise and honest.

Decision Gate Prompt

You are acting as a decision checkpoint.

Before this output is used:
- confirm whether it is safe to act on
- identify risks
- recommend approve, revise, or reject

Return a short justification.

Cognitive Engagement Prompt

You are forcing the user to think.

Before answering:
- ask 3 clarifying questions
- identify what the user should decide themselves
- then provide support, not a full solution

4) Guardrails

• Do not treat AI output as final.

• Keep humans responsible for decisions.

• Use AI to assist thinking, not replace it.

• Challenge outputs regularly.

• Be aware of legal disclaimers on tools.

• Avoid over-automation of judgment-heavy work.

5) Pilot Rollout — 3 hours

1. Pick one workflow heavily using AI today.

2. Map where AI is helping versus replacing thinking.

3. Add a simple challenge or validation step.

4. Require human sign-off on final outputs.

5. Run 10–15 examples and compare quality.

6. Adjust usage rules based on results.

6) Metrics

• Human validation rate

• Error detection rate

• Quality of final outputs

• Time saved vs accuracy maintained

• Instances of blind acceptance

• Team confidence in outputs

• Cognitive engagement score (qualitative)

Pro Tip: The goal isn’t to use AI more. It’s to think better while using it.

🎯 The Arsenal — Tools & Platforms

• ChatGPT / Claude / Copilot · powerful tools, but only when paired with active thinking · https://chatgpt.com · https://www.anthropic.com · https://www.microsoft.com/copilot

• Google Docs / Notion · environments where AI-assisted drafting can be reviewed properly · https://workspace.google.com · https://www.notion.so

• Airtable / Sheets · track usage patterns and quality outcomes · https://www.airtable.com · https://workspace.google.com/products/sheets/

• Internal policy docs · define when AI is assist vs decision-maker · (internal)

• Prompt libraries · enforce structured thinking and validation · (internal)

Copy-paste prompt block:

You are helping me build an AI Usage Discipline Engine.

For the workflow below:
1. identify where AI is currently used
2. classify tasks as assist, draft, or decision
3. identify where human thinking must remain
4. identify risks of overreliance
5. add a validation or challenge step
6. propose a 2-week pilot
7. define success metrics

Workflow:
[insert workflow here]

Return the answer in markdown with sections for:
- Workflow summary
- AI usage map
- Human-only steps
- Risks
- Validation layer
- Pilot rollout
- Metrics

💡 Free Office Hours

If you’re trying to use AI without losing control of your thinking, your team, or your output quality, I run free office hours to help design workflows that actually hold up.

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

700+ teams have Viktor reading their Google Ads every morning.

Your media team opens Slack at 8am. There's a cross-platform brief in #growth: Google Ads spend vs. ROAS, Meta CPA by campaign, Stripe revenue by channel. Viktor posted it at 6am. Nobody asked for it.

Last week, one team's Viktor caught a spend spike at 2am on a broad match campaign and flagged it in Slack: "CPA up 340%. Recommend pausing and shifting budget to the top two performers." That would have burned $3K by morning. The media buyer woke up to a problem already handled.

Your strategist reviews spend trends. Your account manager checks revenue attribution. Same Slack channel, same colleague, before anyone's first coffee.

Google Ads, Meta, Stripe. One message. No Looker, no Data Studio. Anomaly detection runs around the clock. Cross-platform reporting runs on autopilot.

5,700+ teams. SOC 2 certified. Your data never trains models.

"Viktor is now an integral team member, and after weeks of use we still feel we haven't uncovered the full potential." — Patrick O'Doherty, Director, Yarra Web

Start free. $100 in credits →

🕹️ Game Over

AI can think fast. That doesn’t mean you should stop.

— Aaron Automating the boring. Amplifying the brilliant.

Subscribe: link