🎮 The Next Input — Issue #069

In partnership with

⚡ The Briefing — 60 sec

• OpenAI’s Nick Turley says ChatGPT is becoming an operating system. GPT OS? The browser wars just got a new challenger.

• AI cheating scandal hits Australian Catholic University. If you’re going to blame me, at least get it right...

• Zendesk claims its new AI agent can solve 80% of support tickets. Do we even need people anymore? (Yes. For the 20% it breaks.)

🛠️ The Playbook — GPT OS: Building the “Everything Console” for Your Company

Mission Turn ChatGPT into your company’s operating system—an “everything console” that manages operations, executes automations, and provides insight across all teams from a single interface.
Difficulty Expert | Build time 4–6 hours (pilot)
ROI Replaces dozens of disconnected dashboards with one unified command hub, saving ≈ 20–30 hours per week across teams.

0) Why This Matters

Nick Turley isn’t exaggerating—ChatGPT’s evolution from assistant → operating layer is the single biggest shift in enterprise AI since RPA. Instead of jumping between dashboards, spreadsheets, and tools, your team will just ask GPT OS to do things:

1) The Architecture

2) Workflow

Example: “Create a new support workflow and summarize this week’s Zendesk data.”

1. Input:

   • User asks in ChatGPT (connected via Apps SDK).

2. Plan:

   • GPT OS calls the Action Layer → identifies relevant tools (Zendesk, Notion).

3. Execute:

   • Pulls Zendesk metrics → creates summary in Notion doc → alerts #ops Slack channel.

4. Confirm:

   • GPT checks company Rulebook: “Non-financial action? ✅ Auto-approve.”

5. Memory Update:

   • Logs workflow pattern for future reuse.

6. Follow-Up:

   • Suggests: “Would you like to automate this as a recurring workflow?”

3) Core Prompts

Master Prompt – GPT OS Controller

SYSTEM: You are the Operating System for {Company}.
GOAL: Execute the user's request across connected tools.
RULES:
- Always plan before acting (list steps, then perform them).
- Confirm high-impact or financial actions before execution.
- Update company memory with {workflow, summary, metrics}.
TOOLS: [Notion, HubSpot, Slack, Sheets, Jira, Xero]
OUTPUT: JSON log {tool, action, result, timestamp}

Policy Prompt – Action Governance

Check if this action is approved:
Input: "Delete customer record"
Rulebook: 
  - "Financial actions require CFO approval"
  - "Customer data deletion requires Legal confirmation"
Return:
  {"approved": false, "reason": "Requires Legal sign-off"}

4) Guardrails

• Auth Scopes: Per-user OAuth; only read/write data from approved sources.

• Data Sensitivity: Restrict personal data and financial transactions by role.

• Audit Trail: Log every command + result for compliance.

• Fail-safe: Human confirmation for high-risk actions (delete, post, invoice).

5) Pilot Rollout — 4 Hours

1. Connect 3 core apps (Notion, Slack, HubSpot).

2. Create one company Rulebook (JSON file of policies).

3. Build Custom GPT + AgentKit with those integrations.

4. Test flows:

   • “Summarize this week’s deals.”

   • “Create next week’s team meeting doc.”

   • “DM me tasks overdue in Jira.”

5. Refine confirmations and logging.

6) Metrics

• Avg workflows completed per user/day.

• Approval/exception ratio (how often GPT needs a human).

• Hours saved from dashboard context switching.

• Reuse rate of automated workflows.

🎯 The Arsenal — Tools & Prompts

User request: "Send customer follow-up for deals > $10k."
Return JSON plan:
[
  {"app":"HubSpot","action":"query deals","filter":"value>10000"},
  {"app":"Slack","action":"post message","template":"Follow-up sent!"}
]

💡 Free Office Hours

Want to turn ChatGPT into your team’s command console?
Book a free 15-minute Office Hours slot—no sales pitch, just workflows solved.

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

Find out why 100K+ engineers read The Code twice a week

Staying behind on tech trends can be a career killer.

But let’s face it, no one has hours to spare every week trying to stay updated.

That’s why over 100,000 engineers at companies like Google, Meta, and Apple read The Code twice a week.

Here’s why it works:

• No fluff, just signal – Learn the most important tech news delivered in just two short emails.

• Supercharge your skills – Get access to top research papers and resources that give you an edge in the industry.

• See the future first – Discover what’s next before it hits the mainstream, so you can lead, not follow.

Join 100,000+ engineers who read The Code to stay ahead of the curve.

🕹️ Game Over

Deploy your first GPT OS workflow today—tomorrow, your entire company runs from one command line: ChatGPT.
Share your win; you could headline Issue #070.

— Aaron
Automating the boring. Amplifying the brilliant.

Forwarded this? Subscribe here