Lesson 06 · ChatGPT Mastery Pro+ ~14 min read 3 workflows + safety patterns

GPT-5 agents: workflows that run themselves.

GPT-5's agent capabilities let ChatGPT run multi-step workflows autonomously — visit websites, fill forms, take actions. It's the most powerful agentic AI in consumer software today, and the most risky if misused. This lesson covers what agents do well, the supervision patterns that prevent disaster, and the cost-control playbook.

The mental model

Agents are interns with internet access. Treat them accordingly.

GPT-5 agent mode can do remarkable multi-step work autonomously. It can also confidently complete the wrong task, spend tokens fast, and make irreversible changes. The skill isn't using agents — it's supervising them well.

Safety first

Never give an agent access to anything you wouldn't give to a brand-new intern. No banking. No production systems. No accounts where one wrong action matters. Use throwaway browsers, throwaway accounts, public data. The capability is real; the trust required is still being earned.

Workflow 01 Research and synthesis at scale

1

The 30-source research session

Give the agent a research question. It visits 20-30 sources, synthesizes, and reports back.

The prompt that works

Research promptResearch the current state of the AI training market for enterprise B2B. Visit at least 20 sources. Synthesize: market size, top players, pricing models, recent funding. Cite all sources.

Best use cases

  • Pre-meeting prep with depth
  • Industry analyses
  • Competitive intelligence
  • Trend reports
Agents can spend a lot of tokens on long research. Set a budget. Spot-check sources for credibility — agents will sometimes cite SEO-spam.
Time savings: Hours of manual research → 10-15 min of agent run + your review.

Workflow 02 Multi-step web tasks

2

Find, compare, recommend

Tasks that require visiting many pages, comparing options, and recommending one.

The prompt that works

Multi-step promptFind the best business credit card for a small SaaS company with $50k/mo spend, lots of software subscriptions, and a need for travel rewards. Compare at least 8 options. Rank top 3 with reasoning.

Best use cases

  • Vendor comparison and selection
  • Product research with specific criteria
  • Travel planning with constraints
  • Anywhere you'd otherwise compare 10+ options manually
Agent recommendations are inputs to your decision, not the decision. Always verify the agent's claims against current source data before acting.
Time savings: Comparison-research hours: minutes.

Workflow 03 Periodic monitoring

3

Set-and-monitor weekly tasks

Schedule an agent to perform a weekly task: check 5 specific URLs for changes, summarize industry news, audit accounts.

The prompt that works

Monitoring patternEvery Monday at 8am: visit our top 5 competitor blogs. Summarize any new posts. Flag anything strategically relevant.

Best use cases

  • Competitor monitoring
  • Industry news roundups
  • Inventory or pricing tracking
  • Periodic compliance/audit checks
Sites change layout; agents break silently. Audit your monitoring setups monthly to catch silent failures.
Time savings: Weekly manual monitoring: automated.

Final challenge: one supervised agent task

Identify one research-or-monitoring task you do manually that takes 30+ min. Use GPT-5 agents for it. Watch it carefully the first time. Verify every claim. Iterate on the prompt. If it works, schedule it. If it doesn't, learn why and try a different task type.

What you can do now

  • Understand what agents do well vs. poorly
  • Apply safety patterns (throwaway accounts, supervised first runs, no irreversible actions)
  • Use agents for research, comparison, and periodic monitoring
  • Set cost budgets and verify before acting on agent output
  • Recognize when scripting (Python/zapier) would be more reliable than agents
Pro
Up next in ChatGPT Mastery

Lesson 7 · Search, image, & DALL-E: the multimodal features most users underuse

Web search inside ChatGPT, image analysis, DALL-E for real work (not just art). The workflows that turn ChatGPT into a Swiss army knife. See pricing →