Dilarang Menggunakan Cara Ini Kalau Mau Kaya dengan AI #KPODS

Overview

The video is a discussion hosted on KPODS with media/AI practitioner Edward Santosa about how to use AI effectively—especially for business and creative work—while staying critical and mindful.

Key Takeaways

1) Most “not working” AI results come from poor prompting and mismatched expectations

• Santosa argues that many people already know what they should do; the real problem is often execution (laziness, indecision, or misunderstanding).
• He claims negative chatter about AI (“AI output isn’t good”) is frequently caused by:
  • not understanding prompting, and
  • not providing the right inputs/templates,
  • leading to generic output or “slop.”
• Core idea: AI results are largely input → process → output—so weak inputs produce weak outcomes.

2) Adoption is high in Indonesia, but “quality of use” may lag

• He cites a Microsoft–LinkedIn study on generative AI adoption:
  • Indonesia: 92%
  • Global: 75%
  • Japan: 32% (lowest)
• His interpretation: Indonesia’s high adoption is likely driven by AI being cheap and accessible, not necessarily by teams using it with high-quality strategy.
• He warns that quantity does not equal quality.

3) Use AI first for efficiency—starting with sales and iteration

• Santosa uses a “water pipe” metaphor:
  • Many entrepreneurs are in “survival mode,” where the “pipe” is still dripping.
  • Priority number one: sales to keep revenue flowing.
  • Optimization comes later, after collecting enough data and feedback.
• He emphasizes that optimization requires data:
  • Without enough trials (e.g., selling to 10 vs. 100 people), you can’t reliably diagnose what’s wrong.

4) AI can replace tasks, not necessarily taste or human judgment

• In design/creative production:
  • AI can speed up technical execution and can produce professional-looking work when guided well.
  • But taste, references, and “humanistic” direction still matter.
• He reframes a common fear:
  • Misconception: “AI replaces humans.”
  • More accurate: “People who can use AI effectively may outcompete those who don’t.”
• Skilled designers who pair AI with strong references can achieve time savings and comparable or higher quality.

5) Efficiency gains can be misused—saved time may turn into more work

• Some workers report saving hours with AI but then being assigned even more tasks.
• The result depends on whether the organization treats AI time savings as real productivity gains or uses them to expand workload.

6) AI for business should reduce risk; humans still matter in sensitive moments

• For customer chatbots, Santosa highlights that AI should support:
  • emotional escalation handling, and
  • handoff to humans when irritation/frustration appears.
• He stresses the ongoing need for a crucial human-to-human component for trust and relationships.

7) Privacy/security: turn off training-on-your-data where possible; avoid sharing sensitive info

• Many AI services have data usage for training enabled by default.
• For corporate usage, he recommends private/internal servers to keep data off external clouds.
• He suggests using dummy data to avoid exposing confidential IDs, account details, or proprietary information.
• He warns not to assume chats are fully private, citing concerns about real-world data leaks.

8) A “fundamental” mindset: use AI as an advisor/counsel, not just an answer machine

• People often misuse AI by seeking only direct answers.
• Better approach: use AI to generate better questions and broader perspectives.
• He introduces/endorses the idea of “AI Council”:
  • Instead of one AI voice, simulate multiple roles (e.g., CEO/CMO/CFO) to examine decisions from different angles.
  • Positioned as a way to support better decision-making for MSMEs.

9) Personal productivity and mindset: train boredom; build range, not just specialization

• Santosa discusses consistency:
  • Boredom can be trained gradually (e.g., reading a few pages daily).
• He promotes a “range”/broad-view learning strategy:
  • Learning multiple skills/spokes can outperform narrowing too early.

10) Business outputs: his organizations and offerings

• His main ventures include:
  • A media platform: “7%” (7%.co)
  • An AI University / AI training and integration business (AIU)
  • An SaaS for insurance agents: Instructuractive.com
  • An AI-run crypto media product: Cryptojen
• His model: start small, learn from real use cases, then scale into training/workshops.

Presenters / Contributors

• Edward Santosa — founder; AI practitioner; guest speaker
• Sandra — host/interviewer (KPODS)