Using AI Skills in Everyday Task

Manda Putra

TRAE AI @ ILC Timoho, Yogyakarta — 02 May 2026

Why this matters

  • Less busywork, more judgment
  • Faster research, better summaries, more consistent output
  • Works for product workflows and HR workflows

Two things today

  1. What are AI Skills and why we should use that?
  2. Using TRAE SOLO to create automation and using skills for better result

Audience check

  • Product: faster discovery, better specs, clearer communication
  • HR: better screening summaries, policy drafts, report briefs
  • Everyone: more time for human decisions

The story

I’m a bakery owner.

Parsley bakery, Yogyakarta.

I want a brief report of Google Maps reviews.

The pain

  • Reviews are scattered and noisy
  • Reading every review takes time
  • You still need a human decision at the end

What are “AI Skills”?

  • Reusable micro-workflows
  • A capability with a clear input/output contract
  • Helps you get consistent results, not random answers

AI skills ≠ prompting

  • Prompting: one-off conversation
  • Skills: repeatable building blocks
  • Better for teams: easier to reuse and review

Why use AI skills?

  • Consistency: same format every time
  • Delegation: you provide goal + constraints
  • Speed: less time on repetitive work

Good model: human + AI

  • Human: sets the goal and decides
  • AI: drafts, summarizes, structures
  • Human: verifies, edits, ships

Guardrails (important for HR)

  • Privacy: don’t paste secrets, keys, or private personal data
  • Accuracy: treat output as draft, verify sources
  • Bias: ask for assumptions and blind spots

From skills → systems

If a workflow is repeatable, automate it.

TRAE SOLO mental model

  • Agent session: understands your task
  • Skills: plug-in capabilities
  • Automation: schedule it to run without you

Better results = better inputs

  • Context: who, what, where, why
  • Constraints: tone, length, format, audience
  • Review loop: assumptions, edge cases, “what could be wrong?”

Demo goal

Weekly Google Maps review brief for Parsley bakery (Yogyakarta)

  • Short and actionable
  • Written for a busy owner
  • Ends with a checklist

Demo: the skill

Google Maps scraper skill

https://skills.sh/gosom/google-maps-scraper/google-maps-scraper

Demo: prompt pattern


Context: I run Parsley bakery in Yogyakarta.
Goal: Generate a weekly brief from the latest Google Maps reviews.
Constraints: 1 page max, bullet-driven, actionable, non-technical.
Output sections:
1) Summary (3 bullets)
2) Themes: praise & complaints
3) Notable quotes
4) Action checklist (this week / next month)

Demo: make it automation

  • Schedule: weekly
  • Same report format each run
  • Only review + decide (human stays in control)

What the report looks like

  • Summary: what changed since last week
  • Top complaints: what to fix
  • Top praise: what to preserve
  • Action list: do this week, plan next month

CTA

Try 1 automation this week.

  • Review brief (business)
  • Hiring pipeline snapshot (HR)
  • Meeting recap + action items (product)

Closing

  • Build AI skills as habits
  • Automate the repetitive
  • Keep the human decision-making

Q&A