Home / Academy / STEP-BY-STEP COURSES

GONDAL AI ACADEMY / PRACTICAL STEP-BY-STEP COURSES

Turn voice notes & ideas into a finished blog post, step by step

Writing is hard; talking is easy. If your best ideas arrive while walking, driving, or making tea, this course is your pipeline: record them, let the AI do the typing chores, and keep your voice — the thing readers actually follow — fully in charge.

What you need: a phone that records voice notes (every phone does), one idea you genuinely care about, and an AI chat. Most phones and AI apps transcribe recordings for free — your AI chat can also work from a pasted transcript.

The steps

1. Dump the idea out loud. Hit record and talk for three to five minutes as if explaining the idea to a friend. Do not script it, do not restart it. Rambling is fine — the structure comes later. Transcribe the recording (your phone's voice recorder, a free transcription app, or the AI app itself).
   Checkpoint

   You should now have: one messy transcript, in your real words, covering the whole idea from start to finish.

2. Clean the transcript — keep your voice. The AI's first job is janitorial, not creative. The moment it starts “improving” your wording, your voice — the entire point — starts leaking out:
   Copy-paste prompt

   Clean up this transcript: fix the punctuation, remove filler words and false starts, but keep my exact wording and my order of ideas. Do not rewrite it into formal English. Here it is: [paste transcript]

3. Find the shape. A ramble becomes a post when it has a headline, a point, and sections. Extract them from your own words:
   Copy-paste prompt

   From this cleaned transcript, give me: one headline under 12 words, a one-sentence summary of the post, and an outline of 3–5 sections — using my own phrases as the section titles.

4. Draft section by section. One section at a time keeps the AI close to your material. A single giant request produces a generic article that sounds like everyone and no one:
   Copy-paste prompt

   Draft section 1 from the transcript, about 150 words, in my voice. Use ONLY ideas from the transcript — add no new facts, examples, or claims. If something needs checking, mark it [CHECK].

   Checkpoint

   You should now have: a full draft assembled from your sections, with every [CHECK] flag still visible.

5. The honest edit. Read the draft aloud — the same test as the email course. Cut about a fifth of it. Verify every [CHECK] yourself against a trustworthy source. Then hunt the generic bits:
   Copy-paste prompt

   Read this draft and list anything that sounds like a generic AI article instead of a real person. For each spot, suggest tighter wording using my own phrasing from the transcript.

6. Finish the package. A post needs its wrapping: something for the readers in a hurry, and something for the people who share it:
   Copy-paste prompt

   From this finished post, write: (1) a two-sentence introduction I can use when sharing it on social media, and (2) a five-bullet summary for readers in a hurry. Both in my voice, both using only what's in the post.

Common mistakes (skip these)

• Publishing the raw transcript. Spoken rhythm is not reading rhythm — the cleanup and shaping steps are the course.
• Letting the AI add facts and examples you never said. It makes the post smoother and less true. Your material only.
• Skipping the read-aloud. Your ear catches what your eye forgives — especially the sentences that aren't yours.
• One giant “write my blog post” prompt. Generic in, generic out. Section by section is slower and ten times better.