← Content Writing & SEO VA Lesson 3 of 6 · ~5 min

Writing WITH AI honestly: tells, facts, editing passes

The workflow nobody admits to

Almost every working content VA drafts with AI now. The difference between professionals and hacks isn't whether they use it — it's that professionals are honest about it, edit hard, and take responsibility for every fact that ships. Clients aren't paying you to type; they're paying you to guarantee the output. This lesson is that guarantee, as a workflow.

Key idea: AI drafts, you deliver. The moment your name is on the doc, every sentence in it is yours — including the ones the machine hallucinated.

The draft-with-AI workflow

Feed the briefPaste the full brief from lesson 2 — keyword, intent, audience, outline, must-cover terms. Garbage prompt, garbage draft.

Outline firstHave the model propose the outline; fix it before drafting. Reordering sections now costs seconds, not an hour later.

Draft section by sectionOne H2 at a time, not the whole article — quality collapses at length, and you stay in control of flow.

Expertise passAdd what only a human source has: first-hand details, real examples, correct nuance, a practitioner quote. This is the E-E-A-T layer.

Fact passVerify every number, name, date, and claim. Separate pass, full attention — details in the next section.

Voice passStrip the AI tells (below), match the client's voice, read it aloud once.

The AI-tell checklist

These are the fingerprints editors are paid to find and remove:

Do (reads human)

Varied sentence rhythm — some short. Some that run longer because the idea needs it
Specific nouns and numbers: "a 3-bed in Frisco listed at $410k," not "a property"
First-hand texture: what it looked like, what went wrong, what surprised you
One clear claim per paragraph, defended
Contractions and natural transitions

Don't (reads machine)

"In today's digital landscape," "it's important to note," "furthermore"
Uniform paragraph lengths and rule-of-three everything: "fast, reliable, and efficient"
Hedging every claim: "can potentially help improve"
Section summaries that restate the heading
Confident claims with zero sources attached

Your lesson 1 audit doc already started this list. Keep growing it — a personal AI-tell checklist is a portfolio artifact, and running it is literally the job description of an AI content editor.

Fact verification: the discipline that keeps clients

Language models generate plausible text, not true text. They will invent statistics, misattribute quotes, cite studies that don't exist, and do it all in a confident tone. The playbook's failure-mode list says it plainly: one hallucinated fact destroys client trust and E-E-A-T. One. Google's quality systems reward experience, expertise, authoritativeness, and trust — and a fabricated statistic in a client's niche is a detectable, reputation-level failure.

The discipline: every number, date, name, and factual claim gets traced to a primary source — the actual study, the official docs, the company's own page. "A 2024 Stanford study found..." that you can't locate in ninety seconds gets cut or replaced, no exceptions. If a claim is load-bearing and unverifiable, it doesn't ship.

Reality check: the fact pass is a separate pass. Editors who "check facts while polishing voice" miss things — your brain can't hold skepticism and style at the same time. Two reads, two jobs.

What this workflow makes you

Run this loop well and you are, functionally, an AI content editor — the one lane in this market where headcount is actually growing. Clients producing 40 AI-assisted articles a month don't need another writer; they need the person who makes those articles accurate, on-brand, and safe to publish. That person edits three to five articles in the time a from-scratch writer produces one, which is exactly why the role pays $6-9/hr mid-band (~₱34k-44k/mo) and climbs from there — anchor here if the strategist tier feels far off.

Do this now

Twenty-five minutes. Run one full rescue:

Generate an 800-word AI draft on a topic you actually know.
Mark every AI tell using the checklist — count them.
Highlight every factual claim, then verify three of them to a primary source. Note what you found (expect at least one surprise).
Rewrite the worst 200 words to publishable, human quality.

Save before and after versions side by side as "AI edit sample — [topic]." A raw-versus-final comparison is one of the three portfolio artifacts this course builds, and it's the single most persuasive one for editor roles.

Tip: use your ← → arrow keys.