AI-assisted content, human quality control: our disclosed workflow

We use AI. Here's exactly how.

Most agencies either hide their AI use or brag about it. We do neither — we disclose it, precisely, because you deserve to know how content published under your name gets made.

Our workflow, in full:

1. Weekly 30-minute founder interview. You talk; we listen. Your stories, your opinions, your client wins, your answers to buyer questions. This is the raw material, and it's 100% human — it's you.
2. AI-assisted drafting. Our team uses AI tools to accelerate the first draft: structuring the interview transcript into posts, generating variations of hooks, adapting one idea across formats. AI handles speed and structure.
3. Mandatory human quality control. Every draft is reviewed by a human editor before you ever see it. The editor checks facts, kills generic phrasing, restores your voice, verifies claims, and rejects anything that doesn't meet the bar. This step is not optional, not sampled, not "spot-checked." Every piece, every time.
4. Client approval. You review every piece before it publishes. Nothing goes live without your explicit sign-off. If a draft doesn't sound like you, you reject it and we rewrite.
5. Scheduled publishing, engagement, and reporting. Approved content publishes on schedule; we manage engagement and DMs; you get the monthly pipeline report.

AI is the assistant. Humans are the authors. You are the final authority. That sentence is the whole philosophy — and it's stated publicly on our standards, not buried in a contract.

For how the interview becomes a month of content, see our pillar guide to content repurposing and one interview, thirty pieces.

Why pure-AI content fails

The temptation is obvious: AI can produce a month of posts in an afternoon, for nearly nothing. Some agencies sell exactly that. Here's what their clients get:

Confident hallucinations. AI fabricates statistics, misattributes quotes, and invents case details with total confidence. A coach who publishes a made-up stat under their name isn't just wrong — they're exposed. One sharp reader who checks the source ends the credibility permanently. Our human QC exists largely to catch what AI invents.

Generic advice. AI writes the average of everything it's read, which means it writes what every coach already posts: "5 tips for better leadership," "consistency is key." Generic content gets scrolled past by buyers who've seen it a hundred times. Authority comes from specificity — your stories, your contrarian takes, your client outcomes. AI can't generate those; only your interview can.

Voice drift. Publish pure-AI content for three months and your voice converges toward the generic AI voice — the same cadence, the same structures, the same hollow enthusiasm as ten thousand other accounts. Buyers can't articulate it, but they feel it: this doesn't sound like a person. In a market where trust is the product, sounding like everyone is the same as sounding like no one.

Unverified claims. AI will happily write "our clients see 10x results" or invent a methodology you don't have. Under your name, that's your liability — including FTC liability if the claims touch testimonials or earnings. Our FTC disclosure guide covers why this matters legally. Human review is a compliance function, not just a quality function.

No judgment about what not to publish. AI doesn't know which client story is too sensitive to share, which opinion would alienate your best referral partner, or which topic is beneath your positioning. Humans know. The approval gate — your approval — is where judgment lives.

The pattern: AI is excellent at producing words and terrible at deciding which words should exist under your name. That's a human job, and it always will be.

The disclosure ethics: why we tell you this

There's an active debate about whether AI assistance must be disclosed. Our position is simple: when content is published under your name to build trust with your buyers, the process behind it is material information. You should know. Your buyers, arguably, should be able to know too — which is why we publish this workflow publicly instead of hiding it in an onboarding doc.

The ethics break down into three commitments:

Disclose to the client. You know AI assists our drafts before you pay us a dollar. It's on our standards page, in this article, and in our sales conversations. No surprises.

Never misrepresent authorship. The ideas, stories, and opinions are yours — from your interview. AI structures and accelerates; it doesn't invent your point of view. The content is yours in every way that matters, and your approval makes it officially so.

Hold the human line. Disclosure without standards is just confession. The ethical core isn't that we told you about the AI — it's that no AI output reaches your audience without human verification and your approval. The workflow is the ethics.

We'd hold any agency to the same standard. If you're evaluating us against competitors, ask them the three questions from the FAQ above: Do they disclose AI use openly? Is human review mandatory before publishing? Do you approve everything? Their answers — or their evasions — tell you everything about how seriously they take your reputation.

What human QC actually checks

"Human quality control" sounds vague, so here's the actual checklist our editors run on every draft:

• Voice match. Does this sound like the client? We compare against the interview transcript and prior approved content. Anything that reads generic gets rewritten.
• Fact verification. Every statistic, quote, date, and factual claim gets checked. If it can't be verified, it's cut or rewritten without the claim.
• Claim review. No unverified outcome claims, no invented client results, no earnings promises. Anything touching testimonials or results gets checked against our FTC standards.
• Specificity test. Would this post work with any coach's name on it? If yes, it's too generic — back to the interview for the specific story or detail that makes it yours.
• Sensitivity screen. Client stories anonymized appropriately, no confidential details, nothing that could embarrass a client or damage a relationship.
• CTA and accuracy. Links work, calls to action are correct, names are spelled right, formatting is clean.

A draft that fails any check goes back for revision. A draft that can't be fixed gets killed. This is the unglamorous work that pure-AI workflows skip — and it's the work your authority depends on.

Where AI helps most — and where it's banned

Not every task in the workflow gets AI assistance. Here's our internal line, because "AI-assisted" is meaningless without boundaries:

AI assists: structuring interview transcripts into post drafts, generating hook and headline variations for human selection, adapting an approved idea across formats (post to carousel to newsletter excerpt), summarizing long transcripts for the editor's review, and checking grammar and readability. These are speed and structure tasks — the draft gets better faster.

AI never decides: which stories to publish, what claims to make, whether a draft sounds like you, whether a client detail is too sensitive to share, or whether a piece is ready. Those are judgment calls, and judgment stays human. AI also never publishes, schedules, or sends anything — every output waits for human review and your approval.

AI never invents: no fabricated statistics, no invented client outcomes, no synthetic testimonials, no made-up quotes. If the interview didn't contain it, the draft doesn't contain it. The editor's fact-verification check exists to enforce this, but the rule comes first: AI works from your raw material, never from its own imagination.

This boundary is what separates disclosed, responsible AI use from the "AI content agency" model where software writes and publishes unattended. The first is a productivity tool inside a human system. The second is a liability machine with a login.

The bottom line

AI is a powerful drafting assistant and a dangerous publisher. The agencies worth hiring know the difference and build their workflow around it: disclosed AI assistance, mandatory human quality control, and client approval as the final gate. That's our machine, stated plainly, held publicly.

If you want the full system — weekly interview, AI-assisted drafts, human-edited and client-approved content, engagement and DM management, monthly pipeline reporting — the Authority Engine tier ($1,500/mo, $900 setup) is built for it, with DM responses within 2 business hours and comment responses within 4 business hours.There is no call guarantee. Social Signal removed its earlier call guarantee because it was unproven; instead you get the call-tracking dashboard — every call the system produces, reported in numbers, so you can judge it yourself. Start at the pilot.

Frequently asked questions

Do you use AI to write content?

Yes, openly. AI assists our drafting — it helps structure ideas, generate variations, and speed up first drafts. But every draft then passes mandatory human quality control, and nothing publishes without the client's approval. AI is the assistant, never the author. We disclose this on our standards page and in every client relationship.

Why not just let AI write and publish directly?

Because AI makes confident mistakes: fabricated statistics, generic advice, voice drift, and claims no human verified. For a coach, whose business runs on trust and authority, one published hallucination can undo months of credibility. The cost savings of skipping human review are tiny compared to the cost of publishing something wrong under your name.

Will my content still sound like me?

That's exactly what the human layer is for. Our editors work from your weekly interview — your actual words, stories, and phrasing — and shape AI-assisted drafts back into your voice. Then you approve every piece before it publishes. If a draft doesn't sound like you, it doesn't ship. Your approval is the final gate, always.

Is AI-generated content bad for SEO or platform reach?

Platforms and search engines increasingly reward original, helpful, human-judged content and discount generic filler — regardless of how it was produced. The question isn't whether AI touched the draft; it's whether a human expert verified, improved, and approved it. Our workflow is built around that distinction.

How do I evaluate whether an agency's AI use is responsible?

Ask three questions: Do they disclose AI use openly? Is there mandatory human review before publishing? Do you approve everything before it goes live? If the answer to any is no — or if they dodge the question — walk away. Responsible AI use is disclosed, supervised, and client-controlled. Anything else is a liability wearing a productivity costume.

Keep reading

Pillar guideContent Repurposing System PlaybookOne Interview, Thirty Content Pieces GuideOur Standards: The FTC Honesty Pledge GuideFAQ: Pricing, Terms, Timelines