Almost every call opens the same way. Somewhere in the first 10 minutes the person says a version of this: "I should tell you upfront, I don't really have a content library."
Then they talk for 40 minutes. They walk me through their framework. They tell me about the client who nearly quit in week 3 and what they said to keep them. They repeat the line they use on every sales call, the one that closes it. They mention the onboarding doc they rewrote 4 times because people kept misreading step 2.
That is the library. It was never missing. Nobody had ever collected it.
TLDR
- People disqualify themselves because they think a content library means a folder of finished blog posts.
- It does not. It means evidence of how you think, in your words, under real conditions.
- There are 3 Sources and you already have all 3: what you say, what you write, what you sell.
- 10 to 20 pieces across those 3 is enough to encode a voice. Not 200.
- The blocker was never volume. It was that nobody told you the recordings counted.
The Disqualification Happens Before Anyone Counts Anything
Here is what I think is actually going on. "Content library" sounds like an asset an organization has. It sounds like a shelf. If you are a coach running your practice out of a calendar and a Zoom account, you look for the shelf, you do not see one, and you conclude the thing is not for you.
So you go back to opening ChatGPT on a Sunday night and asking it to write a post about consistency. And it writes a post about consistency, and it sounds like every other post about consistency, because you gave it nothing of yours to work from. I have written before about why generic output is a context problem and not a model problem. This is the upstream version of that. You cannot supply context you do not believe you have.
The fix is not more discipline. It is a different definition. A library is not finished content. A library is proof of judgment. It is a record of the calls you make about what is good, what is wrong, and how you say things when nobody is editing you.
By that definition, you are wealthy and you have been for years.
Source 1: What You Say
Every call you have already recorded. Coaching sessions, group calls, discovery calls, podcast appearances, workshop replays, the voice notes you send clients at 11pm.
This is the strongest of the 3 Sources and the most neglected, and the reason is boring: talking does not feel like producing. You finish a call, the recording drops into a folder, and the folder becomes a place recordings go to die. One client I work with runs 16 to 20 coaching calls a week. Every one of them was sitting there unreachable until we built the layer that made them queryable.
What comes out of speech that never comes out of writing:
- Your actual rhythm. Sentence length, where you pause, what you repeat when you want something to land.
- The explanation you have refined 200 times by saying it out loud to confused people.
- Objections in a client's own words, which is the best headline research there is.
- The asides. The jokes. The part where you say "look, honestly" and then say the true thing.
You write like a professional. You talk like yourself. The system wants the second one.
Source 2: What You Write
Not your polished website copy. The working stuff.
Client emails. Slack and WhatsApp replies to a member who is stuck. The long comment you left under someone's post because they were wrong and it bothered you. Newsletter archives. Proposals. The Notion doc where you dumped a framework at 1am before you knew what to call it.
Written material carries something speech does not, which is structure. When you write, you order things. You decide what comes first. That ordering is a rule, and rules are what a Reference Library is built out of.
It also carries your anti-patterns, and those matter more than people expect. The words you would never use. The claim you refuse to make. The compliment you find embarrassing. A system that only knows what you like will drift. A system that knows what you reject holds a line.
Source 3: What You Sell
Your program. Your course modules. Your worksheets, your onboarding sequence, your sales page, your pricing rationale, the Q and A recordings from the last cohort.
This is your most refined thinking and it is the source people forget fastest, because once a launch closes the material stops being visible. It sits behind a login nobody revisits after week 6. Meanwhile the same person is scrambling for something to post on Tuesday.
What you sell tells the system 3 things nothing else can:
- The promise. What outcome you are actually accountable for.
- The sequence. The order you teach things in, which is a hard-won opinion disguised as a curriculum.
- The vocabulary. Your named steps, your framework labels, the words your clients repeat back to you.
Content built from what you sell converts differently, because it is upstream of the offer instead of stapled to it.
Why 3 Sources Beats 1 Big Archive
You could hand me 300 blog posts and I would build you a worse system than the one I would build from 15 well-chosen pieces across the 3 Sources. That sounds backwards. It is not.
A pile of one kind of material teaches the system one register. Feed it only finished posts and it learns your performing voice, which is the voice you were already unhappy with. The 3 Sources triangulate. Speech gives rhythm. Writing gives structure. Offers give stakes. Put them together and you get something that can tell the difference between a LinkedIn post and a client email, because it has seen you do both.
This is the whole logic behind the Encode step. We gather 10 to 20 pieces across the 3 Sources, produce a Voice Print, and build a Reference Library out of examples, anti-patterns, banned words, and structural rules. You end up holding a Rule Registry you can actually read. For most founders it is the first time their standard has been written down anywhere.
What "Enough" Actually Looks Like
Here is the specific answer, because vague answers are how people stay stuck.
- 4 to 6 recordings. Mixed. A sales call, a coaching call, a podcast if you have one.
- 4 to 6 written pieces. At least 2 of them unpolished. Real replies, not marketing.
- 3 to 5 pieces of what you sell. A module, a worksheet, the sales page, an onboarding email.
That is it. That is 10 to 20 pieces. If you have been in business 2 years you cleared this bar a long time ago and did not notice.
And if you genuinely have almost nothing recorded, the answer is not to go build an archive first. The answer is to record the next 3 calls you were already going to have. The material generates itself once you stop treating it as exhaust.
The Part That Actually Stops People
It is not volume and I do not think it is really doubt either.
I think it is that going through your own material is uncomfortable. You listen back and you hear yourself say "um" 40 times. You read an email from 8 months ago and wince. Nobody wants to spend an afternoon in a room with their own unedited output.
Which is fair. It is also the whole job, and it is the part I do not think most people will do, and that is precisely why the ones who do it look like they got lucky with AI.
The raw material is not the constraint. It never was. Coaches, consultants, and creators all arrive at this conversation carrying 3 full sources and a story about being empty-handed.
If you want to test that cheaply: send me 1 piece of your work. Something you said, wrote, or sell. Pick whichever is easiest to find. You get a finished asset back in your voice within 48 hours, and you find out in 2 days whether your library exists instead of wondering about it for another year.
FAQs about the 3 Sources
What are the 3 Sources?
What you say, what you write, and what you sell. Recordings of calls and talks, working written material like client emails and newsletters, and the contents of your paid offer including modules, worksheets, and sales pages. Together they give an AI content system your rhythm, your structure, and your stakes.
How much content do I need to build an AI content system?
10 to 20 pieces spread across the 3 Sources. Roughly 4 to 6 recordings, 4 to 6 written pieces, and 3 to 5 pieces from your paid offer. Volume past that point adds very little. Range across sources adds a lot.
What if all my calls are confidential?
Common with consultants and it is workable. Your own talk tracks, your explanations, and your framing survive redaction. Client specifics get stripped. What the Reference Library needs is how you think and how you speak, not who you said it to.
Does a messy archive disqualify me?
No. Mess is the normal starting condition. The structuring pass is the work, and it is the layer most people skip when they try to wire tools together themselves.