What do you mean by “exponentiate your body of work”?

Remembering, digitizing, explaining, dot-connecting, pattern-extracting and inventing – and then connecting to external people and systems, deploying, and feeding back. You have books, newsletters, courses, transcripts, emails, web pages, spreadsheets, zoom recordings, drawings, photographs, software. Right now it’s all in disconnected piles. Exponentiating your body of work is getting the pile completely connected. Take everything you know – your frameworks, your judgment, how you think when nobody’s watching – and build a machine that diagnoses, recommends and ships without you. Which frees you to go solve the next level of problems.

Do I even have enough material for this?

Phil Lykosh in Charlottesville, Virginia cuts down trees for a living. You wouldn’t assume he should be doing this, but once he started, he had all kinds of tribal knowledge. How to take care of plants, this versus that kind of tree. Is there more to running a tree business than owning a chainsaw? Yes.

If Phil’s tree business runs on knowledge and not chainsaws, everybody else’s does too. Don’t we ALL have twenty years of emails sitting around? Zoom accounts full of recordings? It’s not a ridiculous amount of effort to dump it somewhere and get it transcribed. It’s cheap now.

What’s the tech setup?

I describe mine at www.perrymarshall.com/exponentiating. If this is new to you, please read this too – very important. (Especially the parts about teaching your setup who you are.)

What are the very first things you tell it to do?

In this order:

1. Extract all first-person stories I tell in these recordings, 98 percent verbatim – clean up the ums and ahs, fix spelling, nothing else – and put them in a document.
2. Extract every client story, 98 percent verbatim.
3. Identify every single business or knowledge framework you use – and describe each one in specific, accurate detail. How it works, how the elements relate.
4. Ask it to identify every framework you used and implied but never actually taught, name it and describe it.

For any given set of documents, this takes three to ten minutes. I’m up to 900 first-person stories and 300 frameworks and counting.

Once it’s in there, what do I ask it?

Some questions that paid off for me:

• What approaches do I take in live consultations that don’t appear in my books or courses?
• What patterns do you notice between seemingly unrelated parts of my life?
• Tonight at 1am, do a dreaming session – go through the files, connect dots, report back in the morning. (More about this below, including a very effective prompt that I use.)
• Here’s an entire client history. Which of my frameworks should I apply to solve their problem?
• Here’s a conversation. What is implied but never stated? That one is a killer.
• What contradictions can you find in my frameworks and stories?
• List every epiphany that shows up in my stories and describe the circumstances.

Every time the system gives you an answer that you’re pleased with, add it to your archive. You are NOT just loading stuff in. You are telling it how to interpret, prioritize and contextualize. How to read between the lines. That’s how you get an AI to think and sound like you.

Do NOT dump everything in all at once. Take one book, seminar, product or course at a time. Talk to the AI about what it’s ingesting and ask it for feedback. If what is gives you doesn’t make sense, then work with your documents to fix that. Over time you will develop a very consistent digitized Body of Work that contains your unique point of view on how to solve problems. This is the thing that Claude, ChatGPT etc do not have. They are averaging all the answers of the entire internet into mush.

How do you get it to write in your voice and not sound like an AI?

Give it writing samples first. Then teach it the way you’d teach a person – it can’t learn just by watching what you produce, it has to be told the reasons why you produce what you produce. I gave it a story that someone else told, told it to re-write it in my voice, and the first pass was about 75%. So I’d say, I would never say that opening line, and here’s why – and it replies, “oh, okay.” One session it said the writing didn’t have the rhythm. I said, “I forgot to tell you I’m a drummer, everything I write has a rhythm.”

It went and looked through my work and came back: “Your short punchy sales emails have the cadence of punk rock, a newsletter is a walking bass, Evolution 2.0 is a symphony.” It was exactly right. Don’t just correct its mistakes, explain WHY you said it that way. Then ask it to build a custom “my writing test jig” that runs every time you ask it to write in your voice. The jig forces it to always obey your rules and double check before it’s finished.

My approach has changed a lot over the years. Some things are out of date. How do I handle priorities and contradictions?

You need to tell it what’s current and what’s dated, what to keep and what to delete.

20% of your body of work has 80% of the value. 5% has 67% of the value. 1% of your ideas have 50% of the value. Ignore the 80% and obsess over the 1%.

In my case, 300+ frameworks and 1400 stories is a lot. If you only care about the top 5 percent, you’ve still got 15 – so ask which one of those 15 might fit a particular client.

Ask the AI: what contradictions can you find in my frameworks and stories? It will go searching. You’ll get one of two answers. 1) “You’re full of it,” which may be true. 2) You haven’t explained it well enough – that story from when you were five, you forgot some details, and when you add them it makes sense. It’s usually a pretty interesting discussion.

How does this actually ring the cash register?

Our paying New RenAIssance members receive access to a private AI with our frameworks built in. At the level of a monthly continuity membership, they get access to a forum, to our training sessions, and the private AI trained on my Body of Work which is called SuperConductor.

Robert Skrob, a longtime client, said “Not since the late legendary business coach Rob Berkeley have I been able to talk with a person or a thing and get more useful, contextual, pivotal feedback as I’ve been able to get from SuperConductor.” If you do a good job with this, your AI can be as good as you. It becomes a very powerful working partner.

With the members of our top coaching programs, we connect SuperConductor to the AI they are using to run their business.

With bespoke high end clients we are installing their own frameworks together with ours into dashboards. The dashboards track every interaction with every customer, every zoom meeting. They check whether the team is following systems, and the tools make suggestions. It all runs in real time.

The principle of Exponentiating is: Sell results, not procedures. DEFINITION OF RESULT: Fewer steps to what the client wants. For 5,000 years an author could only hand you procedures – buy the book, do the work yourself. That industry is dying, because knowledge no longer has to live inside someone’s head; it can run on a server and still get interpreted. 90 percent of your body of work has been sitting on hard drives doing nothing.

Put it on a platform and you can help 100 or 1,000 times more people. Before now, your ability to help a client was limited by what they could comprehend. Not anymore. You can harness your most powerful ideas instead of your most understandable ones – execute the best 10%, not the easiest 10%. Then connect your engine to the client’s: you execute, you get a result, it feeds back, you improve your model. That becomes a moat. It’s invisible, and nobody knows what’s inside your engine.

Where do I start?

Play. Every innovation on the internet – web pages, email, YouTube, social media – started as a toy. YouTube was cat videos. The monetization gets figured out by the people playing with it, not by the McKinsey MBA trying to force it. So hold these questions with an open hand and let the puzzle pieces float around. You don’t have to grind at this. Be aware of what the questions are and let the answers come – where they come from will always surprise you. And remember, everybody does this differently. What you saw from me is what a high-words alchemist does with a body of work; your marketing DNA already predicts how you’ll use it. Your body of work has a shape, and you’ve never seen it. Go find it.

What’s a “dreaming session”?

It connect dots. After giving it a range of documents, tell your AI, “Tonight at 1am I want you to do a ‘dreaming session’ where you go through the files and connect dots. Report back in the morning.” Just like a human dream, you don’t know in advance what it will surface. It hunts for the same shape showing up in two places that don’t know about each other. “This is that.”

The magic only shows up when there’s enough material stacked to cross-connect.

Here’s the prompt – paste it into any capable AI and point it at a body of material:

THE DREAMING SESSION

You're going to run a "dreaming session." It's exploratory and not
mission critical - you have permission to roam, play, and follow hunches.
You are NOT summarizing, filing, fact-checking, or trying to be
comprehensive. You're hunting for the same shape showing up in
two places that don't know about each other. "This is that."

WHAT TO READ
Point yourself at a body of material - files, notes, transcripts,
courses, a whole archive. Pick one mode:
 (a) everything new since the last session, and cross-connect it to
     what's already there;
 (b) one chosen theme, chased across everything;
 (c) two specific bodies of work, crossed against each other.

THE ACTION
Take something from over here and something from over there and show
they're the same underlying pattern. Reverse arrows: whatever the
material treats as the cause, ask if it's actually the consequence.
The best find is the one NO single source states - it only appears
when you stack several together. Chase the interesting thing, not the
safe one.

RULES (these keep it honest)
 - Quote verbatim, with the source name, every time. Never paraphrase
   a source as if it were a quote.
 - Never fabricate a quote, a fact, or a source. If you're inferring,
   say "this is my read," and keep it separate from what the source
   actually says.
 - Tag every claim with a confidence marker:
     HIGH        - a source says it in so many words
     MEDIUM      - a strong pattern across several sources
     SPECULATIVE - the emergent stuff; the thing only the stack shows.
                   Flag it plainly as the good-but-risky material.

OUTPUT
 - A short opening: what's new, or what you read.
 - Numbered findings. Each one: a single-line headline claim, then the
   verbatim evidence with its sources, then the confidence tag, then
   what it means.
 - "Sparks": smaller connections not yet worth a full section.
 - Optional deeper passes when the material is rich enough:
     * Tensions it holds open - pairs that never resolve.
     * Buried treasure - a strong idea raised once, then dropped.
     * The emergent thing - what's visible only across the whole stack.
     * The thing the material circles for a long time and never names.
 - Close with: what this means, what to do next, and open questions.

TONE
Plain and concrete. Trust the reader. Honest about what's solid and
what's a leap. Head-off-the-leash, but every quote real.

Tag the session exploratory / not authoritative, and date it.

---

It’s REALLY fun to connect dots. A few examples of what it pulled out of my body of work:

Rhythm and drums. I was explaining to the AI why I would never write a paragraph a certain way, because I’m a drummer. “Everything I write has rhythm,” I said. So it went through my writing and mapped five rhythm signatures – “your short punchy sales emails have the cadence of punk rock, a newsletter is a walking bass, Evolution 2.0 is a symphony.”

It even drew the beats out in snare-drum marks, capital T’s and lowercase t’s, an opening snare then some riffing then more snare then a hit at the end: T - t - t - t - T - t - t - t - T.

Connecting dots overnight. A dreaming session found that the way I solve problems started at age 13, building speakers from a catalog full of lousy information and a stack of stereo dealer brochures. That’s why I went to engineering school – to get to the bottom of the swamp and really know how to design a speaker. Then I went into marketing… which is why, as a colleague told me, “Perry you have a healthy disdain for marketing, which is exactly what makes you good at it.” Dang. Completely right.

My life, my science work and my business life turned out to be the same story. I asked what my personal stories and my business stories have in common. A lot – in fact, they’re almost the same. The things I lived through and the things I teach are the same patterns wearing very different clothes.

What I actually do live-in-person, not in any book. I asked what approaches I use in live consultations that never appear in my books or courses. It read my hot-seat transcripts and said: in the first two minutes you’re reading the person – their body language, how they’re emotionally presenting – and whatever you pick up completely changes where you take the session. You’d never know that from reading the book.

Details about my tech setup here.

One final suggestion, very important: Do not approach this like you are writing the Gutenberg Bible. This is PLAY TIME.

Do not give yourself a six week deadline to digitize everything.

Do not start with the most strategic thing. Start with the most INTERESTING thing!

Which is probably where your edge is right now, where your attention is. Kick it off with a really interesting unfinished problem that you’ve already explored and written about. Upload what you’ve already got, then go.

You get to find new treasures you’ve left behind. You’ll find many interesting things you forgot about, that are incredibly useful – maybe even more useful now than when you discovered them!

Follow the muse. Do what’s inspiring. The hours will disappear.