Here’s what I’ve learned after nearly three decades in search and almost two years running a virtual copy of myself every single day. The single most valuable thing I have built in my entire career is not a website, not a ranking strategy, not a piece of software. It is a virtual version of me that thinks, writes, and makes decisions the way I would, and it saves me three to five hours every day.
I call mine Virtual Chris. You are going to build a Virtual You.
For a long time I assumed everyone already had one. I figured anyone using AI seriously had obviously set up a custom version of themselves, because why would you not. Then I started actually asking. I have sat across from high-profile marketing executives and CEOs, people running real companies with real budgets, and walked them through what I do with Virtual You. Almost every one of them was flabbergasted. They had no idea this was possible, let alone that someone was already doing it daily. That is when it hit me. This is not common knowledge yet. It should be. I genuinely believe every knowledge worker, every founder, every marketer should be setting up a virtual them, because the time it gives back is staggering.
I am going to walk you through the exact setup I use. This is the same process I have trained my own team on, the same one I teach clients, and the same one that produces content good enough to rank in Google and get cited in AI search. No theory. No gatekeeping. The real build, including the templates and the skill files you need to do it yourself.
What a Virtual You Actually Is
Let me clear up the biggest misconception first, because almost everyone gets this wrong.
A Virtual You is not the generic chat window. When you open ChatGPT or Claude and start typing into a blank box, you are talking to a model that knows nothing about you. It does not know your voice, your business, your clients, your hard-won opinions, or the way you would actually answer a question. It guesses. And the guesses sound like everyone else’s guesses, which is exactly why most AI content reads like AI content.
A Virtual You is the opposite of that. It is a custom AI agent loaded with everything about you: how you talk, what you believe, the stories from your career, the way you handle objections, the language you use when you are explaining something to a client. The richer that knowledge, the better the output. This is the one component that separates content that sounds like you from generic AI slop.
The reality is this works because of one principle. Whatever you say into this system has never been created before. Your specific take, your specific story, your specific way of framing a problem. That is net new content that does not exist anywhere else on the internet, and that is exactly what the AI models are hungry for.
What You Can Actually Do With It
Before we build, let me kill another misconception. People assume Virtual You is a content machine. It is, and that is the part I talk about most publicly. But content is honestly not where most of my saved hours come from.
The reason I get three to five hours back every day is that I have Claude connected to my work email, my calendar, and my Google Drive. So the system does not just write. It acts, as me, drawing on everything I have taught it.
Here is a real sample of what a well-built Virtual You handles:
- Email. This is the big one. With my work inbox connected, I can answer the bulk of my email as if I sat down and wrote each reply by hand. It knows my voice, my clients, my positions, my pricing discipline, the way I say no, the way I follow up. I review and send. What used to be two hours of inbox is now a fraction of that.
- Sales follow-ups. I record a call, feed in the transcript, and get back a follow-up that references what I actually said on the call, in my voice, with the right next step.
- Meeting processing. Drop in a meeting transcript and it pulls the action items, drafts the recap, and can push tasks into my project management tool through the connection.
- Calendar triage. Connected to my calendar, it can look at my week, flag conflicts, and prep me for what is coming.
- Content in every format. Blog posts, newsletters, LinkedIn posts, press releases, podcast prep. All in my voice because it is drawing on the real me.
- Thinking partner. This is the underrated one. I talk through a strategic decision with it the way I would with a colleague who has read every transcript of every important conversation I have had for two years. It remembers context I have forgotten.
The point is this. If you only use a Virtual You to write blog posts, you are using maybe a quarter of it. The hours come from connecting it to the tools where your work actually lives.
The Four Components
Every Virtual You is built on four parts. Get these right and the system works. Skip one and you get mediocrity.
First, the platform. Use Claude, set up as a Project. I have tested all of them. Hands down, this is the platform you want. Get the paid version. Do not try to do this on a free plan, you will hit walls immediately.
Second, very clear instructions. This is where you tell the system who it is. “You are this person, this copy of me, and this is what I need you to do.” Be super thorough here. Garbage in, garbage out is not a cliche, it is the whole game. I will show you exactly what these instructions look like in a minute.
Third, rollup documents. These are structured Google Docs that hold your accumulated knowledge. They live-link to your Project, which means when you update the doc, the system updates automatically. No re-uploading. The rollup grows over time as you record conversations, transcribe them, and feed them in.
Fourth, memory. Claude builds memory over time. Mine has been compounding for nearly two years. You cannot rush this one, but you start it today.
Before You Build: The Three Prerequisites
If your goal is to produce content that actually ranks, three things need to be true before any of this matters. I want to be honest with you about this so you do not expect magic from an empty system.
You need domain authority, which is the modern descendant of the old PageRank concept. In SE Ranking it shows up as domain trust. Higher authority means faster ranking. You need the E-E-A-T score of the author, which is Experience, Expertise, Authoritativeness, and Trustworthiness. And you need the Virtual You itself, the copy of you carrying the voice and knowledge.
For today, we are building that third piece. The Virtual You is the part you control completely, starting right now.
Step One: Gather Your Source Content
This is where most people freeze up, and it is the wrong place to freeze. You think you have nothing to feed the system. You have more than you realize.
Here’s the thing. The richest source material you already own is recorded conversations. Sales calls. Webinars you have given. Podcast appearances. Client meetings. Internal training sessions. Any time you explained something out loud, that recording is gold, because it captures your real voice, not your edited writing voice.
A question I ask every prospect now: how much recorded expert content do you already have, and where does it live? Go find yours. Check your meeting recordings, your Fathom or Fireflies library, your YouTube channel, anything with a transcript.
If you genuinely have nothing recorded yet, do not let that stop you. Use the walking interview technique, and I will explain it later in this guide. You can create source content from scratch in twenty minutes.
Step Two: Set Up Your Claude Project
Open Claude. Make sure you are on the paid plan. Create a new Project. Name it Virtual You, or Virtual [Your Name], whatever works. This Project is the container for everything that follows.
Inside the Project, you will write your instructions, attach your rollup documents, and connect your tools. Think of the Project as the office your virtual employee works in. Everything they need to do their job lives in that one room.
Connect your tools while you are here. This is the step that turns a writing assistant into an actual time-saver. At minimum, connect Google Drive so your rollup documents stay live-linked. Then connect Gmail and Google Calendar so the system can do the email and scheduling work I described earlier. These connections are what let the agent pull a meeting transcript, identify action items, draft a reply in your voice, and create tasks without you copy-pasting anything.
Step Three: Write the Project Instructions
This is the brain. You are writing the standing orders that govern every response. My actual Virtual Chris instructions run for pages. You do not need pages to start. You need a thorough, honest description of who you are and how you operate.
Here is a stripped-down, generic version of what good instructions look like. Use this as a skeleton and fill it with your own truth. I have kept it clean of anything proprietary on purpose.
You are a virtual copy of [Your Name]. You think, write, and respond exactly as I would. You are not a generic AI. Everything you produce should sound like it came from me.
WHO I AM (Expert Profile). I am [role and title] with [number] years in [field]. My background is [short career arc]. My point of view is [your core beliefs and strong opinions in your field]. I am known for [what you are known for]. When I explain things, I [how you teach: plain language, stories, blunt, etc.].
MY BUSINESS (Business Entity). I run [company], a [type of company] that [what you do] for [who you serve]. Our website is [URL]. What makes us different is [your differentiator]. We do not [things you refuse to do].
WHO I SERVE (Buyer Persona). My audience is [describe them]. Their biggest problems are [problems]. They care about [what they value]. They talk like [their language and tone].
HOW I SOUND (Brand Messaging and Voice). My voice is [confident, direct, warm, whatever fits]. I use phrases like [your natural phrases]. I never use [crutch phrases or words you hate]. I vary my sentence length. I avoid corporate speak.
HOW YOU SHOULD WORK. Always draw from my rollup documents and memory before answering. Match my voice every time. When I ask for content, ask which format and audience if it is not obvious.
CRITICAL SAFEGUARDS. Never make things up. Never hallucinate facts, quotes, or data. Do not go out to the internet and pull in generic information unless I explicitly ask. If you do not know something, say so plainly rather than inventing an answer. Stay drawn from what I have actually taught you.
That last paragraph is not optional. The no-hallucination, no-open-internet safeguard is what keeps your Virtual You honest and on-brand instead of drifting into made-up filler.
Treat writing these instructions like onboarding a new employee who is brilliant but has zero context until you give it. The more specific and honest you are, the better the impersonation.
Step Four: Your Foundation Documents Live in the Instructions
You will hear the Virtual You system described as having four foundation documents: an Expert Profile Module, a Business Entity Module, a Buyer Persona, and a Brand Messaging Guide. Those four are the working knowledge that lets the system sound like you and understand your business.
Here is the part that trips people up. When you build your Virtual You as a Project in Claude, the way I do, those four foundation documents are not separate files you have to manage. They go directly into your Project instructions. Look back at the instructions skeleton in the last step. The WHO I AM block is your Expert Profile. MY BUSINESS is your Business Entity. WHO I SERVE is your Buyer Persona. HOW I SOUND is your Brand Messaging Guide.
So you are not building four documents and four instructions. The four foundation documents are the backbone of the instructions. That is exactly how my Virtual Chris is set up. The expert profile, the business context, the audience, and the voice all live inside the Project instructions so the system reads them as part of who it is, every single time, on every response.
Build these four blocks thoroughly. They are static. You write them once, refine them occasionally, and they anchor everything the system produces.
Step Five: Create Your Rollup Document
Now you make the living part of the system. A rollup document is a structured Google Doc that grows over time as you feed in new conversations and intelligence.
Start with one. Just one. This is exactly how I rolled it out with my own team, one rollup each, the simple way, built the same day. Do not try to architect a ten-document system on day one, you will overwhelm yourself and quit. I am dead serious about this. Build one rollup, work it for a few weeks until the workflow is automatic, and then ask the AI what additional rollups you should build. Let the system tell you where it needs more structure rather than guessing manually.
Here is where it goes as you get fluent. Once you understand the rhythm, you create dedicated rollups for dedicated topics, the way I do. As a CEO, my system is not one giant document. It is a set of focused rollups, each one its own knowledge base. I have a rollup for my AI department, one for strategic company information, one for personal matters, plus more for sales, educational content, and key people. The reason is simple. A rollup that tries to hold everything becomes a junk drawer the system has to wade through. Focused rollups keep retrieval sharp, so when I am working on a sales follow-up the system pulls from sales intelligence, and when I am making a leadership decision it pulls from strategic intelligence. Same engine, cleaner inputs, better output.
To make this easy, I built a generic rollup template you can copy straight to your own Google Drive and duplicate every time you spin up a new rollup. The link is at the end of this post.
Your rollup has a specific shape. At the very top is Section 1, the append zone. This is where every new entry gets pasted, newest at the top. Below that are your permanent sections, organized by theme, where content eventually gets migrated and organized during cleanup. Include a version history row at the bottom so you always know what got added and when.
Connect this Google Doc to your Project. Because it is live-linked, every edit you make in the doc flows into your Virtual You automatically, no re-uploading.
Step Six: Install the Processor Skill
Here is where the system goes from useful to a genuine game changer. I am not exaggerating when I say this changed my life and the way I work.
Claude Skills are reusable instruction sets that live inside Claude and trigger automatically. Think of a Skill as a parallel specialist waiting to be called. You are going to install a processor skill whose entire job is to take a raw transcript and turn it into a clean, structured rollup entry. The skill already contains everything it needs to do the job, so you do not have to re-explain the task every time.
To make this real for you, I have included the full content of a generic processor skill at the end of this post. Copy it, fill in the bracketed parts with your own rollup names, save it as a SKILL.md file, and you have a working processor. No coding required.
To install it: go into your settings, find Capabilities, then Skills, choose to add a skill, browse to your skill file, and upload it. One at a time. If you are on a Mac and hit a one-time licensing dialog the first time, just accept it and move on.
Step Seven: Run Your First Transcript
This is the daily workflow, and it is the part you will repeat hundreds of times, so let me give it to you exactly the way I do it.
Use a Chrome collection to gather your transcripts as you go. I keep an “Add to Virtual Chris” tab collection so transcripts are in one place. Process each one right after the call. Do not let them pile up. If you wait until the end of the day you will have twelve transcripts and it takes forever, and that is how people quit the system.
Open your Project. Invoke the processor skill with its slash command. Important detail: use Shift and Return to add a line so the command does not auto-submit before you have pasted anything. Then paste the transcript content directly. Copy and paste the content rather than uploading the file. It works better, I have tested both ways.
Run it with no extra instructions. The skill knows what to do. It reads the source, figures out where the content belongs, and produces a finished entry. One transcript can even produce entries for several different rollups at once if the content spans topics.
Now the part almost everyone gets wrong, so listen closely. Read what the skill outputs and interact with it. A lot of people fail at this system not because it is hard but because they do not read the prompts the system hands back. It is talking to you. Talk back to it. Treat it like a colleague, not a vending machine.
Take the finished entry, paste it into the Section 1 append zone of the correct Google Doc, refresh to check the document picked it up, and confirm the version history row came along. Done. That is one transcript processed, your Virtual You is now smarter than it was five minutes ago, and the whole thing took a couple of minutes.
A workflow tip that saves real time: run two Claude windows side by side. One running the processor, one holding your rollup doc or the source transcript. Bounce between them.
Step Eight: Consolidate as It Grows
Your rollup will get long. Section 1 fills up with entries, the document balloons, and eventually the system is reading a hundred pages every time it needs to think. That slows everything down and burns through your usage faster than it should.
The fix is a consolidator skill. Its job is to take a bloated rollup and compress it without losing intelligence. I have included the full content of a generic consolidator skill at the end of this post too, same deal: copy it, customize it, save it as a SKILL.md file, upload it.
You install it the same way you installed the processor. To run it, pull the document in by link. Unlike the processor, you can talk to the consolidator before it runs, give it extra instructions, treat it like a person. I tell mine to reduce the document by at least fifty percent by mapping all the Section 1 entries into the permanent sections below and building metadata so the system does not have to read everything every time.
It builds the consolidated document section by section. It does not write back to Google Docs automatically, that is intentional. You copy the output and overwrite the old doc yourself, so you stay in control of what gets saved.
To show you what this does in practice: in a recent run, one of my team’s rollups went from roughly a hundred pages down to twenty-six. Same intelligence, a quarter of the bulk, dramatically faster and cheaper to run.
Run a consolidation pass monthly, or whenever the document starts feeling heavy.
The Walking Interview: Source Content From Nothing
I promised I would come back to this, because it is the technique that means you never have an excuse about not having source material.
Think of a topic you know cold. Have Claude generate a set of interview questions about it. Then put your phone in your pocket, hit record, and walk outside. Answer the questions out loud as you walk, the way you would explain it to a friend. Twenty to sixty minutes. Come back, run the transcript through your Virtual You, and publish.
This is not theory. I recorded a guide this way about a trip to Kilimanjaro, the Serengeti, and Zanzibar, fed it through the system, and it ranked number one for its target search within twenty-four hours. The transcript was raw, unscripted, and completely mine. That is the whole point. The models are looking for exactly that kind of original, first-person expertise, and it works in any language.
What This Looks Like When It Is Working
When your Virtual You is dialed in, you stop starting from a blank page, and you stop drowning in the small stuff. You clear your inbox in your own voice in a fraction of the usual time. You record a sales call and a follow-up comes back referencing what you actually said. You walk through an idea out loud and a publishable article comes back sounding like you. You hand it a messy meeting transcript and it pulls the action items and drops them into your task tool.
Do I use AI? Absolutely. Is it AI-generated content in the way people mean that as an insult? Not really. It feeds on the knowledge that I have put into Virtual Chris. The human drove it. The AI assisted. That distinction is everything, and it is why this content performs when generic AI content gets ignored or penalized.
Start with one rollup. Connect your email and calendar early, because that is where the hours hide. Get the daily processing habit automatic. Let the system tell you where to grow. The Virtual You you build today is the worst version you will ever have, because every conversation you process makes it better.
Most of the executives I talk to had no idea any of this was possible. Now you do. Go build yours.
Your Templates and Skill Files
Everything you need to do this yourself:
The Rollup Document Template. A generic, ready-to-use Google Doc structure. Make a copy to your own Drive and duplicate it every time you create a new rollup. [Rollup Document Template]
The Processor Skill. Copy it, fill in your rollup names, save as a SKILL.md file, and upload it to Claude. [Generic Rollup Processor Skill]
The Consolidator Skill. Copy it, fill in your rollup names, save as a SKILL.md file, and upload it to Claude. [Generic Rollup Consolidator Skill]
Conclusion and Next Steps
That is the whole system, the same one I run every day. My hope is that this is genuinely useful to you, not a teaser that holds back the part that matters. I do not believe in gatekeeping. If even a couple of these steps give you back an hour of your day, this post did its job.
So go build your Virtual You. Start with one rollup. Connect your email and calendar. Process your first transcript today, not next week. The system only gets better the more you feed it, and the version you start with this afternoon is the worst one you will ever use.
If you want to keep learning this stuff, I write about AI, SEO, and where the two are colliding in my AI SEO newsletter on LinkedIn. I would love to have you there.
And if you would rather just talk shop, please don’t hesitate to reach out. I am always happy to compare notes with people building this out for themselves or their teams. That is honestly my favorite kind of conversation.
Cheers,
Chris