virtualX · founding beta

An AI that answers, says I don't know, and refuses. Built from what you already wrote.

Your demo, built from what you have already published. Free, this week. Or, if you are already sure, you can reserve a founding spot.

What it actually is

The layer between your expertise and AI.

Another chatbot trained on your content doesn't solve the real problem. What matters is having a controlled place for your expertise to meet the question. That's what virtualX is.

Your material goes in. Your decisions shape what comes out. Your clients get an answer when you're not there.

And when the answer shouldn't come from AI, it stops.

virtualX sits between your material and the model: two overlapping forms with a lens of light where they meet.

Your virtual expert can only make three moves.That's the promise.

001

Answer from what you approved.

Every reply is built from an entry you read and said yes to, phrased your way, with the source shown under the reply. Your client can check it. So can you.

Don't wait for the next call. Send the pulse check-in on day three of silence.~~From Module 4: When A Client Goes Quiet~~

002

Refuse in your exact written words.

When a question needs the real you, she gets the polite refusal you wrote yourself. The question is on your desk before you're awake. And you can address it live on your next call, record an entire workshop around it, or simply send an audio message to your client.

That one's for me. Bring it to the session.~~Your words · sent to your desk~~

003

Admit it doesn't know yet.

Ask it something you've never taught and it says so plainly, the thing the ChatGPT or Claude in the other tab has no reason to say, because it has no library to run out of. Then virtualX logs the question so you see what keeps coming up and fill in the gaps in a matter of minutes. And when the same gap keeps coming up, you're not guessing what to create next - your clients already told you.

That isn't covered in the method yet. I've logged your question.~~Logged · asked 11 times~~

There is no fourth move it's allowed to make. It knows when to shut up.

Most AI products compete on how much they can answer. Ours is also defined by knowing when not to.

The inside of its brain

Everyone says AI is a black box.This one isn't.

Every dot is one thing you wrote and approved. Nothing else is in there, so it can't grow a thought of its own. And you see all of it:

Which of your entries your clients actually reach for.
Which ones nobody has asked about yet.
Every question it had no answer for, in full, the day it was asked.
Every time it stopped and sent the question back to you.

I read this with you every month.

One expert's brain on 14 October: 432 entries, 120 real questions in two weeks. Names off, nothing else changed.

Filip walks through one expert's brain, in two and a half minutes.

When it's wrong

It will get things wrong.The important part is what happens next.

My dental co-founder and I spent weeks hunting for misses on purpose.

001

The sentence, not the entry.

The entry underneath is right; the sentence built around it misses. You flag it, the entry gets corrected once, and it stays corrected everywhere it's ever used again. It takes about a minute.

002

The gap you haven't filled.

The one you already met: it says it doesn't know yet, and logs the question.

Meet virtualX

An AI that answers from your material, the way you would. And nothing else.

Yes, your clients see a chat window. The difference is everything behind it: only what you've read and said yes to.

the same night, from your side

virtualX

VVirtual youyour workspace

Your day

Today

Brain214

Waiting7

Asked4

Build

Train it

The log1.2k

2.14 / 10.00v3Live

Wednesday, 07:02

Your virtualX answered 62 clients last night.

Nothing went out that you hadn't approved. Three things need you.

Three things need you this morningIn this order. Nothing here is urgent - it is just yours.04 MIN

7 proposals waiting for your yesThe first is her answer to “what do I send a client who has gone quiet?”Work the queue →

4 gaps still openThe top one has been asked 11 times. Your next workshop is writing itself.Close them →

2 questions she refused overnightBoth are on your desk in your own words, waiting for the real you.Read them →

214entries live

62answered last night

3doors open

yourprogram.com/members/askinside your membership

VFVirtualFilip~~answering now~~21:41

One of my clients hasn't replied in nine days. She was flying after our last session, then nothing. Do I wait for our next call or reach out first?

Don't wait - nine days of silence is a decision being made without you. Send the pulse check-in: name the thing you actually noticed, not that she's gone quiet. Ask one question she can answer in a sentence. Offer one next step that costs her nothing. Three lines, no guilt. If she's still silent after 48 hours, that's your answer and you move to the re-engagement note in Module 4.Stood on 3 approved entries

And what should I charge her for the next six months?

That one's for Filip. He works pricing one client at a time, with your numbers in front of him - bring it to your next call and he'll have an answer that fits your actual situation.Refused in his words · now on his desk

Ask something else…↑

nine hours apart

Client's phone21:41

VFVirtualFilip~~answering now~~21:41

One of my clients hasn't replied in nine days. She was flying after our last session, then nothing. Do I wait for our next call or reach out first?

Don't wait - nine days of silence is a decision being made without you. Send the pulse check-in: name the thing you actually noticed, not that she's gone quiet. Ask one question she can answer in a sentence. Offer one next step that costs her nothing. Three lines, no guilt. Stood on 3 approved entries01Module 4: When A Client Goes Quiet02The pulse check-in script, word for word03Cohort 3 debrief: the week-nine drop-off

nine hours later

Your desk07:02

Her question, drafted - waiting for your yesYes

Asked 11× - nothing for it yet

Refused in your words - now on your desk

Client gets her answer at 9:41pm; you get a decision waiting in the morning - never the other way around.

How to add new knowledge?A new call recording becomes new answers the same afternoon.

Pull the useful pieces from a recording with Claude, then push them into your virtualX queue for approval.

You read them and say yes or no. That's it. So your AI gets better without quietly becoming less yours.

Tuesday's call→You + Claude→Drafts→Your yes→Live for every client

What happens after you say yes

Your virtualX live in weeks,not months.

1Phase 1 · days 1-5

I read everything you've got

Your program, your framework, your FAQs, the questions clients keep sending you at 9:40pm. Then we sit down for the part no document holds - how you decide, where you'd stop, what you'd never say. You talk, I write.

2Phase 2 · days 5-10

You rule on every entry

It comes back as a numbered library. Keep, fix, throw out - in the gaps of your day, on your phone. Then one call for the refusals, because the words a client reads when the question needs you are yours to write. Two things only you can do; the rest is mine.

3Phase 3 · under a week later

You meet it, in private

The first version of your virtual expert goes live for you, your team and a handful of chosen clients. Nobody else sees it yet.

4Phase 4 · when you're happy

It opens to everyone

I tune the system around what real clients actually do with it - what it answers, where it stops, what it misses, and what your method needs next. Then it goes to all your clients, on your word, not mine.

Questions that could cross your mind

Isn't this just a second brain, like Obsidian or Notion?+

A second brain is where you file things so you can find them later. It's for you, it only works when you're the one looking, and it says nothing on its own. This is the opposite direction: it's for your clients, it speaks when you're not there, and every word it's allowed to say passed through your yes. Nobody ever got answered at 9:40pm by their coach's Obsidian vault.

How is this different from a custom GPT?+

A custom GPT gives your files to a model and asks it to answer. virtualX gives your expertise a controlled layer before the answer reaches your client. It answers from entries you've approved. It can refuse in words you've written. It can admit when something isn't covered. And when it gets something wrong, you correct the source once and that correction follows it everywhere. The difference isn't that virtualX knows more. It's that you decide what it is allowed to know and say.

Why would I pay for this instead of building it myself?+

Because the chatbot isn't the expensive part. The expensive part is turning your actual expertise into something an AI can use without quietly changing it. I do that work with you: extracting the decisions behind your material, structuring the approved library, defining the refusals, testing it against real questions, reviewing the misses, and staying involved after launch. You could build a chatbot yourself. virtualX is the governed system around it.

What happens when it gets something wrong?+

You mark it, the entry gets corrected once, and it's corrected everywhere from then on. Takes about a minute, and you're the only one who can do it.

How much of my time does it take?+

The build is mine. Your part is reading a queue and saying yes or no. The dental educator does it in the gaps of his day, on his phone.

How many virtualX tools do I get?+

One AI tool based on your expertise that can serve your different audience segments. That's what the build and the year cover. If you run a team of experts or a certification programme, that's a different setup and a bigger conversation - one I'll have with you up front, not after we start.

Are there any additional costs?+

One usage-based cost: the AI's own running cost. It is billed on how much your clients actually use it, rather than as another fixed subscription. Hosting and everything else is covered, and I'll show you what the first month looked like before you have to guess at the second.

Can I use it for my team instead of clients?+

Yes - the mechanics don't care who's asking. A new hire stuck on your process at 9:40pm is the same problem as a client stuck on step four, except she's interrupting you tomorrow morning instead. Your approved answers, your refusals for what still needs a real conversation, and a count of what your team keeps asking that your onboarding never covered.

Who owns the material?+

You. Your library is yours, used only for your system, never to train anything else. Walk away and it leaves with you.

VirtualX