You've been promised software before. We know.
Half the companies we meet have a tool they paid for and stopped using — while the knowledge they were trying to protect kept walking out the door anyway. So we don't ask you to take our word for anything. Here's exactly how we work, what the AI does, what it never does, and how we make sure this one actually lands.
AI does the reading. You keep the judgment.
There is a lot of noise about AI right now, and most of it over-claims. So here's the plain version — the things we say, and the things we will never say to you.
If a vendor tells you their AI is never wrong, they are selling you something. We'd rather tell you what it actually does — and then show you, on your own documents.
Four things that stay true, always.
These aren't features we might add. They're the rules the platform is built on.
Your documents, and nothing else
Answers come only from what you've uploaded. Not the public internet, not someone else's data, not a guess.
Every answer shows its source
The document and the section, right there. If you don't trust an answer, check it in one click.
You decide who sees what
Access is governed by role, department and sensitivity. The board pack doesn't surface for the shop floor.
Your experts own their own words
Nothing captured in someone's name is published until they have reviewed and approved it. AI drafts, a human approves — that order never reverses.
Most projects don't fail on features. They fail on follow-through.
A licence key and a login is not an implementation. Here's what you get from us that a download link can't give you.
A named owner
One person from our side is responsible for your rollout. Not a ticket queue. A name and a phone number.
A date, agreed up front
A go-live date and what "done" means, written down before we start — so nobody is guessing three months in.
We're here
Built in Nepal by a team you can meet. When something breaks, you're not filing a ticket into a distant time zone.
Your HR and L&D team already has a full-time job.
Every tool promises to save them time and then hands them a second one. So we do the setup, not you.
What we do
- Help you load your documents into your own private space, under NDA
- Set up your access rules with you
- Run the capture sessions with your senior people
- Draft the first courses from your own material
- Train your team, and stay close after go-live
What you do
- Share the documents you already have — under NDA, on your terms
- Give us a few hours with the people who know most
- Tell us who should see what
- Review and approve before anything publishes
- Keep the judgment. It's your company.
Your documents stay yours.
You're about to put your policies, your procedures, and the hard-won judgment of your best people into someone else's platform. You should be asking hard questions. Here are our answers, before you have to ask.
Private to your organization
Your content lives in a space only your organization can see. It is never pooled with another company's, and never shared.
Never used to train public AI models
Your policies and internal documents are not sent off to train someone else's model. They are used to answer your people's questions, and nothing else.
You grant access. You revoke it.
Every document carries access rules you set — by role, department and sensitivity. The board pack never surfaces for the shop floor. Change your mind, and access changes instantly.
NDA before anything moves
We sign a non-disclosure agreement before we see a single document — including during a pilot. Confidentiality isn't a clause we add later; it's where we start.
If you want your security team in the room before we start, good. Bring them. We'd rather answer their questions early than lose your trust later.
The questions we always get asked.
Our most experienced people are close to retirement. Can you actually help?
Yes — and this is the moment to start, not the week before they go. Structured sessions draw out what they know that was never written down, they approve every word in their own name, and it becomes something their successor can simply ask a question of. Start it while they're still in the building.
Will our senior people actually want to hand over what they know?
It's the first thing people worry about, and the answer is usually yes — if it's framed right. What we capture is published in their name, with their approval, as their contribution to the company. Most experts want their work to outlast them; what they resist is being copied without credit or asked to write a manual in their spare time. We do the drawing-out; they review and approve.
Isn't this just a chatbot on top of our files?
The asking part looks similar — that's the easy half, and plenty of tools do it. The difference is what's in there to ask. Most of what runs your company was never written down, so a tool pointed at your existing folders can only return what already exists. We capture the judgment first, then make it askable. There's no shortcut to a corpus that doesn't exist yet.
We tried a system like this before and nobody used it. Why is this different?
Because adoption is a delivery problem, not a software problem. We load your documents ourselves, set it up with your team, and prove it works in one department before you commit to anything wider. If people aren't using it in the pilot, you haven't bought it.
Can the AI make things up about our company?
It answers from the documents you upload, and it shows you the source of every answer. If the answer isn't in your material, it tells you rather than inventing one. That's the whole design.
Where do our documents actually live, and who can see them?
In a space private to your organization — never pooled with anyone else's. Access is governed by the rules you set, by role and sensitivity, and you can change or revoke them at any time. Your documents are never used to train public AI models. We work under NDA from the first conversation.
Does it work in Nepali?
Yes — properly, not as an afterthought. Your people can ask questions and take training in Nepali or English, which matters when your head office and your floor staff don't work in the same language.
Do we have to buy the whole platform at once?
No. Most companies start with the part that hurts most — usually getting their policies and documents answerable — and grow into the rest as it proves itself.
How long until we see something real?
A pilot runs about 30 days end to end. You'll see your own documents answering questions well before that.
Bring one policy. Watch it answer.
The fastest way to know whether this is real is to see it working on a document you wrote.