How does uploading documents work for the assistant?

Upload regulations, procedures, handbooks, internal memos — anything. The engine parses, indexes, and the assistant starts answering from them within minutes.

How do you prevent hallucinations?

The assistant only answers from retrieved passages in your documents, with a similarity threshold. Below that threshold, it tells the user the answer isn't in the knowledge base — and we flag the question for your review.

Can I buy just the assistant, without courses or assessments?

Yes. The assistant is available standalone.

If we have our own materials, do we have to buy the Kampster library?

No. Disable the Kampster library for your team and run only on your own content.

Where is our data stored?

EU data residency is available, with a Data Processing Agreement on request. Your content is tenant-isolated and never used to train any model.

Which models does Kampster use?

We use models from several leading providers — picking the right one per use case. We can confirm the specific configuration during onboarding.

What happens if the assistant can't find an answer in our docs?

It says so explicitly — no guessing. We flag the question for your team to review, so you know where the knowledge base has gaps.

For Individuals For Business For Universities For Government

For Government

Your agency's rulebook, answering itself.

AI assistant that answers from your own regulations and procedures, with citations.

Company Assistant

Acme Inc · your tenant

Grounded

Can I paste customer data into our AI tools?

Yes for sanitised data. No for raw PII — names, addresses, or payment info must be redacted first.

Source AI Use Policy v3.2 · Section 3.2

Zero data retention

stays in your tenant

EU AI Act is now in effect — is your team ready?

Step 1

Upload your regulations.

Statutes, procedures, handbooks, internal memos — drop them in. The engine indexes everything in minutes.

Knowledge base Indexing

Drop your docs here

PDF · DOCX · MD · TXT

AI Use Policy v3.2PDF · 18 pagesIndexed
Vendor OnboardingDOCX · 6 pagesIndexed
Brand & Voice GuidePDF · 24 pagesIndexing…
Engineering AI PolicyMD · 9 pagesQueued

Ready in minutes

Step 2

Ask anything.

Your staff ask in plain language. Every answer is grounded in your documents, with citations pointing back to the source.

Ask anythingCited

Who approves a new AI vendor?

Procurement first, then Legal for the DPA, then InfoSec for the security review.

Source Vendor Onboarding · 2026

Always cites the source

Try a question your staff would ask.

Sample answers from a sample agency. The real one runs on your regulations, in your tenant, with citations.

Sample Agency · Sandbox

Demo

Pick a question below. The answer comes back with citations to the source document.

Pick a question

Demo sandbox — not connected to live data.

An assistant that knows your agency

Your regulations. Their answers.

No outside knowledge. No leaked records. No invented rule. The assistant answers only from what's in your tenant — your regulations, procedures, handbooks, and training content.

Grounded in your regulations — nothing else.

Every answer is retrieved from your own content. If it isn't in your knowledge base, the assistant says so instead of guessing.

Cites the source.

Every answer links back to the exact document and section. Staff can verify before they act.

Speaks your team's language.

Ask in any language, get the answer in the same. Regulations in English, answers in Romanian — fine.

Adapts to context.

Tone, examples, and local references shift to fit each region your agency serves.

Tenant-isolated.

Your content stays in your tenant. Never used to train any model. EU data residency available.

More than translation

Cultural adaptation

Personalization engine that adapts to each region — examples, scenarios, and tone.

Live in every market you operate in.
Tell us where you operate; your team learns in their own language.
Same standard, local context.
Examples adapt. The bar doesn't. A pass in Vienna means the same as a pass in Bucharest.
Adapted, not translated.
Role context, examples, and tone reshape per locale — not just the words.
Updated everywhere, instantly.
Update a course once; every locale updates with it. No translation tickets, no version drift.

Wherever your team works

Web, iPhone, Android.

Same lessons, same data, every device. Native apps for mobile, web for desk work.

Web

iOS

Android

3.2 hours

the average knowledge worker spends searching for internal information every week.

Slite · Enterprise Search Survey 2025

20–30%

of mid-sized companies' operational expenditure lost annually to rework and fragmented knowledge.

McKinsey · Bain · PwC · Gartner (2024–2025)

27%

faster internal-info retrieval among assistant users vs. unassisted baseline.

Microsoft · Work Trend Index 2025

Security & trust

Your data is safe.

Enterprise-grade controls, EU data residency, and zero data retention — built in by default.

GDPR-aligned
EU data residency available. Data Processing Agreement on request.
SOC 2 controls
AES-256 at rest. TLS 1.3 in transit.
ISO 27001
Annual third-party penetration testing.
Privacy by design
Tenant-isolated. Your content stays yours.
Zero Data Retention
Nothing stored beyond the session. Never used to train anyone's model.

Trusted by leading universities

Tested with students at world-class universities.

Real-world validation with participants from top global universities.

Johannes Kepler University, Austria

Focusing on the technical UX of adaptive learning; testing how students interact with AI-driven interfaces to ensure seamless algorithmic integration.

London School of Economics, United Kingdom

Focusing on behavioral usability; analyzing how interface design affects student decision-making, engagement, and long-term learning outcomes.

The Chinese University of Hong Kong

Focus on usability and learning effectiveness of the platform, including analysis of user behavior, AI interaction quality, and overall learning experience, with emphasis on personalization and content accuracy.