On-premise product
KnowledgeAI Enterprise: the answers are in your documents.
A chat that queries manuals, contracts and procedures with RAG technology and always cites the source. Installed on your servers, with permissions per role and full traceability.
Answer · sample
- Question: warranty on spare parts?
- Source: General terms 2026, art. 7
- Visible to the sales team only
Use cases
Where KnowledgeAI saves time.
Company knowledge becomes accessible with a question in natural language.
Internal knowledge base
Procedures, policies and best practices queried without searching through documents. Useful for onboarding new hires.
Technical documentation
Technicians and designers find specifications, manuals and drawings without leafing through hundreds of PDFs.
Contracts and legal
Clause search, contract comparison, deadlines and renewals with the source always cited.
Human resources
Answers on leave, benefits and expense procedures without going through the HR office.
Analysis and reports
Management queries reports, market analyses and historical data to decide quickly.
Internal customer-service support
Operators immediately find guides, case history and escalation procedures.
How RAG technology works
It searches your documents, then generates the answer.
RAG (Retrieval Augmented Generation) combines semantic search in your documents with text generation: every answer comes from real sources.
01
Indexing
PDF, Word, email and databases are processed and indexed in a local vector store.
02
Semantic search
For each question the system finds the most relevant passages, even without the exact words.
03
Generation with context
The model answers using only the information found in your documents.
04
Answer with citations
Every answer lists the source documents, verifiable with one click.
Comparison
KnowledgeAI versus a generic assistant.
| KnowledgeAI Enterprise | Generic AI assistant | |
|---|---|---|
| Source of answers | Only your documents, with citation | General knowledge, no sources |
| Where data goes | Stays on your servers | Sent to external services |
| Access control | Per role, team and document | None |
| Traceability | Full log of questions and documents consulted | None |
Features
Built for company use.
Permissions per role
Every user sees only authorised documents. Integration with Active Directory, LDAP and SAML.
All formats
PDF, Word, Excel, PowerPoint, email, HTML and Markdown, with text recognition on scans and access to SQL databases.
Connectors
SharePoint, Google Drive, Confluence, Notion and Dropbox, with automatic indexing of new documents.
Full traceability
Log of every question, answer and document consulted, exportable for GDPR audits.
Multiple channels
Web app, Slack, Microsoft Teams, embeddable website widget and REST API.
Usage analytics
Most frequent questions, documentation gaps and usage trends.
Pricing
Three formats based on users and archive size.
Prices exclude VAT. No cost per question.
Team
up to 50 users€30,000
For a team or department.
- 10 GB of indexed documents
- PDF, Word and Excel
- Web application
- Email support
- Team training (4 hours)
Business
up to 200 users€55,000
For companies needing access control.
- 100 GB of documents
- Permissions per role
- All formats and text recognition
- SharePoint and Drive, Slack and Teams connectors
- Team training (8 hours)
Enterprise
unlimited usersfrom €85,000
For organisations with advanced requirements.
- Unlimited archive
- Single sign-on and Active Directory
- Custom models
- Guaranteed service level and dedicated support
- Pre-configured server included
Combining KnowledgeAI with VoiceAI On-Premise gives a 15% discount on the total.
Explore VoiceAISpecifications
What the installation requires.
Hardware requirements
- NVIDIA RTX 4090 or A100 GPU, 24 GB VRAM (48 GB for larger models)
- 128 GB RAM (256 GB recommended)
- 2 TB NVMe SSD storage
- 32+ core CPU for indexing
- 1 Gbps local network
- Requirements grow with document volume
Supported integrations
- Repositories: SharePoint, Google Drive, Confluence
- Chat: Slack, Microsoft Teams
- Authentication: Active Directory, LDAP, SAML, OAuth
- Databases: PostgreSQL, MySQL, SQL Server
- ERP: Zucchetti, TeamSystem, SAP
- REST API, webhooks, GraphQL
Frequently asked questions
How much does KnowledgeAI Enterprise cost?
From €30,000 (Team, up to 50 users) to €55,000 (Business, up to 200 users with permissions per role) up to over €85,000 for the unlimited Enterprise version. Prices exclude VAT, no cost per question.
What does “answers with cited sources” mean?
Every answer lists the documents it was drawn from, with a link to the original passage. If a piece of information is not in your documents, the system says so instead of making it up.
Do the documents stay in the company?
Yes. Indexing, search and answer generation take place on the server installed on site or in a private cloud you control.
How long does activation take?
On average 4–8 weeks: installation, connection to repositories, initial indexing, permission set-up and team training.
Let’s talk about your business.
45 minutes with a PugliAI consultant: process review, priority opportunities and a written report with an ROI estimate. Free and without obligation.