Private AI over your internal documents. Evidence first.
Product 1 is a private AI platform that answers questions and drafts documents from your own sources, with citations that lead back to them. It's designed for deployments you control, with inference that can run on a model inside your environment.
You must give 90 days' written notice before the end of any contract year 1. The 2025 amendment adds that notice is only valid if it is sent to the supplier's legal department by registered letter 2.
No early-termination fee applies if notice is given on time 1.
Either party may terminate this Agreement by giving ninety (90) days' written notice prior to the end of any Contract Year, without penalty.
Notices shall be sent by registered letter to the Supplier's Legal Department.
Illustrative screen with fictitious documents.
Companies want AI. Their most valuable information can't reach it.
- It's confidential.
Contracts, regulations, specifications and code can't simply be sent to a public AI service.
- A generic assistant isn't enough.
It answers from general knowledge when the answer has to come from your own sources.
- Nobody remembers why.
Systems like SAP record what changed. The reason is scattered across tickets, people and emails.
Evidence first. Then the answer.
- 1
Ask
Someone in your company asks a question in plain language.
A question, in their own words - 2
Retrieve the evidence
Product 1 searches the relevant passages inside the project that person is allowed to use.
Only the projects they have access to - 3
Answer with citations
The model writes from that evidence and shows where each statement comes from.
Every statement linked to its source
Your documents don't retrain the model. They only become searchable inside their own project.
Citations don't make AI infallible. They let your team check each answer against its source before acting on it.
One platform. Several ways to use it.
Answers over internal knowledge
Policies, procedures, contracts and specifications, with every answer pointing to its source.
Technical knowledge
Questions about code, requirements and technical documentation, for development and support teams.
Documents from evidence
Structured documents drafted from your own material, so your team starts from a draft with sources instead of a blank page.
Domain applications
Plugins add the knowledge of a specific domain on top of the common platform, so each new use case doesn't start from zero.
Private is a set of requirements. We define them with you.
Product 1 is designed for customer-controlled AI deployments. Inference can run on a model runtime inside your environment, and the indexed data stays in your deployment.
- No public AI service
- No internet access
- Encryption at rest
- Data residency
We ask which of these you need and validate them with your team.
On infrastructure you control.
Authentication, roles, permissions per project and an audit trail of key actions.
For stricter requirements.
Networking, encryption-at-rest and egress controls are defined and validated for your deployment.
A common foundation. Domain applications on top.
Product 1 handles the parts every AI use case needs. Domain applications add what is specific to each field.
- Projects and permissions
- Search and citations
- Answer and document generation
- Deployment
Spin and Spin+ are developed with Eurolabs, our partner company in Luxembourg.
Enterprise experience. Not an AI start-up guessing.
Product 1 is developed by ITBASS, an SAP technical boutique based in Andorra. Our team has spent more than 18 years in complex SAP landscapes, from ABAP and S/4HANA to integration. That's where we learned how much valuable knowledge never reaches the people who need it.
Start with one workflow. Prove it on your data.
- One real domain of information
- One or two data sources
- A small group of users
- Representative questions from your team
- Success criteria agreed before we start
- Quality of answers and of the citations behind them
- Time saved on real tasks
- Whether your privacy and deployment requirements are met
Questions we hear first.
What is Product 1?
A private AI platform that answers questions and drafts documents from your internal information, with citations. It's built for deployments you control and extended through plugins for specific domains.
Is it just another chatbot?
No. A chat is one way to use it. The core is retrieving evidence from your sources, generating from it and showing where each answer comes from, inside projects with their own permissions.
Does our data leave our infrastructure?
Product 1 is designed for customer-controlled deployments: inference can run on a model inside your environment and the indexed data stays in your deployment. Exact networking, encryption-at-rest and no-egress requirements depend on the deployment posture you choose, so we validate them with you explicitly.
Do our documents train the AI model?
No. Indexing makes documents searchable inside their project. It doesn't retrain the model.
Can it connect to our systems?
The platform is connector-based, so many systems are natural candidates. We scope the specific system, its API and its permissions before committing to an integration.
Can the AI take actions in our systems?
Only through controlled tools you authorise. We treat execution as an authorised capability, not as open access for an AI model.
How do we start?
Tell us about the information you want to use with AI and what stops you today. If there's a fit, we agree a focused pilot with success criteria up front.
Have information your AI can't use yet?
Tell us what you want to ask of your internal information, and the constraints that stop you today. A member of our technical team reviews every request.