Employee trust is part of the work.
Organizational data is sensitive. Septima protects it in two ways: your data never has to leave your servers, and your people are anonymous before any analysis begins.
Your data stays on your servers.
Septima can run the whole analysis inside your environment, on the AI model you choose. The leadership readout is the only thing that leaves.
Your environment
- Your dataMessages, meeting transcripts, documents
- Names removedBefore any analysis
- AnalysisA frontier or local AI model
- Leadership readoutThe only thing that leaves
Your infrastructure.
We can run data ingestion and analysis on your own servers, set up the way your systems work today.
Any AI model.
Frontier models give the strongest results. Open-source models running on your own hardware keep everything in house. Septima works with either.
Agreed sources.
Together we decide which messages, meeting transcripts and documents are in scope.
Your standards.
We set up each engagement to meet your security, privacy and regulatory requirements.
Your people stay anonymous.
Septima reads how groups work. It never evaluates individual employees.
Anonymization comes first: we remove names and personal details before any analysis begins. Findings describe teams and groups, and Septima declines questions about named individuals.
Questions your legal, security and people teams will ask.
Straight answers to the questions that come up before an engagement. Where a detail depends on your engagement, we agree it before work begins.
- What information does Septima analyze?
- Company-approved messages, meeting transcripts and documents. Together, we agree which sources are included before the diagnostic begins.
- Does Septima evaluate individual employees?
- No. Septima examines patterns at the group level, and findings describe the organization and its groups only.
- How is the information anonymized?
- Anonymization is the first step. We remove names and other personal identifiers before any analysis begins.
- Where is the information stored, and who can access it?
- Wherever your policies require. We can run ingestion and analysis entirely on your own servers, so your data never leaves your environment. Septima receives the finished readout, which we share with you.
- How long is it kept?
- Your source data follows your own retention rules: if you delete messages after 90 days, they are gone after 90 days. We don’t keep a store of your raw data. You and Septima both keep the readout, and a new question later means a fresh analysis on current data.
- Which AI models are involved?
- The one you choose. Frontier models give the strongest results; open-source models running on your own hardware mean nothing reaches an AI provider.
- What do we agree before work begins?
- The scope, the information to be included, where the analysis runs, the delivery timetable and commercial terms.
Before a diagnostic begins.
We work with your organization to agree the sources, decide where the analysis runs, and answer the questions your legal, security and people teams need answered.