Security, compliance & ownership

AI helps. You decide what it can do.

For small teams and larger organizations: EU data storage, appropriate access and control over what people and AI do.

Discuss your requirements
CompanistOSSearch your workspaceLI

Settings · Departments

Access that fits your role

Department: Sales

LisaManages the department
Owner
SamWorks on assigned tasks
Member

Action for review

Change
Publish a website version
Proposed by
Lead Developer
Status
Awaiting approval

This change is submitted to you first.

EU storage and deliberate processing.

Your organization’s data remains stored within the EU. Processing is aligned with the services and models your organization chooses.

  • EU-only processing with selected EU services or local models.
  • Define data flows, storage locations and retention per application.
  • Choose which sources an agent can access and what can be shared.

Roles, permissions and departments.

People and AI work within the same boundaries. Access matches the role, department and task.

  • Set up departments, teams and responsibilities.
  • Give employees and AI agents only the access they need.
  • Restrict sensitive customer information, conversations and files to authorized users.

Monitoring and logging.

See what your agents do, what changes and where attention is needed.

  • Track tasks, progress, errors and exceptions.
  • Keep actions and approvals traceable with logging.
  • Match alerts and checks to each process’s risks.

Semi-automatic, with your approval.

Let AI prepare work and carry out permitted steps. You decide when a person needs to review first.

  • Keep transcripts, summaries and tasks reviewable.
  • Have agents prepare emails, website changes or other actions as proposals.
  • Require human approval for sensitive, high-impact or irreversible actions.

Learn together, with respect for privacy.

Self-learning systems improve with feedback and knowledge within agreed boundaries. Learning together starts with choosing what to make available.

  • Decide which feedback and approved sources your agents may use.
  • Share experiences and ways of working in the community without sharing confidential conversations or company information.
  • Define whether data may be used for model training and what agreements apply to the selected service.

Model-independent. Local too.

Companist connects your work with a suitable AI model. You are not tied to one provider.

  • Choose models based on quality, cost and data policy.
  • Use local models when your data or policy requires it.
  • Match capabilities and performance to the selected model and setup.

Your environment, your data.

Your organization retains ownership and control of its data and materials.

  • Agree on data access, exports and portability.
  • Define ownership and transfer of code, websites and brand files.
  • Decide how data is retained, deleted and transferred when moving away.

A foundation you can build on.

Your work needs to continue when a service changes or a supplier becomes unavailable. Model choice and an appropriate setup help reduce that dependence.

  • Match backups, recovery tests and recovery times to how critical your processes are.
  • Define export formats, access and handover of data and services.
  • Plan for disruptions and for moving to another model or service.

Compliance starts with clear agreements.

Technology and organizational policy need to align. We define the requirements for your deployment up front.

  • Document processing agreements, suppliers and data flows.
  • Match retention and access policies to the purpose of processing.
  • Discuss required infrastructure documentation and audit information for the deployment.

Additional control for business databases · Pilot preview

Existing applications.
New possibilities with AI.

With Sentinel by NewHeap, all applications running on a supported SQL Server or PostgreSQL database can become AI-ready for use within your organization. No new or additional development of those applications is needed. Once configured, AI agents can use permitted data with query controls and protected results.

Explore Sentinel and the pilot options →
  • Define access per application, user and data column.
  • Mask, pseudonymize or block data according to policy.
  • Require a separate approval for changes where policy requires it.
  • Record decisions without SQL text, parameter values or result data in the audit log.

An additional pilot option, not included by default. Not yet released for production. We review deployment, supported database features and costs with your IT team up front.

Before you decide

Your IT and privacy team’s questions. Discussed up front.

We review the requirements for your use case together. You do not need to work out data, model and continuity arrangements after setup.

  • Processing agreements and suppliers per service
  • Storage and processing locations, including any transfers
  • Model training, retention and deletion
  • Roles, audit logs and human approval
  • Backups, recovery, support and service agreements
  • Data exports and handover when leaving
Discuss your security questions →

The agreements behind the technology.

The specific setup, models, data flows and retention are agreed per organization, so the solution matches your security and compliance requirements.

Privacy informationTerms & agreementsCompanist Guard →

A practical first step

What would you hand over first?

Tell us where time or clarity gets lost. We will show you how Companist can make that work simpler.

Book 30 minutesRequest a callback →
Your data stays yours

You retain ownership, access and control.

Data stored in Europe

Processing varies by service and selected model.

ISO 27001-compliant infrastructure

Security is part of the foundation.

Model independent

Choose suitable models, with Local AI where needed.