X
About Us
DataBliss builds EdGrove, a modular school-management platform, alongside horizontal business apps and gated AI-agent tooling. Based in Warsaw, working with schools and organisations across Europe.
Contact Info
- Warsaw, Poland
- [email protected]
AI, data & automation
Copilots, data pipelines and AI agents that work behind proven checks.
AI, data and automation for education
We help schools and organisations put AI and automation to work safely: admissions copilots and briefings, data pipelines that turn spreadsheets into official records, and AI agents that build and run software behind proven checks.
Key features
AI copilots & briefings
Data conversion & pipelines
Workflow automation
Gated AI agents
Where we apply it
AI inside our products
Gateway
Admissions copilot, AI insight briefings, policy-based content drafting and narrative dashboards.
TaskFlow
AI agents connect to the project board through MCP, with consent per client.
Slate (coming soon)
An AI tutor for students is being built into the student portal.
Gridly (coming soon)
Smart scheduling suggestions and timetable optimisation.
Data and integration
- Turning MIS exports into official formats, as Dziennix does for the Polish e-register.
- Directory and device syncs from Google Workspace, Mosyle and Apple School Manager, as in Vault.
- Calendar feeds, imports and exports that keep systems in step without retyping.
AI agents with a gate on them
- Our gmind and Loopcraft tooling runs AI agents unattended only behind a proven check, a budget cap and a kill switch.
- Every run is recorded: what was asked, what was produced and how it was judged.
- People stay in control of merges, releases and anything that touches production.
We work across several leading providers and choose per task; the checks stay the same whichever model does the work.
We agree the data scope of every AI feature with you before it is switched on, and send no more than the task needs.
Yes. We apply the same gated approach to document handling, reporting and internal tools.
With one repeatable process. We define how a good result is checked, run it alongside your team, and only then let it run on its own.