Our commitment to AI
AI where scale is needed.
People where judgement is needed.
Keeping our customers' infrastructure running means handling volumes that grow every year: events, requests, systems, documents, code. We chose to do it with artificial intelligence, under clear rules, so that decisions stay fast, effective and always owned by someone.
Commitments
Six commitments we make to our customers.
Being AI driven
Artificial intelligence is part of how we design, run and improve our work.
AI-supported processes wherever scale is needed
Where volumes grow, AI helps us keep decisions fast and effective without adding steps.
AI-supported MSP and cybersecurity services
AI enters managed and security services where it makes our engineers' work faster and more accurate, without changing contractual commitments.
People first
Validation and decisions stay with people, where they truly matter.
Agentic software development
Our software is built with AI agents, under the responsibility of our developers.
Innovation and technology with AI
We build our own technology, like the intacta.ai platform, with AI supporting every stage.
Principles
The rules that apply to every automation.
AI proposes, people decide
Irreversible actions, security and choices that affect the customer always go through a specialist.
Every decision has a name
Who approved, when and why is always traced.
Customer data stays the customer's
No customer data goes to an AI service without an explicit agreement.
Switched on step by step
First AI suggests, then prepares, then acts after confirmation. A step up only comes after measuring.
Every automation can be switched off
Each automated flow has an off switch and a manual fallback procedure.
We measure
Times, accepted proposals and corrections show where AI helps and where it needs rethinking.
People first
Four levels of autonomy, chosen case by case.
Not everything has to be automatic, and not everything has to go through a person. For each activity we choose the right level and change it only when the numbers justify it.
Suggests
AI proposes, the person acts. We always start here.
Prepares
AI prepares the full action; the person reviews and launches it.
Acts after confirmation
AI acts after an explicit yes, recorded with the name of who confirmed.
Acts and reports
Only for reversible, measured actions; the person gets the report and can undo.
Agentic software development
Agents write the code. Developers answer for it.
Our software, starting with the intacta.ai platform, is built with AI agents across the whole development cycle. Developers spend their time understanding the problem and verifying the solution.
Specification
The problem, the constraints, the expected result. Written before the code, and approved.
Mockup and code
Agents propose the interface, write the code and tests, document the choices.
Automated checks
Automated tests must pass: without a green result, no release.
Review and release
The developer reviews the change, tests it and approves the move to production.
What we delegate to agents
- Writing code from approved specifications
- Tests and documentation
- Interface mockups
- Repetitive checks before release
What we don't delegate
- Understanding the problem
- Architecture and security choices
- Approving every release
- Responsibility for the result
Want to know more?
We'll tell you how we work with AI and what it changes for the services we provide you.