Trust
Your cloud, your data, your AI models.
OnTrack runs inside your own cloud and uses AI models your organisation approves. Every change records where it came from, which rule applied and who approved it.
Where it runs
Inside your own cloud.
Your cloud account
OnTrack, its agents and its access to AI models all run inside your own cloud. Your project data stays there.
Your AI models
Agents use the AI models your organisation approves, under your own policies.
Your region
Your data sits in the cloud region you choose.
Your channels
Web, Microsoft Teams, Slack and WhatsApp can each be switched on or off.
Who decides
Agents act only within your rules.
Approval levels you set
Your rules decide what agents can just log, what they can apply and tell you about, what needs approval and what gets escalated.
Decisions go where they do today
A launch delay goes to the project lead. Something that affects the programme goes to the programme director, and funding goes to the portfolio board.
News never changes a plan by itself
Company-wide news creates questions and suggestions that go through the same approval levels.
Pause and undo
Anyone who owns a rule can pause an agent or undo a change an agent made.
Autonomy is earned, not granted.
New agents only suggest, and people decide. OnTrack tracks how often each agent's suggestions are accepted. Only your PMO head can let an agent make a type of change on its own.
Approval levels in practiceFull history
Every change can be traced.
A complete record
Every change shows where it came from, which agent suggested it, which rule applied and who approved it.
Undo for every field
Whether a person or an agent made the change.
"Why?" on demand
Every agent suggestion explains its reasoning and how confident it is.
Disagreements go to the owner
If two people report different facts, the Gatekeeper keeps both, asks each for evidence and passes it to the item's owner.
Independent checks
The Auditor reports to the PMO head and never edits project data.
Agents only show what you're allowed to see
Ask about a project you can't see, and your question goes to its owner instead of being answered.
Questions security teams ask
Straight answers.
Does our project data leave our cloud?
Which AI models does it use?
Can an agent change our plan on its own?
Can we see what an agent did and why?
Bring your security questions.
We'll walk your IT, risk and security teams through the set-up, the approval levels and the change history.