Register: Establish ownership
Every AI agent gets a named human owner and a documented, versioned list of what it is permitted to do. Know which agents are operating, who is accountable for them and what boundaries have been approved.
Vigil from Datakosa
AI governance. Accountability. Control.
Know what your AI is allowed to do. Know who approved it. Prove what happened.
AI is moving faster than traditional governance can keep up. Employees are using AI tools. AI agents are taking actions. But can your organisation demonstrate who authorised those actions, what each agent was permitted to do, and whether it stayed within those permissions?
Vigil gives your organisation visibility, control and verifiable accountability over AI activity, running entirely on your own network.
Your AI Is Working. But Who Is Accountable?
AI adoption is accelerating across departments. From internal assistants to autonomous agents, organisations are increasingly relying on AI to perform tasks and make decisions.
But important questions remain:
- Who approved each AI agent and its activities?
- What exactly is each AI agent permitted to do?
- Can you identify when an agent operates outside its approved permissions?
- Who is notified when something goes wrong?
- Can you produce reliable evidence of AI activity for your board, auditor, insurer or client?
An AI policy document may explain what should happen. It cannot, by itself, prove what actually happened.
Vigil brings AI permissions, oversight and accountability into day-to-day operations.
Introducing Vigil
An accountability layer for your AI ecosystem.
Vigil is an AI governance platform that operates on your organisation's own network.
Every AI agent is registered with a named human owner and a recorded, versioned list of permitted activities. Vigil checks agent behaviour against those permissions in real time and creates a tamper-evident record of its checks and decisions.
When an agent steps outside its approved boundaries, Vigil identifies the exception and alerts the responsible person.
Your organisation retains control of its AI governance records and infrastructure.
Your AI. Your network. Your rules. Your evidence.
How Vigil Works
Four essential layers of AI accountability.
Check: Monitor against permissions
Vigil checks AI agent behaviour against its recorded permissions in real time. When an agent attempts to operate outside its approved scope, Vigil identifies the exception.
Record: Maintain verifiable evidence
Every check and decision is recorded in a tamper-evident, hash-chained audit record. The record stays on your organisation's own network, providing evidence that can be reviewed by authorised stakeholders.
Hold: Keep human authority in control
An authorised person can place a scoped, time-limited hold on AI use. The hold is recorded, ensuring that important governance decisions are visible and accountable.
Built for Organisations That Cannot Afford AI Blind Spots
Financial Services and Insurance
Strengthen oversight of AI activity in environments where risk, accountability and evidence matter.
Healthcare
Support governance over AI usage in organisations handling sensitive information and important operational workflows.
Legal and Professional Services
Maintain clearer ownership and accountability as employees adopt AI tools for research, analysis and everyday work.
IT Services and Technology Companies
Bring structure and accountability to growing AI ecosystems, internal agents and AI-enabled workflows.
Enterprise and Large Organisations
Establish consistent AI ownership, permissions and oversight across teams and business functions.
BPO and Business Process Services
Improve visibility into AI-enabled workflows and establish clear responsibility for agent activity.
Your Network. Your Data. Your Control.
AI governance without handing over your records.
Vigil is designed to run on your organisation's own network, not in a vendor-hosted SaaS environment.
- On-premises deployment: Vigil runs within your own infrastructure.
- Your records stay with you: Governance records are held on your network.
- No vendor-held data: Vigil does not hold your governance data in its own cloud.
- No standing vendor access: Your organisation controls access to its environment.
- Offline operation: Vigil is designed to continue operating without an internet connection.
Your organisation retains control over its infrastructure, records and access.
Policy Is Not Proof. Evidence Is.
An AI policy tells employees and teams what should happen. But when your board, insurer, auditor or customer asks whether AI activity was actually controlled, you need more than a policy document.
You need to demonstrate:
- Which AI agents were registered.
- Who was responsible for each agent.
- What each agent was permitted to do.
- Whether its activity stayed within those permissions.
- What exceptions were identified.
- Which governance decisions were recorded.
Vigil helps turn AI governance from a written policy into an operational, verifiable process.
Be Ready for the Questions That Matter.
Imagine your board asks: Who approved the AI agents operating in our organisation?
Your insurer asks: How do you know your AI tools are operating within approved boundaries? A customer asks: Can you demonstrate how AI is governed when processing our business information?
Would your organisation have a clear, verifiable answer?
With Vigil, you can establish named ownership, documented permissions, real-time checks and tamper-evident records, all within your own network.
See Vigil in Action
Start with a 5-day guided trial.
Every organisation has a different AI environment, different workflows and different governance requirements. That is why we offer a 5-day guided trial tailored to your organisation's requirements.
Our team will work with you to understand your environment, identify the relevant use case and guide you through Vigil's capabilities.
Experience how your organisation can:
- Register AI agents and assign accountable owners.
- Define and document permitted activities.
- Check agent behaviour against approved permissions.
- Identify activity outside defined boundaries.
- Maintain tamper-evident governance records.
- Exercise authorised, time-limited holds.
At the end of the guided trial, you can discuss a paid deployment based on your requirements.
No generic sales pitch. Let's explore how AI accountability can work in your environment.
Frequently Asked Questions
What is Vigil?
Vigil is an AI governance and accountability platform that registers AI agents, records their permitted activities, checks behaviour against those permissions and maintains tamper-evident records.
Is Vigil a cloud-based platform?
No. Vigil is designed to run on your organisation's own network rather than a vendor-hosted SaaS platform.
Does Vigil store our data?
Vigil's governance records remain on your organisation's own network. Almost Magic Tech Lab does not hold your governance data in its own cloud.
How does Vigil monitor AI agents?
Vigil checks AI agent behaviour against the agent's registered and approved permissions in real time.
What happens if an AI agent operates outside its approved permissions?
Vigil identifies the exception and alerts the responsible person.
Can an authorised person stop AI activity?
An authorised person can place a scoped, time-limited hold on AI use. The hold is recorded.
Who is Vigil designed for?
Vigil is designed for organisations using AI tools and agents that need clearer ownership, permission boundaries, oversight and verifiable accountability.
How can we evaluate Vigil?
You can request a 5-day guided trial tailored to your organisation's requirements. Following the trial, continued use moves to a paid engagement.
Request a Guided Trial
Tell us about your AI environment and we will arrange a 5-day guided trial.
AI adoption should not mean losing visibility or accountability. Give every AI agent a named owner. Define what it is allowed to do. Check its behaviour. Keep the evidence.
vigil@datakosa.ai
Vigil by Almost Magic Tech Lab. Marketed by Datakosa.
Request Your 5-Day Guided Trial
Take Control of Your AI Governance.
Make AI accountability part of your operating environment with Vigil.
Vigil from Datakosa. AI governance on your network.