Vigil
The architecture behind Vigil, built for enterprise
That work is underway, and a continuation patent covering it is pending.
What carries over
Vigil scores a message in three stages. Sentinel reads the message on the phone and scores it there. Scout asks about the things only the person holding the phone would know, and the score is reconsidered in light of the answer. Sage explains what was found and recommends what to do about it.
That architecture is protected by U.S. Patent No. 12,500,909, our system for bi-directional message scoring using feature extraction, contextual refinement, and synthesis. It is granted, and it covers Vigil as it ships today.
Not another filter
A filter decides for the employee. Vigil is built for the employee to interact with, which is what leaves them better at spotting the next one.
Every organisation already has filters, and the messages that matter are the ones that got through them. The interesting problem is not classifying mail more aggressively. It is what happens in the seconds after somebody opens a message that made it to their inbox, and whether they end that moment knowing more than they did.
What changes for enterprise
Two halves. Employees get the app they would have had as consumers, deployed through the device management an organisation already runs, scanning corporate mail rather than personal mail. Administrators get the half that does not exist on the consumer side: what arrived, what it scored, and what people did about it, across everyone.
Sage is where the difference concentrates. On a phone it recommends. In an organisation the same recommendation can also open a ticket, notify a channel, or enrol someone in the training they were about to need. The advice does not change; what can be done with it does.
The scoring stays where it is. Sentinel runs on the device, and an integration that scores mail before delivery is an addition to that rather than a replacement for it.
None of this is built yet, and this page will say so until it is.
Ask first, then explain
The version of this we want to build asks the employee what they think of a message before it tells them, then shows them what Sentinel found and where the two differed.
An organisation that does this gets something a phishing simulation cannot give it: not who clicked a fake, but how well people read the real mail they were already receiving.
Where it stands
- VigilShipping. Granted patent.
- EnterpriseIn development. Continuation patent pending.
If you are looking at this for a team, we would rather hear from you while it is being built than after. What you need now is what shapes it. enterprise@vigilcode.com