Frequently asked questions

VAL is the AI Agent working alongside your security officers. VAL answers their questions from your own post orders, documents patrols and incidents while they happen, and turns a shift into reporting your supervisors and clients can act on.

• Vigilis AI — The platform.
• VAL — The intelligence layer running on it.

No. VAL makes the officer on post more capable.

The person closest to the risk usually has the least support. Every dollar spent upstream on cameras, access control and analytics narrows down to one person, 90 seconds, and what they see, decide and record. VAL is there for that moment.

What you do with the hours VAL gives back is your decision.

• Enterprise security teamsWho want the field feeding their operating picture in real time instead of arriving as a PDF the next morning, although we send PDFs as well.
• Guard companiesWho want their officers documenting better and their clients seeing it.
• Leadership, clients and other partiesWho need proof the work happened, and can see it in VAL directly.

It does guard tours. That’s the smallest thing it does.

Guard tour software exists to prove an officer walked a route. Fine, and VAL does that minutes after launch on day one: patrol tracking, incident reports and photo analysis. Table stakes.

The industry has spent 30 years making the perimeter smarter and very little making the person standing inside it smarter. Your officer is the most important sensor on site and the last one to get real technology. VAL changes that. They talk to VAL in their own language, get an answer grounded in your approved post orders for that exact site, and finish the shift with the record already written.

What this means — your documentation improves itself, gaps surface the same week instead of at the next audit, and the person closest to the risk finally has the support everyone upstream already has.

No. VAL runs in a browser. An officer scans a code or opens a link, adds VAL to the home screen, and it behaves like any other app. It works on iPhone and Android. There is no update for officers to install: when we ship a fix, everyone has it that day.

Most reach working competency in about 15 minutes. They speak or type in plain language, and spelling and grammar don't matter. If an officer doesn't know how to do something, they can ask VAL. Where voice doesn't suit the environment, text does everything voice does.

Yes. Officers speak, type and get answers in their own language, and the report reaches the client in the language the client reads. The interface labels are in English today.

Patrols, checkpoints, photos and incident capture keep working. VAL stores everything on the device and syncs the moment coverage returns, with nothing for the officer to do. Live conversation needs a connection, and VAL says plainly when it is offline.

No. An officer photographs the checkpoint and the photo carries the location and the time. If your sites already use tags, VAL reads them. We won't ask you to put hardware on every door.

Every checkpoint carries a location, a time and a photo, and all three land on the patrol report. If someone files a patrol from the wrong end of the county, the report shows that too. Basements and dense multi-story buildings degrade location accuracy, so we walk your sites with you before we scope anything.

Your documents. VAL answers from the approved post orders and procedures for that specific site, and sites stay isolated from one another. When VAL doesn't have an answer, it says so, sends the officer to a supervisor, and logs the gap so someone can close it.

Not on its own. VAL drafts the change and sends it for review. A supervisor edits, approves and merges it, and their name stays on the record. Nothing reaches the live document without a person putting it there.

You do, and VAL makes that easier than it is today. Every question an officer asks that your documents can't answer becomes a logged gap with a drafted fix waiting for a supervisor. Most teams find gaps in their first month they didn't know they had.

VAL stores every change to every record with who made it, when, and what it said before. Narratives are never overwritten. Patrols and photos carry their location and time. When your counsel needs the file, you take a snapshot, which locks the document set as it stood and exports it.

Usually, and there are three ways in. Some systems we have already built for. Anything that can send an email, a text or a file can reach VAL today, which covers most alerting systems without anyone writing code. Everything else goes through our documented API. We don't charge for integration work.

VAL doesn't watch live video. Your camera system sends its alert and the image, and VAL reads the image, decides whether it needs attention, files it and routes it to the right person. Any system that can send an alert by email works today, and we publish setup guides for the common ones.

No. Scheduling, payroll and time and attendance stay where they are. VAL reads your staff, sites, clients and schedules once a day and never writes anything back.

VAL is clean and auditable from your first day. The years before that stay with your current vendor, and getting them out is a contract question more than a technical one. Two things are worth doing before you give notice: pull the full export while you are still a customer in good standing, and put data-extract language in any renewal you sign between now and then.

You do. Your data stays isolated to your organization, we never use it to train a model for anyone else, and you can export it whenever you want. If you stop working with us, you take it with you.

Yes. You can export your data at any time, and when you leave — or on request — we remove it. A few records may be held briefly where the law or an active investigation requires, and our privacy policy sets out those windows.

Your data stays in the United States, encrypted in transit and at rest. Nothing leaves the country. VAL runs in the cloud today, and there is no on-premise deployment.

Every request is scoped to your organization and access follows the role. Our privacy policy sets out what we collect, who processes it and how long we keep it. SOC 2 Type II is in progress, and we are glad to put our CTO in front of your security team to walk through the controls we operate today.

Access is scoped by role, and every action is audited — reads as well as writes, by your staff and by the limited Vigilis AI staff who operate the platform. Each entry carries who, when, why, and exactly what was read or changed, so there is a full trail for any access to your data by anyone. We don’t browse your data, and we never sell or share it.

The models behind VAL are run by established U.S. providers under enterprise terms. They process a request only to answer it; those terms prohibit them from training on your data or reusing it, so nothing you or your officers say improves a model for us, them, or anyone else. Our privacy policy names each provider, what it handles, and how long anything is retained, and we’ll put the data-processing terms in front of your security team.

We monitor our own systems and every supply-chain provider we depend on, continuously, for vulnerabilities and alerts, with a full array of security and AI scanning across our logs, our product, and every data change. If something needs attention, we see it early. Reach our security team any time at security@vigilisai.com.

We monitor continuously and audit every action, so an incident can be traced to exactly what happened, when, and to what. If one affected your data, we would notify you promptly with what we know and what we’re doing, and work it through with you. Reach us at security@vigilisai.com. The specifics live in our agreement.

A site is generally running inside a week and officers are competent in about 15 minutes. Getting your documentation in order takes longer, and it is where most of the value shows up.

Pricing scales with the deployment: sites, users and the use cases you turn on. We are early and we structure agreements to grow with a customer. Tell us what you are running and we will put a number in front of you.

We will walk a real post with you, using your sites, your officers and your use case. It takes 25 minutes and we don't bring a deck.