VisaRoxy Book a demo

Book a demo

Show us one workflow. We will show you the agent.

The walkthrough is a working session, not a slide deck. Bring a workflow that costs your site time and we will configure the agent for it live — conditions, thresholds, schedule, escalation and all.

  • 45 minutes
  • Screen share
  • No slide deck
  • No obligation
Two operations professionals reviewing loading-dock camera footage together during a practical walkthrough.

Request a walkthrough

Tell us what you want to see.

The more specific the workflow, the more useful the session. If you have footage you can share during the call, that is the fastest route to a real answer.

Optional. An estimate is fine.
One or two sentences is enough. For example: vehicles cutting through the pedestrian walkway between the dock doors, or a lane that sits idle and nobody notices for an hour.

Every request is read by a person, not routed into a queue.

What happens on the call

Forty-five minutes, four parts

  1. 1

    Your workflow, in your words

    Ten minutes. You describe the problem as your team experiences it. We ask what a good outcome looks like and who currently owns it.

  2. 2

    Live agent configuration

    Twenty minutes. We build the agent on screen: zones, the condition, the persistence threshold, the schedule, the deduplication window, and the escalation path.

  3. 3

    Events against real footage

    Ten minutes. If you can share a clip, we run it through and you judge the events. If you cannot, we use representative footage and are explicit that it is representative.

  4. 4

    Deployment and next steps

    Five minutes. Where inference would run, what the site needs to provide, and what a pilot would cover. No pricing pressure on the call.

Who should be on the call

Two or three people is the useful size. The operational owner of the workflow is the most important attendee, because they are the person who will judge whether the events are right. Someone from plant IT or network engineering saves a follow-up call about how the runtime reaches the cameras. A security or compliance representative is welcome and usually has the questions that decide the deployment model.

Preparation

What to have ready. Nothing is mandatory.

A useful session needs a workflow and someone who knows it. Everything else makes the conversation faster, and none of it is required to book.

One workflow, described plainly
The single most useful thing you can bring. A sentence is enough: what happens, where on site, how often, and what it costs you when it happens.
A clip or two, if you can share one
Thirty seconds of the situation from a camera that sees it. This is the difference between a demonstration and an answer, because you get to judge events on your own footage.
Roughly how many cameras cover the area
Not a full asset list. A count and a rough sense of the camera positions is enough to say whether the view supports the rule you want.
Who currently gets told
The person or role who would receive an alert today, if one were raised. Escalation design usually matters more than threshold tuning.
Where events should end up
The alerting, ticketing or dashboard tool your team already uses. Bring the name and we will tell you whether it is a webhook, a short adapter, or a conversation with a vendor.
Any constraint on video leaving site
If your organisation has a rule about it, say so early. It determines the deployment model and therefore the shape of the whole pilot.

Pricing and pilots

How pricing works.

Three plans are published, so you can budget before you speak to anyone: see the plans and figures.

The plans are a fixed-scope pilot, a per-site plan for live alerting, and an estate plan for organisations running five sites or more. What you are invoiced is confirmed in a written quote after a scoping session, because three things move the figure: how many streams you want analysed, where inference runs, and how many agents are in play. A site analysing eight streams on its own hardware is a different shape of deployment from one analysing four hundred streams in a managed tenancy.

The engagement normally starts as a paid pilot on one site. A pilot has a defined scope, a defined period, and a review at the end where the agent's events are assessed against your own judgement. That structure gives both sides a real cost basis before anything larger is committed, and it means the second conversation is about evidence rather than about projections.

We do not publish accuracy figures, detection rates or performance guarantees, because they depend on camera placement, lighting, the object classes involved and the rule you write. Those are established against your footage during a pilot. If a vendor quotes you a percentage before seeing your site, that number is not about your site.

What the pilot includes

Scoping, camera onboarding, agent configuration, the shadow period with event review, and the go-live session. The event history stays yours whatever you decide afterwards.

Before you book

Questions about the demo itself.

If your question is about the product rather than the session, the features page and the security page go deeper.

Is the demo a sales presentation?

No. It is a screen share where we configure an agent against the workflow you describe. You will see the interface, the rule being written, and the events it produces on footage. If the workflow you bring is not something the platform handles well, the useful outcome of the call is hearing that.

Do we need to share footage in advance?

No, and you should only share what your own policies allow. A clip you can screen share during the call is enough. If you cannot share anything, we will use representative footage from a comparable environment and say clearly that it is representative rather than yours.

What if we do not have a camera covering the area?

Say so on the form. Camera placement is the constraint that decides most deployments, and it is worth establishing early whether the view supports the rule. If a camera needs to move or be added, that is a site decision and we would rather flag it before a pilot than during one.

Can we record the session?

Yes, and it is usually worth doing, because the agent configuration we build is the thing your team will want to refer back to. We do not record calls ourselves.

How long before we can start a pilot?

The walkthrough is usually followed by a scoping session with the operational owner of the workflow, and onboarding starts once that scope is agreed and the runtime has network access to the cameras. The part that most often takes time is the internal decision about deployment model, which is why security and network colleagues are worth including early.

What happens to the details we submit?

They are used to respond to your request and nothing else. The privacy policy sets out what the site collects and how to have it removed. Note that this form is not yet connected to a server, so a submission through it does not currently reach us — see the contact route below.

Contact

If the form is not the right route.

This website is a static site, and the demo form on it is not yet connected to a server. You can use it to see exactly what we ask for, but a submission through it does not currently reach anybody. Until that changes, email is the reliable path.

[CONTACT EMAIL]

  • Email the address above with the workflow you want to look atOne or two sentences is enough to start.
  • Include your site type and roughly how many cameras cover the areaIt decides whether the view supports the rule.
  • Send a clip only if your own policies allow itDo not send footage you are not cleared to share.
  • Every request is read by a person, not routed into a queueThe same statement the form makes.

One workflow is all we need to start.

Bring the problem your team already complains about. That is enough to configure a first agent and to see whether the platform earns its place on your site.

Browse the scenarios See how a pilot runs