"I don't care about your security system," says your CFO
Tell finance a new system reduces incidents and watch their eyes glaze. Show them the black holes in the balance sheet and they sit up. Parker Miller on selling video AI internally.
Show finance the black holes draining the balance sheet and they sit up. Tell them your new security system reduces incidents and you have lost the room by the second slide.
In an agentic video system, cameras are endpoints collecting intelligence for agents to act on. But that distinction is not always clear to the person approving the budget. VP-level buyers may see video AI as "more camera spend" even when nobody is asking for cameras. What is being proposed is an investment in the intelligence layer, which can find missed opportunities across loss prevention, security, safety, and operations.
Those problems are usually worse than leaders realise. Parker Miller, Spot AI's VP of Partnerships, has seen enterprises unlock as much as $15 million in savings from agents on cameras they already owned. Dangle that in front of a decision maker and you have a more tantalising proposition than "let's spend thousands on cameras, again."
What you are buying is a digital workforce
When executives consider a video AI system they ask security questions. Does it reduce incidents? Does it replace guards? The answer is yes, but that framing is narrow. It measures the system against what a guard could do, and a guard cannot watch every register and every line at once.
Announcing a camera refresh won't get org buy-in. If you start talking about stopping vehicle break-ins or POS transaction fraud, leaders sit up and take notice.
Frame video AI as a security cost and you get a security-sized budget. Parker's three moves for getting a bigger one:
1. Start from a hard-to-measure loss
Most of the organisations Parker talks to have financial black holes. Leaders know internal theft or SOP adherence is a problem, but if they cannot quantify it they cannot address it. Find where finance is already relying on guesswork. In retail, that is a shrink number they are already tracking. In manufacturing, it is the process where line-down events are most frequent and least documented. That is your baseline, and video will show whether the company has been under- or over-investing.
Once we get a real idea of what's happening by pulling in video data, it's often about three times worse than they think it is.
2. Follow the data to growth
When you report back, skip the table-stakes outcome of incident prevention and go deeper. Build pilots that expand from the first use case. In retail that means going from deterring loiterers to internal theft at the register, verified on video. In manufacturing it means going from missing PPE to the practices that keep the safest teams compliant, and in construction from recording unauthorised entry to stopping it.
When it becomes clear that video AI isn't just a security system, that it's saving the company a ton of money, people across teams start getting interested.
3. Build process expertise over time
Watching for exceptions is the human model. With video AI you can monitor everything and learn from it. Encode new standards based on what actually works: which shifts and seasons need more staff, which lines perform best and why, where loading bottlenecks sit, whether a site is on plan. Parker's view is that the teams treating cameras as a compliance tool will spend the next decade catching up to the ones using them to define "good" now.
You're either on attack mode or on the back foot. Where do you want to be?