Journal

The Highest Cost in Commercial Building Operations Isn't on Any Invoice

System friction is the greatest unmanaged cost in building operations, and right now, managing a commercial building is an endless series of broken handoffs.

A work order comes in, and someone has to key it into the CMMS. Maybe the system recommends a provider; maybe it doesn't. From there, it's a manual scramble to schedule the job and verify it's complete. Then comes the paperwork. Someone matches the invoice to a PO, chases down three levels of approval, which can take up to 11 days, and keys the data into accounting. Add a compliance flag that must be cleared before final payment clearance, and the chaos is complete. By 9 AM, the operations director has opened twelve systems and solved nothing, because human effort is the only glue holding the process together.

Sentinel removes the busywork.
We don’t add another dashboard or rip out your existing tools.
We simply unify the workflow from request to invoice. AI works in the background, centralizing your data and keeping everyone on exactly the same page.

The cost no one budgets for

Every time work moves between systems, someone loses the context and someone re-types the data. Payment slows, errors multiply, and nobody owns the delay because the delay lives in the gap between the systems. Operators have known this for years. The problem persists because most products fix one step and leave the handoffs broken.

AI was meant to fix this; in reality, it’s only made the problem worse. JLL's 2025 research found 88% of investors, owners, and landlords run active AI pilots, averaging five use cases. Among occupiers, 92% run pilots, yet only 5% achieve their goals. The breakdown is an operational failure.

A typical pilot starts by applying technology rather than analyzing and addressing the work itself: a team buys a predictive maintenance product, runs it in one building, and produces dashboards, but the dashboards don't create a work order, assign a service provider, or post an invoice.

AI added to a broken workflow doesn't fix the handoffs; it speeds them up. Six months later, the pilot has produced findings nobody can act on.

The teams getting results work in the opposite order: they connect the workflow first, and AI comes second, as part of operations rather than the headline feature.

The best automation is boring

A tenant submits a request, and the system sorts it by trade and urgency, assigns it to the right service provider, and creates the purchase order. When the work is done, the provider sends the invoice, and the system matches it to the purchase order, checks the line items, routes it for approval, and posts it to accounting.

The AI never surfaces, and it doesn't need to. The operations director gets fewer emails; accounting processes more invoices in less time; the service provider gets paid faster.

Sentinel's term for this is AI in the background. The automation operates within existing workflows and doesn't require operators to learn new ones. It connects owners, operators, and service providers in a single process: the AI reads the data, routes the work, and flags exceptions, while people make the decisions that matter.

A staffing problem with a software answer

Buildings need technicians, engineers, and tradespeople, but the supply keeps shrinking while the work grows more complex. The US skilled trades need 439,000 workers by 2025; it’s a structural shortfall, and it won't close in the next hiring cycle or the one after.

No operations team can hire its way out of a gap that size, which makes automation the only viable staffing strategy.

A team with automated workflows can handle more service volume with the same headcount: operators re-key less data, chase fewer approvals, and match fewer invoices, and the freed hours go to work that needs judgment. The shortage won’t disappear, but it does become manageable.

Operations software approaches service providers as a cost to control: track them, score them, and squeeze them. But Sentinel takes the opposite position: owners and service providers run the same workflow from opposite ends, and the software should treat them that way.

When a provider gets a clear scope, fast approval, and prompt payment, the owner gets faster service and cleaner records. When a provider loses hours to manual invoices, the owner pays too, in slower response times and weaker work.

With Sentinel, both sides are connected and share the same goal.

One job, start to finish

Most products in this category fix one part of the workflow: a work order product, an invoice product, a compliance product. Each makes sense on its own; together, they create more logins, more exports, and more manual steps when the systems fail to connect.

Sentinel treats request-to-invoice as a single job, running from the first tenant request through service delivery, invoice receipt, approval, and payment. Owners and operators can start anywhere in that process with the systems they already have; Sentinel connects what's there without asking anyone to replace what works.

The AI reads patterns, routes exceptions, flags errors, and removes the manual steps that slow every service event. It doesn't announce itself, and it doesn't add another dashboard that needs another review.

Buyers have stopped trusting the label

When every tech company claims to sell AI, buyers stop trusting the claim and start asking for results, which is fair. They never wanted to buy an AI dashboard in the first place; all they’ve ever wanted is a building that takes less manual work to run.

Sentinel's position is simple: the request-to-invoice workflow runs through too many disconnected systems, AI works best when users never have to think about it, and service providers belong in the workflow as partners. The property leaders who rank portfolio optimization as their top goal won't reach it by adding another pilot to a broken workflow. They need less manual work in the technology their buildings already run on.

At Sentinel, we have one belief that rules everything we do: less admin means better buildings.