AI
Cloud-native rental software has a habit of leading with AI for outbound sales — scrapers, BDR bots, generated cold emails. CloudERM doesn't. Our AI investment goes into the workflow your team already runs: inspections, pricing, search, planning. The features below are either shipping today or actively being built; the status tag on each tells you which.
Hour-meter capture
Near-termInspect mobileHour meters are the most-edited field in an inspection — and the most-misread. The crew snaps a photo of the meter on dispatch; an inspection without an hour-meter reading is incomplete; somebody back at the office squints at the photo to type the number into the system. That's three steps where a digit can flip.
InspectERM's hour-meter capture closes that loop. When the picture is taken; the hours are recognized and the field is pre-filled. The operator confirms or corrects and moves on. The math afterward — projected hours, maintenance interval, billing — runs against a number that hasn't been transcribed by hand.
Damage spotting
Near-termInspect mobileOff-rent inspections are the moment damage disputes happen. The traditional workflow is: yard hand stares at two sets of photos taken weeks apart and tries to remember whether that scratch was there at dispatch. Mistakes here turn into billing disputes that take days to resolve.
The Inspect app stages the dispatch and off-rent photos side-by-side, with a vision model flagging panels, decals, hydraulic lines, and lighting points that look meaningfully different on return. The model surfaces candidates — it does not auto-bill. Every flagged area gets a confidence score and the original side-by-side imagery, so the operator decides whether to charge.
Pricing
Near-termWeb appSetting a new asset's rental rate today is a judgment call: a yard manager looks at age, brand, comparable assets, recent utilization, and the going regional rate, then picks a number. That number may live unchanged for years.
CloudERM's rate recommendations factor in current utilization, asset age, accumulated maintenance cost, and comparable rates across the fleet — then suggest a per-asset daily / weekly / 4-weekly rate. The same recommendation runs for sale price. The operator sees what drove the recommendation — utilization trend, age penalty, maintenance burden — and decides whether to accept it. No black-box pricing; no "the algorithm says so."
Search
Near-termCross-cuttingReporting in legacy rental software means picking a canned report and waiting for an export. Want a slice the canned report doesn't cover? You file a ticket.
Natural-language search lets the operator query the platform the way they'd ask a coworker: "show me last quarter's quotes for skid steers in Austin," "which assets had hydraulic-fluid flags last month," "what's the average dispatch duration on the loader fleet by category." The query runs against real platform data — quotes, reservations, assets, inspections, work logs — and returns rows you can drill into or export. SQL not required.
Fleet planning
Near-termWeb appKnowing on Monday morning which categories are going to be tight by Friday is the difference between an okay week and a stressful one. Legacy systems leave this to the yard manager's memory — patterns held informally, lost when staff turn over.
Demand prediction uses historical utilization, seasonality, and the current reservation pipeline to forecast which categories will run hot a week and a month out. The manager sees which add-ons to stock; the dispatch team sees where to rebalance. Each forecast comes with the inputs it used — "loader demand 18% above seasonal baseline because of three large reservations on the books" — so the team can pressure-test the prediction before acting.
MCP server
Coming soonCross-cuttingThe Model Context Protocol is the emerging standard for letting AI clients — Claude Desktop, Cursor, ChatGPT, and others — talk to external systems. We're building a CloudERM MCP server so any MCP-aware client can read your platform data and take action on it.
The use cases are operator-facing, not sales-facing. Ask your AI for the utilization on a category and let it pull live numbers. Have it draft a quote against your real category tree. Summarize last week's work logs by crew. Surface assets that haven't shipped an inspection in 30 days. The MCP server runs against the same authentication and per-permission boundary the web app uses, so an agent only sees what its authenticated user is allowed to see, and every write action is logged.
How we think about it
Cloud-peer competitors built AI to send cold messages to your prospects — scraping building permits, automating BDR pipelines, generating sales emails. CloudERM's AI lives in the workflow your team already runs. Inspections, pricing, reporting, planning. We don't believe your prospects want a robot in their inbox; we don't ship one.
Every AI surface in CloudERM is a suggestion the operator accepts, modifies, or overrides. The damage analysis model flags candidates — the operator decides whether to charge. The rate recommender suggests pricing — the manager picks the final number. Demand prediction surfaces a forecast — the team uses it to plan, not to be governed by it.
When AI suggests a rate, an inspection finding, or a forecast, the inputs that drove it are visible alongside the output. Utilization curve. Age penalty. Maintenance trend. Comparable rates. Confidence score. You see what the model saw; you decide what to do.
Strengths across the platform, with the "AI for ops, not sales" trade-off in context.