Usage gives organizations a single view of how their connected AI systems are actually being used, what those systems cost, and whether that spend is producing value. It brings adoption, utilisation, and ROI together in one place so leaders can make decisions based on real activity rather than assumption.
The page works at several levels. A top-level overview summarises adoption and value across every connected system, while drill-in views break the same picture down by individual system, department, and person. Usage also classifies the prompts flowing through each system into use-case clusters, giving a clear answer to what work people are doing with AI — and feeds that intelligence directly into the Adoption area.
Purpose
Usage helps organizations:
- measure adoption and utilisation across every connected AI system from one place
- see which users and departments are active, ramping, or sitting on unused licenses
- understand the real cost of each system, department, and person, including wasted seats
- discover what people are doing with AI through automatically distilled use-case clusters
- estimate the net value, or ROI, of AI spend and tune the assumptions behind it
- surface the proven use cases and idle licenses that Adoption can then act on
Usage is the measurement backbone of RexCommand’s optimisation tools. It establishes who is using AI, what it costs, and what it is worth, before that intelligence is turned into action in Adoption.
Overview of Page Sections
Headline metrics
The top of the page summarises the connected estate at a glance: active users, adoption rate, total prompts, and unused seats, each with a trend against the previous period and the spend it represents. A time-range control (24 hours, 7 days, 30 days, 90 days, or Lifetime) sets the window for every metric, chart, table, and ROI figure on the page.
Prompts trend
A trend chart plots activity over the selected window, with a separate line for each connected system. It can be switched between prompt volume, active users, and adoption rate. Systems that are detected only through network telemetry, rather than prompt content, are drawn with dashed lines so it is clear which signal each line is based on.
Systems, Users, and Departments
Three tabs break the same picture down by system, by person, and by department. Each is a sortable table covering assigned versus active seats, adoption rate, prompt volume, trend, and spend, with department and person tables also highlighting wasted or underused seats. Selecting any row opens the matching detail view for a closer look.
Use cases
The Use cases tab answers what people are actually doing with AI. Prompts are sorted into activity categories such as internal operations, document drafting, summarization, research, coding, and data analysis. Each category expands to reveal the distilled use cases within it — concrete, recurring tasks complete with a short description, the estimated time saved, the people and systems involved, and a reusable prompt template. From here a use case can be pushed straight into Adoption.
Net ROI
A Net ROI panel converts usage into estimated value. It subtracts the productive spend on used seats from the estimated time saved (valued at your organization’s hourly rate) to show a monthly net value and a return multiple, along with a breakdown of where that value comes from. While time-saved is still an estimate, the panel is clearly marked as a preview, and surveys in Adoption can replace those estimates with validated figures over time.
Value assumptions
The assumptions behind ROI are adjustable. You can set a default fully-loaded hourly cost used company-wide, and add per-department overrides where some teams’ time is worth more than others. These settings feed every ROI figure across Usage.
System, department, and user detail
Each row in the breakdown tables opens a dedicated detail page that carries the same time-range control and a scoped Net ROI panel.
System detail shows active users, prompt volume, assigned and dormant seats, and an activity trend for that system alone, alongside its connector status. Sub-tabs cover the people assigned to the system, the use cases it supports, the departments using it, and a Settings area. The Settings tab is where commercial data lives: cost per seat, billing cycle, purchased seats, contract dates with renewal alerts, and any one-off credits. These inputs drive the spend and ROI numbers across the rest of RexCommand, so keeping them accurate matters.
Department detail shows how a team’s seats landed — active, visit-only, dormant, or never used — together with how frequently people actually open the tools and which assigned employees are wasting their seats. User detail profiles a single person: their licenses, prompt volume, active days, peer ranking, daily activity, and a cost stack showing what they cost across every AI seat they hold.
Connecting an AI system
New systems are added through a guided wizard that walks through choosing a connector, connecting it, entering cost and seat information, assigning people, and reviewing. Connectors come in two tiers. Prompt-tier connectors, such as Microsoft 365 Copilot Chat, bring in prompts, identity, and seat inventory and can classify use cases and measure per-user activity. Visit-tier connectors, such as the FortiGate syslog collector, detect that an AI domain is being used from network telemetry but never see prompt content, so those systems appear in adoption counts without contributing to use-case or time-saved analysis.
How this supports Adoption
Usage and Adoption are two halves of one workflow. Usage decides what is worth acting on — the proven use cases, the idle licenses, the time-saving wins — and Adoption is where those become outreach. The distilled use-case clusters that Usage produces are exactly what Adoption’s nudges, campaigns, and surveys are built from, and a use case can be sent into Adoption directly from its detail view. The loop then closes in the other direction: when a survey validates how much time a use case really saves, that figure flows back into Usage and replaces the estimate used in the Net ROI panel.
Notes
- Assigned, active, and idle are distinct. Assigned users hold a license; active users have actually used the system; idle or dormant users hold a license but are not using it. “Idle” belongs at the system level — within a single use case the right phrasing is that someone has not tried it yet.
- Time-saved and ROI are estimates until validated. They are based on declared use cases and average prompt cadence, and are marked as a preview until surveys confirm the real numbers.
- Personal prompts and visit-tier systems are excluded from time-saved totals — personal prompts are shown but not counted, and visit-tier systems carry no prompt content.
- Use-case classification refreshes on a schedule. The Use cases tab shows when it last refreshed and when the next refresh is due.