Glossary
The terms Plarix uses to scope, build, and measure an AI system against a real process. Each entry links back to where it's used.
Economically-Engineered Agentic System
An AI agent system built and scoped around one specific business process, sized against a measured baseline rather than a general use case.
An EEAS starts with a process, not a technology. Before any system gets built, Plarix measures how the process runs today in cost and hours. The system is then designed and priced against that baseline, and deployed into the tools the company already uses. The name is deliberate: engineered, not experimental, and economic, not just technical.
Why it matters: A generic AI agent is built because the technology allows it. An EEAS is built because the baseline says it will pay for itself.
The ROI Gap
The space between an AI pilot a company has built and the result it can prove.
Most companies can point to an AI pilot. Few can say, in euros and hours, what it changed. That gap between activity and proof is the ROI gap. It's not a technology problem. It's the result of building before measuring. Plarix closes it by measuring the baseline before any build starts, and reporting the result against that baseline after.
Why it matters: A pilot that can't be priced against a baseline can't survive a budget review, no matter how well it works in a demo.
Process Baseline
The measured cost and hours a process consumes before any system is built.
The baseline is the number everything else gets compared against: current cost, current hours, volume, and error rate, measured directly rather than estimated. Every EEAS is priced and evaluated against its own baseline. Without one, there's no way to know whether a system saved anything.
Why it matters: You cannot report a savings number if you never measured what the process cost before you started.
Diagnose
The first stage of the EEAS Method: measuring how a process runs today.
Diagnose is where the process baseline gets built. Plarix observes the process directly, timing and pricing each step, mapping volume and error rate. This stage produces the baseline every later result is measured against.
Why it matters: Skip this stage and there's no way to prove, later, what the system actually changed.
Design
The second stage: modeling whether automating the process is worth the cost.
Design takes the baseline from Diagnose and models what an EEAS would cost to build and run against it. If the projected savings don't clear the cost of building and running the system, Plarix says so before any build starts.
Why it matters: Most AI projects skip this stage entirely and find out whether it was worth it only after the money is spent.
Deploy
The third stage: building and integrating the system into the process it was designed for.
Deploy is the build itself: an agentic system scoped to the one process measured in Diagnose, integrated into the company's existing tools rather than a new platform layered on top.
Why it matters: A system built for a specific, measured process does the job it was priced to do. A general assistant dropped into a company rarely does.
Measure
The fourth stage: comparing actual results against the original baseline.
Once the EEAS is live, Measure tracks real results, money saved and hours returned, against the baseline from Diagnose, on a set schedule. This is what turns a build into a provable result.
Why it matters: A system that isn't measured against its baseline is just a pilot with better branding.
The Scorecard
The running record comparing a process's baseline against its actual results after an EEAS goes live.
The scorecard is Plarix's reporting standard: baseline, current result, and the delta between them, in both money and hours, updated on a schedule. It's built to be handed directly to a budget review.
Why it matters: A savings claim without a scorecard behind it is a guess. A scorecard is the difference between "we think this helped" and "this saved €180,000 and 5,500 hours, and here's the baseline it's measured against."
Definitions maintained by the Plarix team.
Read the blog →