Case studies
A regulatory consultancy
Two classification engines that turn a week of expert hand work into a pipeline at fractions of a cent per item.
Who they are
A consultancy that prepares data compensation assessments for pesticide registrations.
Shared without the customer’s name.
The problem
- Each assessment rests on a list of underlying data requirements.
- Analysts classified those requirements by hand.
- That took about a week of work per assessment, by the customer’s own estimate.
What runs on Studio
A pure automation pipeline, called through the API. No chat and no knowledge base.
- Two classification engines, each an agent: one for product-specific data, one for data requirements.
- Two workflows that run the engines over every item.
The numbers
- agent runs in the week to 3 Nov 2026
- 12,400+
- median time per item
- 2.8 s
- model cost per item
- ~$0.0002
Figures from the customer’s production Studio workspace, November 2026.
Why it matters
Studio turns expert manual classification into a pipeline at fractions of a cent per item, using agents, workflows, tests and tracing.
Start building with Krazimo
Build with confidence.
Deploy with evidence.
Define your guardrails, run your evaluations, and review the results. Bring your agents and workflows to production with the checks in place.