Review quality
Did qualified reviewers reach a supported decision?
Compare review time, correction effort and unsupported approvals on the same set of claims. Retain disagreements and state the sample size.
The Superlimen measurement approach
Measure what happened. Keep the limits visible.
A review decision, an AI citation and a customer purchase are different events. A useful report preserves that distinction.
AI visibility
What appeared in a recorded answer?
Keep the prompt version, API surface, response, timestamp and citation annotations. Report mentions and citations separately, alongside failures and sample size.
Business outcomes
What did the programme cost and produce?
Record costs, review time, conversions and revenue observations. Keep pipeline influence separate from realized revenue and account for implementation and review effort.
A pilot worth evaluating
Agree on the comparison before the first result.
- 01Define the task
Choose a representative claim set and record how your existing process handles it. Include unsupported statements and difficult evidence gaps.
- 02Assess independently
Have qualified reviewers label support, record disagreements and inspect the same evidence versions. Define what counts as an unsupported approval.
- 03Compare effort and quality
Measure median review time and correction effort together with approval errors. Document differences in claim complexity and reviewer experience.
- 04Report the limits
State sample sizes, missing observations and measurement windows. A small sample with no observed error does not establish a guarantee.
What the software records today.
Experiment records support baseline, treatment and control observations, costs, gross margin, revenue, review time, conversions and pipeline influence. Authorized systems can submit normalized observations through the ingestion API.
The visibility module records answers and citation annotations from a configured web-search API surface. It requires separate activation. Its provider and surface are identified in the records and in the technical capability details; observations do not represent every consumer search experience.
Cohort comparisons describe the observations you supplied. They do not establish causality. Superlimen has not published an independently validated review-time improvement, revenue uplift or competitive performance benchmark.