AI workflow evaluation. Know what to change.

One-workflow AI evaluation starts at $2,500 USD for the starter scope shown on this page. We confirm access, scope, exclusions, timing and a written price before work begins. Third-party software and usage charges are separate; ongoing support is optional.

In 30 minutes, we discuss one task, your current tools and the result you need. We check whether the starter fits and what is needed for a written proposal. If paid discovery is needed, we explain its scope and price first. You do not need a prepared brief or a paid assessment to book the call.

We agree a baseline and acceptance checks with the person who owns the work. A reproducible baseline and candidate comparison support a change-or-keep recommendation, with limitations stated. Where relevant, we also compare manual effort, correction work and operating costs. These checks guide the handoff; they are not a promise of revenue or savings.

A scoped evaluation can produce a baseline, a representative test set, controlled comparisons, and a prioritized recommendation. Record failures as well as aggregate results. The recommendation should identify the evidence, sampling limits, expected trade-offs, and checks needed before any production change.

Choose one workflow and one decision, such as whether a model or retrieval change improves it. Define acceptable outputs with the people using them. Include ordinary tasks and known failures, then run the baseline and candidate on the same cases using consistent scoring.

Cost depends on the number of workflows and candidates, test-set preparation, access to usage information, and the amount of human review. Model calls and repeated runs can add evaluation expense. Agree the decision the evaluation must support before expanding the experiment.

Timing depends on whether a useful test set exists, how easily the current workflow can be reproduced, and who can judge output quality. Missing traces or disputed success criteria add preparation work. The plan should separate baseline preparation, comparison runs, and review of the results.

Sanitized examples, exported usage records, and a reproducible test environment may support an initial review. State what that evidence excludes, such as unusual live traffic or permission behavior. Production access and any production changes should be considered separately with explicit authorization.

Keep the evaluation set, scoring rules, baseline, and known failures under an identified owner. Repeat relevant checks when models, prompts, sources, or tools change. Any rollout needs monitoring and a way to reverse the change if live behavior differs from the test results.