People make the final call
CXPR is human-led, science-backed, and AI-accelerated. An experienced reviewer is responsible for the scope, interpretation, priorities, and recommendations. AI can assist the work; it does not replace the expert who checks and approves the client report.
Jaspurr™ is our closed database of best practices. We’re building and curating it to support expert reviews with relevant research, standards, and guidance. A senior UX practitioner selects the relevant principles, checks the evidence, and approves your recommendations. AI helps organize the work.
Where AI can help
AI assistance can help organize public website observations, summarize selected research, explore copy alternatives, and structure draft recommendations. The reviewer determines whether an observation is accurate, whether a source fits the task, and whether the proposed action is useful for the client’s audience and goals.
A generated suggestion does not become a finding simply because it sounds plausible. We distinguish direct observations from interpretations, hypotheses, and areas that have not been assessed.
Evidence that can be checked
We check relevant claims against the public website and their original sources where available. We consider the source’s method, date, population, context, and limitations. Provider case studies are identified as such; an industry analogy is not treated as proof.
Research citations should link to the original source. Numerical results retain their original outcome and context. We distinguish relative changes from percentage-point changes, and do not add overlapping funnel metrics together. A published result from another organization is not a CXPR client outcome or a forecast for your business.
Bias, uncertainty, and human review
A consistent framework, source checks, and human review can reduce errors and challenge assumptions. They cannot eliminate bias, hallucination, or omission. We look for unsupported claims and material uncertainty, identify coverage limits, and revise recommendations when evidence contradicts them.
The Growth Wheel uses Findability, Clarity, Connection, Confidence, and Momentum to structure the review. CXPS is an expert assessment within a defined scope; it is not an independently validated predictor of growth, revenue, or conversion.
Confidentiality and appropriate inputs
A standard CXPR works with agreed public website pages. We do not need your login credentials, private customer data, payment details, or health records. Do not include them in a brief.
If an engagement would benefit from non-public material, we agree what may be used and the handling arrangements before it is shared with an AI service. We do not treat permission to view a file as unrestricted permission to repurpose it. We consider provider processing and settings for the agreed work. Information handling depends on the provider and service arrangements. See the Privacy Policy.
Images and representation
AI-assisted editorial images and writing may be used on this website, with human review before publication. Illustrative people or scenes are not evidence that a depicted person is a client, researcher, or participant in a named study. They must not substitute for real project evidence, testimonials, or measured outcomes.
The same standard applies to reports and recommendations: a polished presentation does not establish that a claim is true. Factual content, source interpretation, and final delivery remain subject to human review.
Decisions, limits, and feedback
CXPR does not autonomously change your live website, approve payments, or make decisions about individual customers. Design, development, and implementation are separate from the standard review. A review does not certify accessibility, security, or legal compliance.
If you find an error, a broken citation, or a concern about AI use, contact hello@thecxpr.com with the page or report section and the issue. We will review it. Missed agreed report deliverables are covered by the Delivery Promise.