Judge the work against the same standard.

Scores a screen or flow against seven product design principles, with findings tied to visible elements.

When to use it: A general design critique of a screen or flow, when you want a scored review rather than open-ended comments.

What you provide: A screenshot of the screen or flow, or a description that names the visible elements and their text.

How it reports: 1-5 per principle; overall 0-100

Audits a screen or flow against five inclusive design principles for cultural fit, checking who the design includes and who it unintentionally leaves out.

When to use it: An inclusion review of a screen or flow: representation, cultural fit, access, language, and user control, judged from what is visible in the design.

What you provide: A screenshot of the screen or flow, or a description that names the visible elements, imagery, and text.

How it reports: 0-2 per principle; overall 0-10

Reviews a screen’s design for accessibility problems that are visible in a screenshot or description, judged against criteria derived from WCAG 2.2.

When to use it: At design time, before a screen is built, to catch the accessibility problems a picture of the design can already show, so they are fixed in the design instead of discovered in the code.

What you provide: A screenshot of the screen at a realistic size, or a description that names the text, its colors, and the controls. State the platform (web, iOS, Android, desktop) when it is known, because target and text expectations differ by platform.

How it reports: No overall score; each finding rated Blocker, Serious, Moderate, or Minor; criteria that cannot be judged from the input are reported as Cannot judge

Reviews every piece of text on a screen and returns a revision for each one that fits the same space and does the same job better.

When to use it: Before design handoff, when a screen’s words need a dedicated pass: headlines, buttons, labels, empty states, and error text that were written in the layout and never reviewed as writing.

What you provide: A screenshot of the screen, or the screen’s text supplied directly (a copy list, a transcript of the design, or the strings from the product). If text is supplied directly, say so in the result.

How it reports: None

Predicts what a first-time visitor will remember after seeing a screen for five seconds, and which notable elements they will miss.

When to use it: A first-impression check on a landing page, a hero section, an app screen, or any single screen whose job is to land its message and its main action within seconds.

What you provide: A screenshot of one screen at its full bounds, or a description that names the visible elements and their text. One screen per run.

How it reports: Memorability 0-10; simulated panel of 5; 3-of-5 rule

Predicts the three places a viewer’s eye will land first on a screen, in order, with positions a tool can draw as an overlay.

When to use it: To check whether a screen’s visual hierarchy sends attention where the design intends, especially when something decorative might be out-pulling the headline or the main action.

What you provide: A screenshot of the screen (required for coordinates). Optionally, the intended attention path: the elements the design wants seen first, second, and third. A text description alone can support the ranked prediction but not the coordinates; say so and omit them.

How it reports: Exactly 3 hotspots; intensity High 70-100, Medium 40-69, Low 20-39

Reviews a design submission, a multi-frame board with its rationale, against seven dimensions of professional design quality, the way a design leader would in a critique.

When to use it: Before a design critique, a handoff, or a portfolio review, when the question is not only whether the screens work but whether the work itself is good: well aimed, well grounded, well built, and well explained.

What you provide: The design as a multi-frame board (a board overview plus the frames in reading order), with whatever rationale the designer has written on the board or alongside it. A single screen can be submitted, but the dimensions that judge intent and evidence will usually be unjudgeable from a screen alone.

How it reports: 1-5 per dimension; overall 0-100 across the dimensions that could be judged

Reviews a research study plan before it is fielded, judging whether the questions are answerable, the method fits them, and the participants and materials will produce evidence the team can act on.

When to use it: Before recruiting begins for an interview study, a usability study, a diary study, or any planned piece of product research, when the cost of fixing the plan is still low and the cost of fielding a flawed one is not yet paid.

What you provide: The study plan as written: its goals or research questions, the method, the participant definition and screening approach, and whatever session structure, tasks, or materials the plan specifies. A short informal plan is a valid input; the review judges what is there against what the study needs.

How it reports: No numeric score; each criterion judged Ready, Needs work, or Missing, with an overall readiness verdict

Reviews a survey questionnaire question by question, finding the wording, scale, and ordering flaws that bias answers or make them impossible to give, and suggesting a corrected version of each flawed question.

When to use it: Before a survey is sent, while the questionnaire can still be changed: for satisfaction surveys, screening surveys, product feedback surveys, and any questionnaire whose results will be quoted as evidence.

What you provide: The full questionnaire text: the introduction shown to respondents, every question with its answer options and scales in order, and any skip logic described. The survey’s purpose and audience, when known, sharpen the review; without them the review judges craft alone and says so.

How it reports: No overall score; every flaw rated Invalidating, Biasing, or Burden, with a suggested revision for each flawed question

Reviews a research readout, the report or summary stakeholders will read, judging whether every claim traces to evidence, whether observation and interpretation are kept apart, and whether recommendations stay proportional to what the study can support.

When to use it: Before a research readout is circulated: the findings deck, the summary document, or the report a team will use to decide something, when the question is whether a reader can trust what it claims.

What you provide: The readout itself, in full: claims, quotes, numbers, and recommendations as the audience will see them. Access to the underlying evidence (session notes, recordings, response data) strengthens the review when available; without it, the review judges the readout’s internal support and labels every claim check accordingly.

How it reports: No numeric score; each criterion judged Holds, Needs work, or Missing, plus a claim-by-claim audit and a reliance verdict