Use methods that need no recruited people, and know what each one misses. In a heuristic evaluation, three to five reviewers check the interface against Jakob Nielsen's 10 usability heuristics. In a cognitive walkthrough, your team steps through one task the way a new user would. An AI user test, such as Menso's, sends a persona through a task on your live site and records every step. Your analytics show where current visitors leave. None of these is a usability test with real people, and none replaces one.
Last updated
| Method | Who does the work | What it finds | What it misses |
|---|---|---|---|
| Heuristic evaluation | Three to five people, each checking the interface alone against Nielsen's 10 usability heuristics | Breaks from well-known usability rules: unclear status, inconsistent labels, unhelpful errors | Whatever no reviewer happens to spot; and how a person who doesn't know the rules really behaves |
| Cognitive walkthrough | Your own team in one workshop, stepping through a single task as a new user would | Steps where a newcomer would not know what to do next, or would not see that it worked | Everything outside the chosen task; the team already knows the product |
| AI user test | A tool such as Menso: a persona works through a task on your live site in a real browser | Where one first-time visitor hesitates or gets stuck, with a replay of every step | How real people feel and decide; one run is one path (n=1) |
| Your analytics | The visitors you already have; no extra people | Which step of a funnel people leave at, and how many | Why they leave; and a page nobody visits yet |
Each reviewer goes through the interface alone and notes every place it breaks one of Nielsen's 10 usability heuristics, then the reviewers merge their lists. Nielsen Norman Group recommends three to five independent evaluators, because any one of them misses issues. In Nielsen's six case studies, a single evaluator found a given major problem 42% of the time on average, and a given minor problem 32% of the time.
A cognitive walkthrough checks how learnable one task is for a new user. The team picks the task, writes out the steps, and at each step asks four questions:
A “no” marks a step to fix. The weak point is the reviewers: they already know the product, so they have to imagine not knowing it.
An AI user is a language model given a persona and a task. It opens your live site in a browser, decides what to click and writes down what it is thinking at each step. Research prototypes such as UXAgent (Lu et al., 2025) run many simulated users against a website before a study with people. Commercial tools differ in what they run on: the comparison page sorts them.
In Menso, you paste a URL, pick a template or let Menso write the task, and get a replay of every step, the friction it hit and a TRACES score. The how it works page covers the details, and the signup flow guide walks through one task end to end.
One public run from the Product Hunt board, with its review note. It used Menso's Purchase intent template: the AI user had to decide whether to sign up. Pricing lists what a test costs.
Every run here used Menso's Purchase intent template: the persona Zainab Salazar, a freelance UX consultant, reads the homepage, the pricing page and the FAQ, decides whether she would buy or sign up, and stops before paying. It is a route-specific diagnostic, not a product-wide score.
TRACES 83.5/100
Zainab scrolled the homepage for five steps without finding a pricing link, and at step 6 typed the pricing address herself. At step 8 she reached the Free vs Enterprise table, then read the help centre and the docs FAQ, and at step 16 decided to sign up for the free tier.
T 4 · R 4.5 · A 4 · C 4 · E 4.5 · S 4 (each out of 5)
Review note: The actor never opened the top-right menu, where Pricing sits; this is not a claim that the site has no pricing link.
Made one of these products? Write to support@menso.io and we will correct or remove the example as soon as we can.
Not strictly: usability testing means watching people use the product. What you can do without recruiting anyone is a usability inspection (a heuristic evaluation or a cognitive walkthrough), an AI user test, or a look at your analytics. Each finds likely problems early. None tells you how your real audience feels.
In a heuristic evaluation, reviewers judge the interface against a list of principles. In usability testing, you watch representative people try real tasks. Nielsen Norman Group describes heuristic evaluation as a way to find likely issues without participants, and says it cannot replace user research.
No. Nielsen Norman Group's advice is not to use synthetic-user research as a replacement for research with real users. An AI user is useful before that: it gives you a first pass by someone who has never seen the product, so the people you do recruit later spend their time on questions an AI cannot answer.
Jakob Nielsen's 2000 analysis found that a first study with five participants finds about 85% of the usability problems, provided the users are comparable and use the site in similar ways. His advice is to fix those problems and test again rather than run one large study.
On Menso, a Fast test is 40 credits, about $0.64 on Pro, and a Deep test is 300 credits, about $4.83. New accounts get 100 credits free, enough for 2 Fast tests, with no card. Other AI user tools and their prices are on the comparison page.