UI/UX and journey reviews
Review a live product.
See where people drop out and which screens confuse them.
Apply for a UI/UX reviewSee where your product breaks. Before your customers do.
Simulated testers use your product like real people. Our team checks every finding. You get ranked fixes, each with proof.
Use cases
What you can do. One test, three ways.
Prototype testing
Test it before you build.
Run the key tasks before you build.
Apply for a prototype testBusiness questionnaires
Ask your audience.
Ask your audience about a name, message or offer.
Apply for a questionnaireWhat your report shows
Thousands of steps. One ranked list. Each finding comes with proof.
Simulated testers try the same task again and again. We group what they ran into, replay it and have an analyst check it. You get the fixes in order.
Validation reportTidewell Invest onboarding
In one real test4,708 steps1,099 screenshots
7 findings, ranked
- Risk answer clears the portfolioBlocker · 14 of 40 runs Observed, replayed and checked by an analyst
- Refresh deletes the ID documentBlocker · 22 of 40 runs Observed, replayed and checked by an analyst
- No arrival date on the depositImportant · 17 of 40 runs Observed
- Two names for one stepImportant · 11 of 40 runs Observed and checked by an analyst
- Phone field rejects spacesImportant · 9 of 40 runs Observed, replayed and checked by an analyst
- Back resets the deposit amountMinor · 6 of 40 runs Observed and replayed
- Fees hidden below the foldMinor · 5 of 40 runs Observed
3 withdrawn: our tool was at fault
- Continue did not respondOur browser clicked too early
- Identity check timed outOur test runner timed out
- Error text on the depositQuote did not match the screen
Blocker
Saving the risk profile clears the chosen portfolio.
14 of 40
simulated testers got stuck here
What the tester sawtheir product
Select a portfolio to continue
The only words on the screen. Nothing says the choice was cleared.
What we foundour analysis
- Expected
- The portfolio stays selected, or the app explains why it changed.
- Actual
- The selection disappears without a message and Continue is disabled.
- Reproduced by replay
- Checked by an analyst
- Choose “Balanced” on the Portfolio step.
- Go back to “Risk profile” and change the answer to question 4.
- Save.
- Return to “Portfolio”.
Fix this first
Keep the selection after saving, or show which answer changed the result.
Simulated testers
Who the testers are. And how we make them.
Each tester is a behavioural profile with a life story, habits and a goal. Open one to see the profile.
A few examples from one test group4 of a full test group, built for an investment app
-
Careful saver
- Digital skill
- Patience
- Privacy concern
GoalMove part of her savings into a low-risk fund.
-
First-time investor
- Digital skill
- Patience
- Privacy concern
GoalStart a monthly plan in one sitting.
-
Busy parent
- Digital skill
- Patience
- Privacy concern
GoalOpen a savings plan for her daughter.
-
Fee-conscious retiree
- Digital skill
- Patience
- Privacy concern
GoalKnow the yearly cost before he signs anything.
- + many more
Any input works. Four ways to build a test group.
- Own distribution data
- Independent review
- Team confirms
- Versioned for retests
- Your CRM
Exact people, only with your permission.
- Your audience brief
Close to your brief, still diverse.
- Population statistics
Everyday people matching the real mix.
- Specific roles
A board or advisor network, from profiles and interviews.
Clients
Teams we helped. What they decided.
“The first test of our new module found over 30 edge cases that did not work. After the fixes, 15% more user actions were completed.”
What Validator didTested the newly prepared module with 100 simulated advisers.
Result
30+edge cases that did not work
0 of 33runs finished the declined-advice path
+15%completed actions after the fixes
“From anonymized data about our users and their telecom spend, Validator built personas and asked what a competing operator would have to do to win them. We took that offer to a competing operator and to our board.”
What Validator didBuilt personas from the anonymized package and asked what would make them switch tariff.
Result
The offer helped start a partnership and an offer for Patron GO users.
“Validator helped us dodge ‘Flatu’. For some people it sounds like flatulence, and we did not want a name that makes anyone smile for the wrong reason. We are Flatari.”
What Validator didTested 5 name candidates with 250 simulated personas in a simulated survey.
Result
250simulated respondents read the 5 names. Many flagged ‘Flatu’ as negative: for some it sounds like flatulence.
Name dropped. The company is Flatari.
“Our customers told us the process is hard. Now we can see exactly where people get stuck, and which problems come first.”
What Validator didSent 100 simulated customers through car, motorbike, van, camper and travel insurance to see where people get stuck.
Result
101possible bottlenecks found, 59 of them met by 3 or more simulated customers.
The bottlenecks, ranked by how many customers hit them.
“Our simulated personas read ‘Premium’ as a paid plan and many would tap ‘Later’. ‘Patron Extra’ won in every group.”
What Validator didTested the name options with 30 simulated personas in 5 groups.
Result
5 groupsand in every one ‘Patron Extra’ beat ‘Premium’ for a free tier.
Patron Extra won in every group and was recommended.
“We ran Validator on our own landing page and simplified it.”
What Validator didTested how visitors understand our complex products.
Result
Validator can help you do the same for your users. Test yours
Why Validator
Fast. Backed by proof. Close to real people.
Hoursfor 100 simulated testers per run
Not weeks.
A full run of 100 simulated testers finishes in hours. No recruiting, no scheduling, and you can repeat it on every release.
- Screenshot
- Quote
- 14 of 40
- Replayed
- Analyst
1,099screenshots in one real test (4,708 steps)
Every finding comes with proof.
Each finding shows the screen, the exact quote and how many runs hit it, for example 14 of 40. We replay it in a fresh browser and an analyst confirms the important ones. In one study we withdrew 54 of 548 findings because our own tool was at fault.
Goalsevery tester has one, and a level of patience
Not scripts.
Testers take wrong turns, go back and give up, so you see where different kinds of people lose the way, not just whether a script passes. It is a strong signal, not a verdict. When real users are the better next step, we say so.
For regulated teams
Safe for internal systems. Made for regulated industries.
AI testing without giving up control: your data stays in our controlled environment, serves only your test and is deleted on the schedule we agree.
So banks, insurers and other regulated teams can bring Validator to their internal systems, not only to public apps.
How your data moves
You
Test or staging environment
Test data only
Data stays with us
Validator + our team
Simulated testers
A person reviews every report
You
Report
Only people you authorise
Agreed with your security team before any test
- Isolation
- Network access
- Data handling
- Retention
-
Your data stays with us.
It does not leave our controlled environment, is used only for your test and is deleted on the schedule we agree.
-
Internal tools in scope
Staging, test environments and internal tools: we test what sits behind your login.
-
Test data only
No real customer data and no passwords in the brief. Access is agreed with you first.
-
We respect GDPR
GDPR-aligned processing, with a data processing agreement on request.
-
NDA and non-compete
NDA before we start. Your product is never shown to others or used for another client.
-
Reviewed, then shared
A person reviews every report. Only people you authorise can open it.
About us
Built by the team behind emable. At home in financial distribution and AI.
emable builds connected solutions for financial distribution in Europe, on a turnkey basis. The same team delivers them, runs them and keeps them running.
Validator comes out of that work: we test our own AI products and apps the way we now test yours.
Talk to the team
A product of emable.ai
emable
One team builds and runs it all
- CRM and sales engineFor advisers, teams and their clients
- Intranet and mobile appsFor the people in the network
- MCP hubConnects company systems to AI assistants
- European distribution dataDECZPLFRIT
- ValidatorBorn from testing our own AI products and apps This page
- Own product, not a resellerOur engine, our team, our method
- Works with the NUDA groupOur first client and partner
- Built for regulated financial distributionAdvice and insurance flows
- Tests run by people who ship softwareThe team that builds emable products
Talk to us
Tell us what to test. We reply with a scope.
Describe the flow and who uses it. Our team reads your brief and proposes what to test before anything runs.
What happens next
-
Step 1
Brief received, reviewed by our team
A person on our team reads every brief.
-
Step 2
Scope for your approval
We propose tasks, edge cases and simulated tester profiles. Nothing runs before you approve.
-
Step 3
Test and report
Simulated testers work through your product in a real browser. You get findings ranked by impact, each with evidence.
Q&A
Before we start. The short answers.
Questions
All eight answers
Apps and questionnaires
UI, UX and user journeys of live products and clickable prototypes, tested by simulated customers in a real browser. And business questionnaires about a name, message or offer. A person on our team checks every result.
Behavioural profiles
In some cases better than a human test, and not a replacement for people. Each tester has a behavioural profile: experience, digital skill, patience and privacy concern. That is why every app finding comes with proof from a real browser and the important ones are checked by a person. Next, we will publish our own benchmark against real people.
Test data only
We test on test or staging environments with test data only. During a test, screenshots and screen text are processed by our AI model provider. Reports are visible only to people you authorise, and a person on our team reviews every report before you see it. Validator is separate from other emable products and data.
Access agreed first
Yes, once access is agreed with your team. The brief has no password field. Before any test of an internal system we agree isolation, network access, data handling and retention with your security team, and we use test accounts you provide.
Simulated survey
Yes. Describe your audience and ask about a concept, a name, a message or an offer. Simulated personas answer a questionnaire and you get a GO, CONDITIONAL or NO-GO verdict with the reasons and every answer. It is a simulated survey and we label it as one.
The limits
Your real conversion, how many customers will leave, or what a regulator will decide. It shows where people are likely to get stuck and why, so you know what to fix now and what to check with real users.