Human taste for AI made designs
Ask real people whether something an AI designed looks good, before it goes in front of customers.
How it works for this decision
Your agent sends
{
"question": "Does this design look ready to ship?",
"context": "An AI design tool made a landing page for a children's dentist: black background, red headings and a skull in the logo.",
"options": [
"Ready to ship",
"Needs work"
],
"verifiers": 3
}A verified person sees
Does this design look ready to ship?
An AI design tool made a landing page for a children's dentist: black background, red headings and a skull in the logo.
You get back
{
"answer": "Needs work",
"agreement": 1,
"answers": 3,
"votes": {
"Ready to ship": 0,
"Needs work": 3
}
}The result shown is an example. Try answering the task yourself.
Why a human check
AI tools now make logos, landing pages, ads and product photos in seconds. They can't tell when something looks cheap, off brand or just wrong, and a model judging another model's taste shares its blind spots.
Taste is a human call. Several people looking at the same design and agreeing it works, or doesn't, is a quick read on what your customers will feel.
- Send the design as an image_url so people see exactly what customers would.
- Compare two versions by asking which one looks better.
- Ask about one thing at a time: does it look trustworthy, does it fit the brand, is the text readable.
Questions
Who are the people judging?
ID verified people from the general public, the same kind of people who will see the design. Send it to your own design team instead with route: "team".
How do I start?
Get your API key at plures.si. Start with your own team: the first 50 decisions each month are free. Add credit when you want ID verified people on the rest, live right away.
More use cases
- Taste and ethicsMoral and ethical judgment callsIs this message fair to send?Fair to sendCrosses a line3 of 3 verified people agreedHave people judge whether what an agent is about to do is fair and right, not just allowed.See how it works
- CustomersCatch upset customers your support agent misreadsIs this customer upset?YesNo3 of 3 verified people agreedHave people check tone and intent before an AI support agent closes, refunds or escalates a ticket.See how it works
- CustomersVerify refund and dispute decisions before money movesWhat is this customer asking for?Full refundReplacement partInformation only3 of 3 verified people agreedAsk verified people whether a refund request is legitimate and what the customer is actually asking for.See how it works