Understanding Viable Simulated Users

Can you actually trust an AI's opinion about your customers?

It's a fair question. And the honest answer is: sometimes, and not blindly.

AI has made building products faster than ever. It hasn't made deciding what to build any easier — teams can ship in days what used to take months, and still have no idea if they're shipping the right thing.

Simulated users are one of the tools we built to close that gap. Used well, they're one of the fastest ways to find out if an idea is worth pursuing, before you spend real time, real budget, or real customer goodwill finding out the hard way.

Here's what they are, what they're not, and how much weight to actually put on them.

What are Simulated users?

Simulated users are AI-generated representations of the people in your target audience.

Each one is built with a specific background, role, set of goals, motivations, and decision-making criteria — designed to react to your product or marketing the way a real customer in that segment might.

They're not generic AI responses. They're simulated audience members created to evaluate your work within a defined customer context.

What do they actually do?

Here's a quick example of how this works in practice.

Say you're building a project management tool for operations managers. You have two positioning angles you're deciding between: one focused on saving time, one focused on reducing team miscommunication.

Rather than guessing — or waiting weeks to recruit research participants — you run both through Simulated users built to reflect that audience. Within minutes, you can see which angle resonates more strongly, where the messaging loses people, and what objections are likely to come up. Then you refine, and test again.

Depending on the test, Simulated users might:

• React to an ad or product message
• Review a landing page
• Compare positioning options
• Evaluate a product idea or proposed feature
• Decide whether they'd take a next step
• Flag what's confusing or missing
• Explain what increased or reduced their interest

Viable analyzes those reactions together to surface patterns across the audience — not just a collection of opinions, but a structured view of how a specific customer segment is likely to understand and respond to what you're testing.

What questions can they help answer?

• Does this idea feel relevant to the intended audience?
• Is the value proposition clear?
• Which message is most compelling?
• What objections might get in the way?
• Which audience shows the strongest interest?
• Where does the experience create confusion or hesitation?
• Is this idea strong enough to justify further investment?

Viable turns the reactions into findings, comparisons, and recommended next steps — so the output isn't more data to interpret, it's a clearer direction to act on.

The gap that makes this useful

We saw this ourselves this week, testing three different ways to describe a new Viable feature to product managers.

The version that scored highest on Viable's own message-audience fit score, a measure of how closely the copy matched what that audience cares about, didn't win. A different, more concrete angle converted twice as well, even though it scored lower on that fit measure.

Stated preference is easy to collect and easy to be misled by. Behavior is harder to fake, and it's the thing that actually predicts whether people convert, sign up, or buy.

That gap is exactly what simulated testing is built to catch.

How much can you trust them?

Treat Simulated-user results as directional evidence, not absolute proof.

Their value is in helping you identify patterns, surface weaknesses, and spot opportunities faster than assumptions or internal debate alone. They can show you where an idea looks strong, where it may struggle, and what to test next.

Results are most useful when:

• The audience is clearly defined
• The test reflects a realistic customer decision
• The experience is complete enough to evaluate meaningfully
• The findings are interpreted in context, not treated as final verdicts
• High-stakes decisions are confirmed with real customer behavior

Rather than just giving you a score, Viable makes the reasoning behind the results visible — explaining what influenced the audience, where reactions differed, and why the platform reached its recommendation.

What are their strengths?

Speed. Useful feedback in a fraction of the time it takes to recruit and coordinate traditional research participants.

Early access. You can test before a finished product, large ad budget, or complete customer experience exists.

Cheap iteration. Revise an audience, message, or page and test again — without restarting a lengthy research process.

Easy comparison. Especially powerful when you need to evaluate multiple audiences, messages, or product directions under consistent conditions.

Assumption exposure. They surface unclear language, missing information, likely objections, and gaps between what your team intended to communicate and what the audience actually understands.

A stronger starting point. By catching obvious issues early, you improve the test before spending money or asking real customers to participate.

What are their limitations?

Simulated users are simulations. They don't have real bank accounts, real workplace dynamics, or the full emotional and situational context of an actual customer.

They can't perfectly reproduce:

• Real-world buying behavior
• Unexpected human reactions
• Organizational approval processes
• The influence of timing, budget, or competing priorities
• Long-term product usage

They're also only as strong as the inputs. A poorly defined audience or unrealistic test scenario produces less useful results.

For these reasons, Simulated users shouldn't be the only source of evidence for a major, irreversible decision.

Their job is to reduce uncertainty — not pretend it no longer exists.

Where do they fit in the process?

Simulated users are most valuable early — and between major real-world tests.

1. Explore with Simulated users. Test early ideas, audiences, messages, and experiences quickly. Identify the strongest opportunities and the most important weaknesses.
2. Improve the concept. Use the findings to sharpen the product idea, positioning, ad, landing page, or experience.
3. Validate important decisions with real customers. When the idea is ready and the decision carries real risk, test with real people and measure actual behavior.
4. Keep iterating. Use Simulated users to explore improvements, investigate new questions, and decide what deserves the next real-world test.

Simulated and real-user testing aren't competing approaches. They serve different purposes — and work best together.

The simplest way to think about it

Simulated users help you answer: "Based on what we know today, how is this audience likely to respond — and what should we improve or test next?"

They don't replace your customers.

They help you reach your customers with a stronger idea, a clearer message, and fewer preventable mistakes.

That's the whole idea behind Viable: prove it before you build it.

If you want to see what a real test looks like: