Korin Mitrová

AI-powered synthetic data is gaining attention due to its ability to simulate people's behaviour and preferences at a lower price. According to a recent survey, 81% of marketers said they use or intend to use generative AI to generate synthetic data. Synthetic data forms the basis for creating 'synthetic users' that model how your actual users could respond or interact. These methods can be handy for rapid testing, pattern recognition, and generating hypotheses when time, budget, or user access is limited.

As this topic gains traction, it's helpful to ask: What are we learning from synthetic data? How reliable are synthetic users' answers? And can they ever replace talking to real people?

In user research, we aim to understand why and how people make sense of their context, which they continuously co-construct while it simultaneously shapes their behaviour, expectations, and decisions. Each of our methods, from lab experiments to ethnographic fieldwork, uncovers different insights and has its limitations. With this in mind, our team has recently compared synthetic and real respondents during research activity, reviewed emerging studies and the latest tools, and examined what synthetic users can and cannot tell us.


How Synthetic Users Are Created

Synthetic users are built by AI models trained on existing data. Some rely on statistically likely patterns to generate "typical responses"; others use large language models (LLMs) trained on vast datasets. In practice, these two approaches are often combined. The process begins with data based on observed human behaviour, which is recorded, cleaned, and usually interpreted before being fed into the model. The theory is that LLMs encode general behavior patterns, with the proper prompts and data inputs, and can mimic how real people might respond.

Companies enrich AI with their internal data, like customer records or past research logs, to make these simulations better reflect a particular audience. However, insights from synthetic data are shaped by the training data, not by observing live events in real-time, treading a fine line between simulating and discovering new insights.


The Reality Check on Validation

Validating synthetic users against real behaviour remains challenging. Minor changes to how they are built or trained can produce different results. There are no clear industry standards for testing the quality of the models, and little transparency regarding how commercial versions are developed.

A 2023 analysis by researchers at Columbia and Stanford found that OpenAI models tend to exhibit liberal, educated viewpoints, which are likely influenced by the training data and human feedback used to fine-tune them. Another study by James Bisbee in 2024 showed that synthetic users tend to exhibit less variety. They can be susceptible to the wording of questions (prompts) and may not remain consistent over time. Similarly, our recent side-by-side comparisons revealed significant differences between commercially available synthetic users and real users, particularly in response to experience-based or category-choice questions. They tended to mimic simple rating-scale answers more closely.

A 2023 study from Stanford and DeepMind trained AI 'digital twins' on 1,000 in-depth interview transcripts to "develop" their personalities. When tested using well-established psychological surveys such as the Big Five, the digital twins' responses matched those of their human counterparts 85% of the time, outperforming simpler persona or demographic-based models.

Rather than true "thinking", the twins relied on pattern recognition, drawing on well-known correlations in Big Five personality psychology and interview cues. While they performed well on standardised, research-backed assessments, they struggled to generalise beyond them (e.g., predicting behaviour in strategic games). Yet this is precisely what strategic product and service design demands: insight into shifting realities that rarely fit standardised frameworks.


Understand the Trade-Offs


Bias in, bias out

A primary concern is that synthetic users are trained on existing data. If that data is skewed or incomplete, synthetic models risk recycling and amplifying it.


Missing Outliers

Synthetic users reflect statistically likely "average" behaviour. However, in user research, breakthroughs often come from understanding the outliers—users with uncommon needs or edge-case frustrations that expose hidden issues.

Think of Netflix: Their game-changing decision to drop late fees (and eventually pivot to streaming) was sparked by listening to a handful of frustrated users, not a broad trend. Synthetic users often miss these margins where innovation might thrive.


Plausible Answers, False Assumptions

Suppose a factor relevant to your research was not included in the AI’s training data, whether because it was overlooked, unmeasured, or not yet recognised. In that case, it will remain invisible to a synthetic user. They cannot anticipate behaviour in new contexts or identify emerging needs. The catch? They may still produce plausible answers but lead to misguided design decisions.

In a simulated pricing test, real users reacted directly to a price. By contrast, a synthetic user inferred meaning (e.g., a higher price equals better quality) because this association appeared in its training data. Even if a key variable isn’t present, they may still generate confident responses; if a pattern is typical, they might wrongly apply it to new situations.


No Contradiction and Emotional Depth

Synthetic users lack connection to the broader social and cultural context real people shape and are shaped by, a relational dynamic inaccessible to models trained solely on data. Humans change their minds, contradict themselves, and find it hard to explain why. Synthetic users don’t experience emotion, hesitation, or cognitive load, nor can they reflect tonal shifts. While this 'smoothness' and ‘sycophantic’ responses lacking critical nuance may seem appealing, contradictions often provide the most valuable insights. Unlike humans, they don't have lived experiences such as tensing up when an interface recalls a past mistake. In our comparative study, participants linked voice interface frustrations to past failures, a link absent in synthetic responses.


The Validation Gap

When a synthetic user supports your assumptions, how do you validate them? In user research, we triangulate findings using diverse methods and follow-up questions. We are looking for gaps between what people say and what they do, as behaviour often diverges from self-reported intentions and values. You get a narrative with synthetic users, but there's no actual behaviour to observe, limiting your ability to verify the "insight."


Missing the bigger picture: why relationships are key

Our study explored whether real-world habits (like using meditation or language apps outside cars) would translate into interest in new settings, precisely, in-car functionalities.

Real users revealed strong connections: established habits and recent relevant experiences in cars (like needing urgent information) boosted interest in related new car functionalities. However, not all strong habits carry over to a car—context matters—and understanding this nuance helps to pinpoint what will and won't work.

Synthetic users showed none of these patterns. Their answers lacked internal coherence—correlations were flat or missing as if different people answered each question (Figure 1).

Without these relational links, it's hard to answer key strategic questions like "Who is this for?" or "What drives their interest?" While simulations might mirror general trends, they miss the coherence crucial for segmentation and targeting. Habits, expectations, and context and their interconnections shape actual user behavior—connections that synthetic data can't yet replicate.


How to Use Synthetic Users

Don't replace—augment. Use synthetic users to complement research. Use them to generate initial hypotheses and prepare more effectively for fieldwork and studies with actual end users.

Beware of the efficiency illusion. Creating a reliable synthetic user set still requires a solid base of actual user data. Use triangulation to identify contradictions, gaps, and creative workarounds that synthetic insights will likely miss. Consider the efficiencies you are gaining and where you are adding necessary (or extra) steps.

Beware of black-box solutions. Ask how and by whom the synthetic users were generated, what datasets the model was trained on, the modeling techniques used, and the tool’s capabilities and limitations. Assess credibility by examining its validation processes and any third-party peer reviews.


Conclusion

While synthetic users offer speed and scalability, they are not yet capable of fully replicating the complexity of human behavior, including emotional and strategic responses shaped by evolving context.

Synthetic user platforms show promise for early-stage tasks such as exploring trends or generating hypotheses, but their value depends on the quality of real-world data. Many of these platforms are still in the experimental stage. As researchers, we assess where and when they can be integrated into our research toolkit. Considering business goals and risks, we evaluate what they can deliver — and, equally importantly, what they cannot.

Without validation and scrutiny, synthetic users may overlook game-changing insights, leading to costly strategic missteps. In high-stakes situations like new domains, critical user flows, or early-stage innovation, insights based on real human experience remain invaluable.

Need help setting up your research activities? Feel free to reach out.




Go further: Download our synthetic data research checklist

This checklist guides the critical evaluation of models that simulate user behavior or generate insights. It emphasizes the importance of high-quality, inclusive data, sound modeling, internal consistency and transparency.











Can AI-powered synthetic users replace research with real humans?

Viktor Šmidák

Dr. Max has the largest network of stores in Slovakia, but it is also known for bringing innovations. How do they manage the satisfaction of 1.5 million customers, how do they use data and what are their plans for the future?

Today's guest on the Minimum Viable Podcast is Jana Kutlíková, who is in charge of marketing and customer experience at Dr. Max.


Do you know your organization's CX maturity level? Find out in few seconds by answering these 6 questions.




Minimum Viable Podcast is brought to you by the CX design studio Lighting Beetle*. It's usually recorded in the Slovak language, however, you can read an English transcription of this episode below. For more English content, including case studies and handouts, visit our Journal.

To start, a little bit broadly, how big is Dr. Max and how many customers does it serve?

So overall Dr. Max consists of online and offline. Within offline, it's our pharmacies. Currently we have 421 and we can say that we are in every district, which we are very proud of. Then within online we have of course coverage within the whole of Slovakia in terms of delivery. In terms of the number of pharmacies, we are the largest in Slovakia.

How many customers do you serve?

We can measure this by the number of receipts. And then of course we measure how many customers we have in our loyalty program. An active customer for us is a customer who buys within twenty-four months and currently we are above the number of one and a half million. Customer care is a motto for us, we try to be number one in care and always bring something more to our patients and customers.

I hear a little bit in that that one of your goals is to cover the whole country, to have dominance in the market.

Yes, it is a long-term goal. Of course, the competition is still there and is big, but for us it's still like bringing something extra. For us it's availability, of course price and always bringing something extra in care. Currently, the pharmacy is the most accessible place of healthcare.

So when I'm sick and I have a runny nose, where do I go, I go to the pharmacy, only then do I visit the doctor. So this is how we approach it, that we are like the first entry point in case of illness.

What types of customers do you perceive, how do you categorize them?

We categorize online customers on the one hand. Secondly, there are our customers, whom we divide into 2 largest segments, namely prescription customers, then there are ordinary passers-by and last but not least, those who are dealing with some non-acute problem - they need nose drops, vitamins and so on. So we also have our pharmacies divided. We have pharmacies that you see on the street, then there are shopping pharmacies and then there are pharmacies that are in hospitals or clinics, so there is also a different group of patients there.

Do you also divide these customers socio-demographically, that you have them precisely named, that this is some kind of profile of his? Or are they rather shopping modes, that these people simply move from one mode to another?

When we talk about pure age, pensioners are more sensitive to the price. So we try to adapt the offer to them. But as for over-the-counter customers, we of course sort them into customer groups based on their shopping behavior, value and basket composition.

When the NPS is below 70, we really actively work with it

How do you know how to put together the product portfolio at a specific store?

So those outpatient pharmacies are more prescribed, there their stock availability is also adapted based on what doctors are in the given hospital. But on the other hand, when our pharmacies are in the shopping mall, we try to have a wider offer there and there, for example, we also place dermocosmetics and there the portfolio is larger. So we monitor the purchases of each pharmacy and we are also enabled by such a system that you can reserve medicines online with us.

This means that the pharmacist will issue it to you within an hour. There you can see which pharmacy has it and if there is a greater interest in some medicine, we will pre-stock this pharmacy.

What data about customers do you process? How do you know they are satisfied? What are your pillars based on which you make action plans?

As part of our customers, we have the Dr. Max Club. Whoever is a member of the club, of course, receives benefits and we "interview" these members after each purchase through a satisfaction questionnaire. Based on this purchase, we ask him about the NPS, whether he was satisfied with the purchase and whether he would actually recommend this pharmacy to his friends. And then there are things like whether he was satisfied with the pharmacist's approach, whether the pharmacy had the product that she needed and so on and so forth. So we have it divided into larger sections.

I personally don't think that a pharmacy, for example, is the place where we would now place tablets. The goal is, of course, to convert as many patients as possible who are not members to become members so that we have data about them. Then we subsequently work more with the phase of those "detractors". We then communicate with those detectors through our call center.

What else can you do with the data other than calling detractors?

Each pharmacy has its own NPS and we are dedicated to improving it. When there is long-term dissatisfaction in some pharmacy, then of course we work with the regional director to continue working with that pharmacy and it is not just some kind of fluctuation. When the NPS is below 70, we really actively work with it to make it as good as possible. We talked a lot about the data and I'm interested in how important this data is for Dr. Max's top management. It's definitely the topic of the week and I can say that we have it in our annual goals.

We have a very nice example from the beginning of last year, when we were changing our loyalty club, while canceling the previous loyalty points. And by forcing those patients to apply points, it happened to us that we sent those x new customers to the pharmacy to apply those points so that they would not lose them. There, such a thing happened that those customers who shopped online were suddenly dissatisfied with the approach of that pharmacy. At that time, the NPS of pharmacies dropped and it turned out that it was because of online customers who were not used to waiting in line, they had to talk to the pharmacist...

Dr. Max Club has 1.5 million customers. What is the mechanics of how you got so many customers into the loyalty program?

Members have better prices, especially for medicines. Another benefit is our private label, where there are 2+1 promotions only with a loyalty card. They can also participate in competitions. However, the main motivator is the price of the product. Being a member brings benefits and based on the feedback data we can offer customers a tailor-made offer.

For example, if someone regularly buys vitamin C, we will send them information about the discount. The classics are birthday and name day discounts. At the end of the year, we send a summary email about the amount saved, which evokes a good feeling in people.

Is the main primary benefit for you customer data or do you track something else as well?

We definitely monitor commercial parameters. Then definitely also within the measurement of market share. I'm not saying completely, but we can look at it this way. You are making quite significant innovations at Dr. Max, at least that's how it seems to me. I will mention the new Super Dr. Max concept, you do sampling directly at the pharmacy, you have robotization...

Why are you actually doing these innovations?

To put it simply, these things you mentioned give us something extra for patients and secondly, time to devote to them. For example, the robotization of expedition work in our pharmacies. Our ROWA robot saves the pharmacist a trip to and from the warehouse, so he can take care of the customer. Secondly, there are examinations, again something extra.

When you go to the pharmacy, you can measure, for example, CRP and only then go to the doctor if the CRP is elevated. The service we would like to have is vaccination. It will also relieve doctors, because it is something we can do at those pharmacies. So when you go to pick up the vaccine, you can get vaccinated right away with us.

Is this more a matter of increasing revenue for you or is it more a matter of customer care?

We don't charge much for the services. It's really like offering the patient something extra.

We talked about these innovations, how do these innovations help you win in the competitive market?

In loyalty. We know how long the waiting times are for those doctors and when I decide that I'm going to solve some problem, I can solve it in one place and thus build loyalty. And on the other hand, we then have data about the customer, so we can always offer him something extra.

What are the responses to these extra services?

We managed to do the CRP, but the other tests still need time. It is a long-term journey and strategy that the pharmacy provides something extra.

I would like to go back to Super Dr. Max, because it is quite an exceptional concept. Try to tell us what it actually is.

Super Dr. Max is our pilot project of the pharmacy of the future. On an area of 715 square meters, we offer more than 6,000 products from various categories - in addition to medicines and nutritional supplements, there is dermocosmetics, children's products, healthy food, products for animals, electronics, sports aids and much, much more. It was also a test of what the customer and patient are willing to buy in that pharmacy, what wide portfolio can they go to.

We are still trying, rotating the products, looking at the customer behavior. There are consultation zones, we offer body composition analysis with the InBody device. Furthermore, what is there is an additional service of dermo-advice, where we do an analysis of the skin and hair. So it's a comprehensive care of my well-being, that you not only sell me a product, but also give me some analysis of how I'm doing.

What are the first reactions from customers? How do you evaluate Super Dr. Max internally now?

For us, it was a test of the assortment, what we learn and what we don't learn. And I can still say that we are still trying. It was quite a challenge for us to grasp and sell it in terms of marketing, that we have such a thing at all, because customers in Slovakia are not used to it at all, which is the opposite abroad. In the future, we will then open something between a classic pharmacy and Super Dr. Max.

What is the primary mode of purchase? Who is the e-shop for?

For everyone. Of course. But what works very well for us and I can say that it is growing are precisely those reservations at pharmacies, where it is a classic omni channel approach. This means I order online and have it in the pharmacy within an hour. Furthermore, we launched Marketplace in May of last year. There we are still learning what is and what is not, how to do it. We also offer reservations for prescription drugs.

Are you also thinking about expanding services within digital?

We have the Dr. Max consultation, where you can ask the pharmacist a question. It's not instant, but a real pharmacist answers you. Even our articles, which are all on the web, are written by experts. We would never let some copywriter or AI write it.

Do you plan to continue to perhaps bring the pharmacist closer, have a faster response, bring the online space closer to the offline one?

My personal goal is to have a pharmacy in the Metaverse. Today, the legislation will not let go, but definitely yes. We also try to involve pharmacists on maternity leave, and this is not possible in the offline space. So this is how we try to motivate the staff to stay with us, even when they go on maternity leave. We would definitely prefer to have it online 24/7, but currently, in terms of capacity, this is the maximum we can give to patients.

What is your biggest challenge that you would like to solve somehow? We call it a design challenge, i.e. a complex problem that stands before you and you would like to get through it.

As I mentioned, to have a pharmacy in the Metaverse. So for that you need a dose of better legislation. If the legislation wasn't so discouraging, that one could really get carried away and make some kind of pharmacy of the millennium, where everything would be grasped in a different way, that would be very nice in terms of what you nicely called design character and so that our pharmacies really offer something extra, so that patients go there for those services.

So that we are the first choice, that I don't just go for some of my vitamins and for some of my prescriptions, but that I go there because I want advice. I want care and it's also a more pleasant environment than in the doctor's waiting room. Raise the awareness of the pharmacy to a higher level as well. It's more than just a pharmacy, but it's somewhere where that comprehensive care is provided for our patients.

Minimum Viable Podcast is brought to you every two months by the design studio Lighting Beetle*, which focuses on creating an exceptional customer experience.

In Minimum Viable Podcast, we use relatable language to dive into customer experience design. We invite unique guests to chat about how Customer Experience influences business success. A special thanks to Mojmír Procházka for his production support.

We are looking forward to every listen, follow, share, and suggestion for improvement. You can send us your tips for interesting personalities with whom we can talk about design to podcast@lbstudio.sk.

Enjoy!

Minimum Viable Podcast #29: CX in pharmacy

Dominika Babulicová

A Service Blueprint is a visual tool that captures all aspects of a service – from customer interactions to behind-the-scenes processes carried out by employees, as well as the tools and systems they use. This canvas is invaluable for planning, optimizing and implementing services, as it provides a clear overview of how the service operates, where bottlenecks occur, and how to improve both customer and employee experiences. We highly recommend printing it out as a A1 poster or pasting it into your favorite digital board as Miro, FigJam, Mural etc.


What is this tool used for?

The Service Blueprint template helps you answer key questions:

 * Where do customers interact with the service?
 * What internal processes are needed to deliver it?
 * Where do problems, errors, or inefficiencies arise?
 * What technologies and tools support the service?


Why is a Service Blueprint useful?

Most companies today focus on improving customer experience (CX), but many overlook the fact that backstage processes have a significant impact on the final outcome. To deliver high-quality services, you need to understand how all parts of the service – from employees to technology to processes – work together.

A Service Blueprint enables you to:

 * Gain a clear overview of the entire service: Map out each phase of service delivery and every interaction to identify potential issues.
 * Optimize processes: Identify inefficiencies and weak points that could cause delays, reduce quality, increase costs, and more.
 * Improve the employee experience: To deliver excellent customer experiences, employees must be equipped with the right tools and processes. Understanding the full service process helps them prepare for different scenarios.
 * Enhance customer experience: By optimizing services and processes, you can ensure a smooth and seamless customer journey.
 * Streamline cross-team communication: A Service Blueprint makes it easier for different departments to collaborate, as all processes are mapped in one place.


How to use a Service Blueprint

Creating a service diagram involves several steps that help you thoroughly examine and map the service. Here’s a recommended process to follow:

 1. Map the customer journey
    Start by mapping how the customer moves through the service from beginning to end. Record every touchpoint – from visiting the website to in-person interactions with staff.
 2. Add internal processes
    Based on the customer journey, add the internal processes required to deliver the service. These may include employee actions, technologies, systems, or tools that support the service behind the scenes.
 3. Analyze and optimize
    Once all processes and interactions are mapped, analyze the blueprint to identify bottlenecks, inefficiencies, or potential issues. Adjust and improve the processes where necessary to make the service more efficient and seamless.









Get the Service Blueprint canvas


Your full name*





Your company email*





Which one best describes you?*




Freelancer (I work as a freelancer or an agency member)


Product team member (I build or market products and services)


Product owner (I manage a product or startup lifecycle)


Corporate leader (I manage a division in a bigger company)


Academic (I am a student or a teacher)





What language do you prefer to read our emails in?*




English


Slovenčina





Handout type*

Validation handbook


Experiment toolbox


Service blueprint EN







By clicking on "Download" you agree to the processing of personal data by Lighting Beetle s.r.o. as explained in our
Personal data protection policy.







Newsletter opt-in






I'd also like to subscribe for the occasional CX Insights newsletter by LB*





Download












Customer experience
that drives your business

Our services create true business value

Service
Design

Digital
Product Design

CX Transformation

We are Lighting Beetle

Journal

Can AI-powered synthetic users replace research with real humans?

Minimum Viable Podcast #29: CX in pharmacy

Service Blueprint

Customer experiencethat drives your business

Our services create true business value

ServiceDesign

DigitalProduct Design

CX Transformation

We are Lighting Beetle﻿

Journal﻿

Can AI-powered synthetic users replace research with real humans?

Minimum Viable Podcast #29: CX in pharmacy

Service Blueprint

Customer experience
that drives your business

Service
Design

Digital
Product Design

We are Lighting Beetle

Journal