Frequently asked questions

ChatGPT reads your code and gives generic advice. Maskerade personas actually browse your site like humans, using a real browser to scroll, click, fill forms, and react. They surface visual and interaction friction (hidden buttons, slow loads, popups) that code-level analysis misses, and their feedback is honest and persona-specific instead of generic. You also get a session replay so you can see exactly where each persona dropped off and why.

Paid panels typically start at €5,000 per study, take weeks to recruit, and introduce bias because participants know they're being watched. Maskerade runs in minutes, eliminates recruitment bias, and is infinitely repeatable so you can iterate continuously instead of running an annual study.

Because you are not your customer. You see your site through your own knowledge, context, and assumptions. Maskerade lets you see it through the eyes of a new mother in week 36, a price-sensitive comparison shopper, or your highest-value buyer, all in minutes. You also get findings segmented by persona, so you can prioritize 'fix this for our highest-value buyer first' instead of one undifferentiated to-do list.

One-off audits work, and we won't stop you. But value compounds with continuous use. Every change you ship (new landing page, new ad creative, new checkout flow, new product launch) is a research moment. With AI-assisted development, teams are shipping faster than ever, which means the bottleneck shifts from 'can we build it' to 'is what we built actually better.' Maskerade closes that loop in minutes. The teams getting the most value run a quick test before every meaningful release, and validate every hunch quickly.

Maskerade works fully standalone, and requires no tags, no SDK, no API integration and no development work on your side. Personas browse your URL the same way any visitor would, so you can expect first insights within minutes of firing up Maskerade.

Yes. Maskerade works on any web-accessible URL: production, staging, or acceptance. Authenticated areas are supported via test credentials, and we can simulate user location by IP for geo-gated content.

You can start for free: a new account comes with 25 credits, and no credit card is needed. One credit is one persona running one test, on a website or on an ad, so a first test with five personas uses five credits. After that, the monthly plans are Starter at €250 for 100 credits, Team at €700 for 300 credits and Business at €1,100 for 500 credits. You can also buy extra credits at any time, at €2.50 a credit, from 20 credits. Prices exclude VAT. For larger volumes, talk to us about Enterprise.

A plan's credits renew each month with your invoice, and what is left at the end of the month does not roll over. Your starting credits and any extra credits you buy last 12 months. You can cancel a plan at any time, and it runs until the end of the month you have paid for.

The most common use cases are pre-launch landing page validation (test copy and design before paid traffic hits), competitive benchmarking, checkout and conversion analysis, messaging and tone testing across segments, and localization QA for translations, currency, and regional payment methods. The strength is diagnosis (finding the 'why' behind user behavior) and helping you decide where to focus.

You truly can start from any business question related to your site you have. But, you don't need to design scenarios from scratch if you're out of inspiration. Maskerade ships with 22 pre-built playbooks for the most common starting points, such as a homepage and navigation check, a release check, a support chatbot check, or a journey that starts on Google or ChatGPT, plus playbooks for energy, telecom, e-commerce, media and B2B. Pick a playbook, plug in your URL, and you're testing within a minute. Once you see what comes back, defining your own custom scenarios and follow-up questions becomes much easier.

Maskerade is a multiplier for the CRO people you already have, not a replacement. Automating exploratory research frees the team from 'why did users drop off' cycles and lets them validate hypotheses before allocating engineering or agency time. Particularly valuable when CRO headcount is scarce.

Maskerade is an analysis layer, not an implementation layer. We don't compete with your agency, we feed them prioritized work. Most findings translate directly into more work for the agency to deliver, so it's typically a net positive. We've seen this pattern with several customers running alongside their existing CRO partner, and we recommend looping the agency in early.

Yes, and the way it handles this is different from dumping your site into an LLM (which would just produce a long list of missing things). Because personas actually browse, they only flag what users actually hit and trip over. So even on an immature site, you get a prioritized 'fix this first because it blocks the user' list rather than a generic content audit. If you already know your priority area (e.g. cart flow), you can point the personas there to make it sharper.

Yes. Today you can paste your ad copy as the starting context and the persona arrives at your landing page already primed by that promise, useful for spotting message-match gaps that hurt conversion. Or you can start the journey on e.g. Google and see them interact with your ad before landing on the landing page.

Yes. With Ad Testing, in beta, you upload an ad as an image or a PDF and no website is needed. An ad critique tells you whether one creative will work, check by check: does it stop people, do they get the message, do they recognize the brand, does the call to action work. An ad comparison shows two to four creatives, yours or a competitor's, to every persona. Everyone votes, so you see which version to run and why.

Yes. Scheduled reports, in beta, repeat a completed website report daily, weekly, every two weeks or monthly. Each new report is emailed to the people you choose and opens with what changed since the last one, so you see what's new, what's fixed and what's still open. Each run uses credits like any other.

Maskerade.ai's personas are built on the foundation of 61 recent scientific papers on human behavior simulation, and state-of-the-art agentic browsing technology. Every persona is defined based on dozens of underlying variables that defines their perception, how they choose, and how they browse.

Yes. Beyond consumer personas, you can brief expert personas to audit pages or provide specialized feedback. Useful for brand consistency audits, SEO reviews, or CRO best-practice checks. Custom personas can be briefed on your company guidelines and will flag deviations. If this is a requirement, we strongly encourage you to reach out to us so we can help to set this up!

Absolutely. We rank findings based on how critical they are towards your business goal. In addition, you can interact (chat) with the findings themselves to deep dive into a specific topic that is important to you, or interact with a specific persona and their browsing session.

Personas can be geo-targeted across 150+ countries and will see locally-adapted content (language, currency, right-to-left where applicable). They act in line with local culture, and they will use the website in the local language (and even flag things like translation or other localization issues).

Data is stored in London, UK and processed in Belgium, encrypted in transit and at rest. We do not train models on your data and do not share insights across accounts. Each run is isolated to your account. For sensitive flows you can point Maskerade at your staging or pre-prod environment.

Yes. Personas can create accounts or use existing ones if you share credentials, fill forms, and complete checkout up to the payment step, where they stop before paying. They carry no payment details, so no real transaction can occur. For sensitive flows you can also run on staging or pre-prod.