Changelog
What’s new in Voicepanel
Highlights of what we’ve shipped in Voicepanel, month by month.
September 2026
LatestOne library for everything you test
Assets now live in one reusable library with their own attributes, your AI tools can build highlight reels for you, and Grade opens for early access.
Asset library
Every image, video, prototype, and page you upload now lives in one library. Reuse an asset in any study without uploading it again. Your team works from one set of assets instead of a separate copy in every study.
Questions that adapt to each asset
Refer to an asset’s attributes in question text, like “Would you pay {{asset.metadata.price}} for this?”, and each respondent sees the value for the version they were shown. One question covers every version, so you don’t need a separate section for each asset.
Typing tools in the study editor
Attach your segmentation typing tool to a study from the editor. Respondents are assigned to a segment as they answer, ready for quotas, branching, and analysis. You set it up yourself, in the same place you build the rest of the study.
Grade, in early access
Grade scores your AI product experience across hundreds of user scenarios, with a rubric calibrated to what real users say and do. You find out where the experience falls short before your users do. We’re onboarding teams in waves, so join the waitlist to get access.
Highlight reels from your AI assistant
Ask Claude or ChatGPT to find the strongest quotes on a topic and turn them into a highlight reel, or pull a study’s complete results into the chat for your own analysis. You never have to leave the conversation to hunt for clips yourself.
Scoped, filtered exports
Choose which responses and which data to include, apply your filters, and download transcripts and analysis together as a single zip file.
Segment comparison on every plan
Comparing segments side by side, question by question, is now included on every plan.
Reuse screeners that worked
Search the screeners from your past studies and drop a proven one into a new study.
Mobile tasks that stay out of the way
On phones, task instructions float in a card over the site or app being tested. Respondents can tuck it into a corner tab and bring it back when they need it.
Labels above side-by-side options
In side-by-side comparisons, each option’s label now sits above it, so respondents can tell the options apart before they scroll.
August 2026
Fully programmable, built for evaluations
The Voicepanel API and official Claude and ChatGPT connectors open the platform to code and AI, and side-by-side evaluations bring rigorous comparisons to product teams.
The Voicepanel API
Everything in the Voicepanel app is now available through a REST API. Create and launch studies from code, sync results to your research CRM, or stream responses into your data warehouse. Studies you create through the API show up in the app like any other, so the whole team can follow along.
Official Claude and ChatGPT connectors
Install Voicepanel from the Claude and ChatGPT connector directories. Design studies with the full context of your conversation, then turn raw results into branded decks and reports. Anyone on your team can use them, not just the people who write code.
Side-by-side evaluations
Show respondents two versions, like two AI answers or two sets of search results, and ask which is better. Blind randomization and stable IDs tie every judgment to exactly what was shown, so results can calibrate rubrics or train LLM judges.
Test any experience you can build
Upload a self-contained HTML page, such as an interactive demo or a rendered AI output, and respondents use it just like a live site. It’s a fast way to test an idea before anyone builds it for real.
Exclusive answer choices
Mark options like “None of the above” as exclusive, so a respondent can’t select them alongside other answers.
Study plans as PDFs
Export a study plan and its screener as a PDF, with answer options and randomization shown, for legal, brand, or stakeholder sign-off.
Custom code frames
Define the themes you want open-ended answers sorted into, and AI applies them to every response. Edit or delete themes as your thinking changes. Your results line up with the categories your team already reports on.
Compare results by asset attributes
Tag assets with attributes such as price tier or model version, then group and compare results by those attributes instead of one asset at a time. Patterns across many assets become easy to spot, such as whether lower-priced options score better.
Analysis-ready CSV exports
One combined CSV covers every distribution in a study, with transcripts, analysis answers, and consistent column labels. It opens cleanly in Excel.
Recordings that survive a bad connection
If a respondent’s camera or microphone stream fails mid-answer, recording restarts on its own instead of losing the response.
Three days to finish
Respondents can now pause and come back to finish a study up to three days later, instead of three hours.
Answers that are never lost
If a save fails or a concept doesn’t load, Voicepanel keeps the respondent’s answers and transcript, and the study carries on.
July 2026
Choose who gets into your study
A new Responses page and AI-evaluated screener answers give you full control over who takes part, and variables make every study more personal.
Your entire recruiting funnel
The new Responses page shows every applicant by stage: applied, screened in, invited, and responded. Compare screener answers in columns, filter to the people you want, and invite them one at a time or in bulk. You can see at a glance where recruiting is slow and where you have more applicants than you need.
Open-ended screener questions
Ask applicants to explain themselves in the screener, and describe in plain English what a good answer looks like. AI checks every answer against your criteria and factors it into each applicant’s match. This catches people who fit on paper but can’t speak to the topic.
Auto-invite, or hand-pick
Each distribution has an auto-invite switch. Leave it on to fill the study fast, or turn it off to review each applicant on the Responses page first.
Screeners from a single sentence
Describe your audience in a sentence, and Voicepanel writes a screener for link and intercept distributions, just as it already did for panels.
Prolific for every account
Prolific’s panel now sits alongside User Interviews and Respondent in the panel picker. Pick whichever panel best fits your audience and budget.
Variables
Insert earlier answers into later questions, like “Why did you choose {{1.1}}?”, or pass in values from the link a respondent arrived on. Variables work in questions, prototype links, and redirects, and the editor autocompletes them. Studies feel more like a real conversation, because the questions build on what each person already said.
“Not an issue”
When the AI study reviewer flags something you did on purpose, mark it as not an issue, and it stays dismissed on future reviews.
Segment and quality filters
Filter responses and analysis by segment or by response quality, and your exports follow the same filters.
Compare groups side by side
Closed-choice charts and open-ended themes can now show segments next to each other, so differences between groups stand out.
Quality scores in exports
Every exported response now carries a quality score, with the reason spelled out for low ratings, so you can set aside weak data in your own tools.
Upload media from your AI assistant
Your AI tool can now upload images and videos straight into a study, and it hands back shareable respondent and dashboard links when it’s done.
Your brand color behind every question
Choose the background color respondents see while they answer, to match your brand.
Images that fit any screen
Choose whether an image fills the screen or fits inside it, and it stays centered on phones, tablets, and desktops.
June 2026
Recruiting gets smarter
Voicepanel can now recommend where to recruit, sort respondents into your own segments, and keep audio clean, while results update the moment responses land.
Bring your own segmentation
Give Voicepanel your typing tool, and it assigns every respondent to your segments the same way your team does. Use those segments in recruiting quotas, branching, analysis, and reports. Your Voicepanel results then line up with the rest of your research and reporting.
Instant panel recommendations
Describe your audience in plain English, and Voicepanel recommends the best panel, drafts a screener, and estimates feasibility before you launch. You find out if an audience is hard to reach before you commit budget to it.
Hand-pick your participants
For User Interviews and Respondent, review applicants yourself and choose exactly who gets invited.
Matrix screener questions
Screen with grid questions, like how often someone uses each of five apps, in single- or multi-select form, with randomized options to reduce order bias.
AI screener review
Before launch, AI reads your screener the way a recruiter would and flags questions that are leading, unclear, or likely to let the wrong people through.
Audio quality monitoring
Voicepanel listens for background voices, noise, and recording problems while people answer. If it hears one, the respondent gets a prompt to fix it or record again. You get cleaner recordings and fewer responses to throw away.
Interviews that don’t stall
If a network request hangs, the study now recovers within seconds instead of leaving the respondent waiting.
Camera and microphone, in any order
Recording now works no matter which order respondents grant camera and microphone access in.
Mobile app in 37 languages
The mobile screen capture app now appears in 37 languages, to match the rest of the respondent experience.
Real-time results
Charts, themes, and exports refresh as each response comes in, so you can check fieldwork mid-flight. Create as many highlight reels and reports as you need, whenever you need them.
Branch on task completion
Branching now works on task results too. Ask people who couldn’t finish a task what got in the way, and skip those questions for everyone else.
Desktop and mobile preview
Switch the preview between desktop and mobile to check how each step looks on a phone before you launch.
May 2026
Voicepanel beyond the browser
Voicepanel now runs inside your AI tools and on respondents’ phones, and concept testing gets its own purpose-built workflow.
Voicepanel MCP
Connect ChatGPT, Claude, Cursor, and other AI tools to Voicepanel. Create and launch studies, recruit participants, and pull results from a chat, with the same permissions you have in the app. It brings Voicepanel into the tools where your team already works.
Labels, search, and a study switcher
Tag studies with labels, search your dashboard, and jump from one study to another without going back to the list.
Mobile screen capture (beta)
Respondents scan a QR code, complete a task in an app or site on their iPhone, and the screen recording uploads to Voicepanel on its own. You see how people use real mobile apps, not just websites in a mobile browser.
Questions read aloud
Voicepanel can read each question aloud in a voice you choose, and it stops as soon as the respondent starts to answer. Studies feel more like a conversation, and they’re easier for people who prefer to listen.
Redesigned respondent screens
The screens respondents see when a study is finished, full, paused, or not supported on their device got a redesign, and you can put your brand on them.
Multi-line typed answers
Respondents who prefer to type can press Shift+Enter to start a new line, for longer, better-organized answers.
Concept comparison
Upload a set of concepts and choose how many each person sees. Voicepanel randomizes and balances exposure, then compares the concepts side by side in analysis. You get a fair comparison without building the randomization yourself.
Study templates
Enterprise teams can save a study as a template and start new studies from it, with the screener and settings already in place. Teams that run the same kind of study often get consistent results without starting from scratch.
Translations you can review
Enterprise teams can check and refine each translation of a study before it goes out to respondents.
Branch on segment
Branching logic can now send each segment through different sections, like asking only power users about advanced features.
Response grid
See every respondent’s answers to every question in one scrollable grid, one row per person. Use it to follow one person’s story from start to finish.
Custom brand terms
Add product names, acronyms, and jargon to a study, and transcription spells them correctly.
NPS and stacked charts
Use a 0–10 NPS scale in rating questions, and view any rating scale as a stacked chart.
April 2026
New research methods
Five-second tests and loops bring new methods into Voicepanel, and sentiment and sharing make results easier to act on.
Five-second tests
Show an image for a few seconds, then ask what people remember and how it made them feel. Built for first impressions of ads, landing pages, and packaging. Results show what people take away before they have time to study the design.
Loops
Run the same section once for each concept instead of copying it by hand, and get separate insights for every loop. Create with AI can set up loops for you.
Create with AI reads video
Upload an MP4, GIF, or WebM when you create a study with AI, and it uses frames from the video to write tasks and questions.
Sentiment analysis
Open-ended answers now show whether people felt positive, negative, or neutral, alongside what they said. You can see at a glance which themes people are excited about and which ones frustrate them.
Share results with stakeholders
Share an analysis by email. Recipients land directly on the results, without setting up an account first. Stakeholders see the findings for themselves, not just a summary in a deck.
“Select at least one” screener logic
Qualify people who pick any one of several answers, like anyone who uses at least one of three competitor apps.
Your data never trains AI
Our privacy policy and DPA now state it outright: Voicepanel doesn’t use customer or respondent data to train or improve AI models. That makes security and legal reviews simpler.
March 2026
Bigger study designs
Test more concepts per study without tiring people out, control exactly how long things are shown, and export full sessions as video.
Concept groups
Put several concepts in a group and show each respondent only a few, randomized and balanced, so one study can cover more ideas. Each person stays focused, and every concept still gets enough views to compare.
Timed exposures
Show a concept for an exact number of seconds before moving on, to measure what people notice and remember.
Create a study from your files
Upload a brief, a deck, or images, and Create with AI turns them into a draft study with tasks and questions. Your study starts from the thinking your team already did, not a blank page.
Personalized prototype links
Add variables to prototype and website links, so each respondent opens the version, account, or language meant for them.
Full session video export
Export an entire response as one continuous video, with every section stitched together in order. Share a complete session with anyone who wants to watch it end to end.
From chat to document
Turn an answer from chat with your data into a formatted document you can send to your team.
Add to existing highlight reels
Add new clips to a highlight reel you already built, so it grows as more responses come in.
Labeled rating scale charts
Rating scale charts now label each point on the scale, so a 4 reads as “Agree”, not just a number.
Open-ended data downloads
Download the summary data behind any open-ended question for your own charts and reports.
An honest progress bar
The progress bar and time estimate now account for looped sections, so respondents always know how much is left.
No timeouts on long studies
Respondents who take their time on a long study now stay signed in until they finish.
February 2026
More to test, in more places
Written concepts and audio join the list of things you can test, respondents gain more languages, and quotas get more precise.
Text concepts
Show copy, a value proposition, or a written concept as its own step, with no mockup required. Test messaging and positioning early, before design work starts.
Internal concept names
Give each concept a name for your team, like “Concept B, blue packaging”. Results use that name, and respondents never see it.
More respondent languages
Respondents can take studies in more languages and regional variants, so one study can run in more markets. People answer in the language they’re most comfortable in, which leads to richer answers.
A better audio player
Respondents hear audio clips in a compact player that can autoplay, move on when the clip ends, and replay inline.
Rating scales stay in order
Rating scale options always appear in their natural order, even when other answer options are shuffled.
Quota-aware invites
Voicepanel checks your subquotas before inviting anyone from User Interviews or Respondent, and only invites people who fit a slot that’s still open. Quotas fill in the right proportions, without extra invites to people you can’t use.
Collect emails from link respondents
Ask people who join through your own link for an email address, so you can reach them for follow-ups.
January 2026
Analysis, rebuilt from the ground up
Results now start with a study-wide summary and section-by-section findings, and you can shape how the AI analyzes your data.
All-new analysis
Every study now opens with a study-wide summary, followed by findings for each section. Tasks and conversations feed into the findings, not just survey questions. You get the big picture first, then the detail behind it.
Analysis that follows your rules
Add your own rules for how the AI writes response summaries, names open-ended categories, and sorts answers into them. Results come back in your team’s terms.
Analyze in the original language
Run analysis in the language the study used, so quotes and themes keep their original wording instead of going through English. Nuance that would get lost in translation stays in your results.
See what sets each segment apart
Segments can now come from multiple-choice answers, and analysis shows where each segment differs from the rest.
Chat with your data, upgraded
Keep several chats about a study, reset one to start fresh, rate the answers, and get better responses from a newer model.
Sign-off before launch
Admins can review and approve a study before it’s published, and Voicepanel saves every published version so you can see what changed. Nothing reaches respondents until the right person has looked at it.
Earlier screen-outs
Respondents who don’t qualify now exit at the first question that rules them out, instead of finishing the whole screener.
Better on-the-fly translations
Questions translated in real time now read more naturally in the respondent’s language.
December 2025
Ready for enterprise teams
Single sign-on, GDPR tools, and budget controls let larger organizations roll out Voicepanel with confidence, while recruiting gets faster and more targeted.
Single sign-on (SSO)
Your team signs in through your company’s identity provider, so access follows your existing IT policies. IT manages who has access in one place, and your team has one less password to remember.
Built for GDPR
Personal details in responses are masked automatically, and our DPA and subprocessor list are public, which shortens privacy reviews.
Budget guardrails
Admins can set a maximum cost per study, so no single study spends more than planned.
Device targeting
Require desktop or mobile respondents, so a mobile app test only reaches people on a phone.
Automatic invites
Panel respondents who pass your screener are invited right away, so fieldwork keeps moving overnight.
Reuse recruiting setups
Copy a distribution, with its screener and targeting, into another study instead of rebuilding it.
Branching logic
Show or skip whole sections based on how someone answered earlier, like sending current customers and prospects down different paths. Each person answers only the questions that apply to them, which keeps studies shorter.
Consent built into every study
Respondents see a clear consent step before a study starts, including a notice when audio or video will be recorded.
Sharper live transcription
A new real-time transcription engine is faster and more accurate, so the AI moderator’s follow-up questions respond to what people actually said.
Upload progress for respondents
Respondents see a progress bar while their recordings upload, so they don’t close the tab before their answers are saved.
Download highlight reels
Export any highlight reel as a video file, subtitles included, for a deck, a Slack thread, or an all-hands.
November 2025
Research that crosses languages
Interviews in other languages now come back in English, and transcripts and highlight reels become much easier to read and share.
Automatic English translation
Transcripts, screener answers, and highlight reel subtitles from non-English interviews are translated into English automatically. Your whole team can read and quote interviews from any market, whether or not they speak the language.
A highlight reel for every finding
Each finding now gets its own highlight reel of the respondents behind it, with redesigned quote slides in Voicepanel colors. Every finding comes with the voices that back it up.
Full-screen transcripts
Transcripts open full screen, with the video alongside and the respondent’s details pinned at the top.
Custom AI rules
Add your own rules, like tone of voice or topics to avoid, and every study the AI drafts or edits follows them. Drafts start closer to your team’s standards, so you spend less time rewriting.
Preview matches the live study
Previews now run the exact same experience respondents get, so what you check is what they see.
Minimize on mobile
Respondents on a phone can minimize the interview panel to see the full site or prototype they’re testing.
Clearer controls for respondents
Buttons now show respondents exactly where to tap to reveal the next part of a question.
A roomier workspace
A collapsible sidebar, clearer tab names, and new outline icons give your studies more space on screen.
October 2025
AI writes your screeners
Screeners now write themselves, highlight reels become editable, and respondents take studies in their own language from the first question.
Screeners written by AI
Describe who you want to reach, and AI writes the screener questions and qualifying answers. Ask it for changes in plain English. New researchers get a solid starting point, and experienced ones save time.
Editable highlight reels
Trim clips and adjust where each quote starts and ends. Reels now also appear next to the insight they support. You control exactly which moments your stakeholders see.
Segment on study answers
Build segments from answers given during the study, not only from the screener, and break results down by them.
Respondents choose their language
Respondents pick a language before the screener starts, so every question, from screener to thank-you screen, appears in that language. People answer more naturally when the whole study speaks their language.
Studies that recover from errors
If something goes wrong mid-study, like a recording that fails to process, respondents move on to the next question instead of getting stuck.
Preview any distribution
Preview the exact experience a respondent gets from a specific distribution, thank-you screen included. Previews never count toward your quota.
September 2025
Highlight reels for everyone
Every account now gets highlight reels, open-ended findings link to the video behind each quote, and you can build your own segments.
Highlight reels
Open-ended findings now come with a highlight reel of the clips behind them, subtitles included. Stakeholders hear straight from respondents, in their own words and voices.
Audio-only highlight reels
Studies recorded without video get highlight reels too, built from audio clips.
Watch the moment behind a quote
Open-ended findings link to the video of each quote, and the transcript scrolls to the exact spot. You can check any claim against what the respondent actually said.
Build your own segments
Define segments from screener answers and demographics in a visual builder, with a catch-all group for everyone else. Use them to filter and compare results across the whole study.
Screener analysis
Screener answers get their own charts and summaries, so you can see who applied, not just who took part.
Typed-only questions
Set a question to accept typed answers only, for answers that are easier to type than to say.
Studies in Thai
Respondents can now take studies in Thai.
Prototypes that keep their place
Figma prototypes no longer reload between steps, so respondents pick up right where they left off.
Scroll through tall images
Respondents can scroll through a tall image, like a full landing page, instead of seeing a cropped view.
Quotas three levels deep
Nest quotas up to three levels, like age within gender within region, to get the exact mix you need.
More countries on User Interviews
Target User Interviews participants in more countries.
August 2025
A new experience for respondents
Every study now runs on a redesigned respondent experience, while quotas by segment, flexible filters, and Word exports make results easier to shape.
A redesigned respondent experience
Every study now runs on a new respondent experience that shows one clear step at a time. It works the same way in study links, panels, and website intercepts. Respondents spend less effort on the interface and more on their answers.
A rebuilt interview engine
Behind the new experience is a rebuilt engine that moves each respondent through the study one defined step at a time. Screeners, tasks, questions, and follow-ups behave the same way in every study. Fewer respondents get stuck partway through.
See yourself while you answer
A small live camera bubble shows respondents how they look on video. On phones, they can drag it out of the way.
Permissions, explained
Before the browser asks for camera, microphone, or screen access, a short screen explains why the study needs it, in the respondent’s language. More people say yes when they know what to expect.
A text box that stays in reach
The box for typed answers stays fixed at the bottom of the screen, so respondents never have to scroll to find it.
A clear message when recording stops
If a recording stops partway through, respondents see a screen that explains what happened.
Faster uploads
Recordings use more efficient video formats, so they upload faster on slow connections.
Quotas by segment
Set a quota for each segment, like 20 current customers and 20 prospects. Voicepanel counts only completed responses and closes each segment when it’s full. You get a balanced sample without watching the numbers yourself.
Filter responses with any rule
Combine filters on demographics, screener answers, and more to focus on exactly the respondents you need. The same filters work while fieldwork runs and in analysis.
Task results with video
Task analysis now shows each respondent’s recording with its start and end times, laid out in a grid that’s easy to scan.
Word document export
Export your analysis insights as a Word document to edit and share.
Screen out incomplete tasks
Mark a task as required, and respondents who don’t complete it are screened out instead of counting toward your study.
Last initials only
Respondent last names now appear as an initial in the app and in exports, which keeps personal data to a minimum.
July 2025
Findings you can check
Every insight now shows how many people back it, chat answers link to their sources, and videos mark where each step begins.
Numbers behind every insight
Each high-level insight now shows how many respondents back it, with the strongest evidence listed first. You can tell a real pattern from a single comment at a glance.
Chat answers with sources
Answers from chat with your data now link to the responses they came from. Follow a link to read the quote in context before you share it.
Step markers in video
The video player marks where each step of the study begins, so you can jump to the part you care about.
Copy quotes in one click
Copy any quote from analysis, ready to paste into a deck or a doc.
Archive old distributions
Archive distributions you no longer use, so your list shows only the ones that matter.
Smarter follow-up questions
The AI moderator is better at knowing when to ask a follow-up and when to move on, so conversations feel less repetitive.
June 2025
Recruit from User Interviews
The User Interviews panel is now built into Voicepanel, and studies gain ranking questions, audio concepts, and randomized sections.
User Interviews panel
Recruit from User Interviews without leaving Voicepanel. Target by skills and occupation, set device requirements, and give participants instructions before they start. Their screener answers come straight into your study, next to everything else they told you.
Ranking questions
Ask respondents to put options in order, and use images or other concepts as the options. You learn what matters most, not just what people like.
Audio concepts
Play an audio clip, like a jingle or a radio ad, as a concept for respondents to react to.
Randomize section order
Shuffle the order of sections for each respondent to reduce order bias. Analysis still lines every answer up with the right section.
Choose what gets recorded
Turn audio, video, and screen recording on or off for your study, so you only ask for what you need.
Segments while fieldwork runs
See each respondent’s segment in the responses table and filter by it while the study is still live.
Transcripts as PDFs
Export any response transcript as a PDF to share or archive.
Average time to complete
The monitor tab shows how long respondents take on average, so you can check your time estimate against reality.
May 2025
Draft a study by chatting with AI
Create with AI turns a short chat into a full study draft, and a redesigned editor makes every study easier to build and adjust.
Create with AI
Describe what you want to learn, add your concepts, and Create with AI drafts a complete study while you watch. It writes the objective, sections, tasks, and questions. Most of the setup is done before you open the editor.
Edit with AI
Ask for changes in plain English, like “add a question about pricing”, and AI updates the study for you. It works on new drafts and on studies you already built.
A clearer study editor
The study editor got a redesign. Each section is easier to scan, each step explains what it does, and settings live in their own window.
Screen recording on any task
Record the respondent’s screen during any task, so you can watch exactly where they clicked and where they hesitated.
Editor role
Invite teammates as editors, so they can build and change studies without full admin access.
No more blank screens
If part of a page fails to load, respondents see a clear message instead of an empty screen.
April 2025
Studies built in sections
Studies now follow a clear structure of sections, and analysis shows each result next to the concept people saw.
Studies built in sections
Build a study as a series of sections, each with its own concept, task, and questions. Drag sections into place and add steps anywhere. The structure of the study now matches the way you plan it.
See the concept next to the results
Analysis and the monitor tab now show the concept each question refers to, so you never have to guess what people were reacting to.
Chat history that stays
Your chat with your data is saved, so you can close the tab and continue the conversation later.
Average ratings
Rating scale questions now show the average score next to the breakdown of answers.
Edit a live distribution
Change some settings on a distribution after it goes live, instead of starting a new one.
Welcome screens in every language
The welcome screen now appears in the respondent’s language, with the line breaks you wrote kept intact.
Conversations that always get going
The AI moderator now asks at least one question in every conversation, so no conversation ends before it starts.
March 2025
AI-moderated conversations
A new conversation block lets the AI moderator run an open discussion, with its own analysis and a quality score on every response.
Conversations
Add a conversation block, and the AI moderator holds an open discussion on your topic, with follow-up questions as it goes. It gets at the why behind an answer, the way a live interviewer would.
Screen sharing on prototype and website tasks
Respondents share their screen while they try a prototype or website, so you see exactly what they did.
Conversation analysis
Conversations get their own analysis, with a summary and every response sorted into themes.
Response quality scores
Every response gets a quality score, so you can spot low-effort answers and leave them out of your results.
Follow-ups in real time
The AI moderator now follows answers as they’re spoken, so follow-up questions come faster.
February 2025
Voicepanel on your website
Website intercepts bring studies to your own visitors, Slack keeps your team posted, and you can ask questions of your transcripts.
Website intercepts
Invite visitors to a study right on your website with a small snippet. Choose which devices to target and how much of your traffic sees the invite. You hear from real users at the moment they use your product.
Chat with your data
Ask a question in plain English and get an answer drawn from every transcript in your study.
Slack notifications
Get a message in Slack each time someone completes your study.
January 2025
Watch the session, not just the answers
A new video player and task analysis show what people did, and view-only access lets more of your team follow along.
A video player built for research
Play responses at different speeds, jump to any moment, and click a line in the transcript to go straight to it.
Task analysis
Tasks get their own analysis, with a summary of how respondents got on with each one.
Screeners on your own links
Add screener questions to the study links you share yourself, not just to panels.
View-only collaborators
Invite teammates who can see studies and results but can’t change them.
Want to see Voicepanel in action?
Get a walkthrough tailored to how your team works.