We've shipped major improvements to how you set up studies and how our AI moderator screens participants. The result: faster study configuration and more reliable participant quality.

AI-Powered Screener Recommendations

When you describe your target audience in the study chat, the AI now generates 6 highly specific screeners tailored to your research topic. Instead of generic demographic filters, you get precise qualification criteria based on your actual study needs.

The recommendations pull context from your entire study conversation—understanding your research objectives, target audience, and methodology to suggest screeners that actually matter for your insights.

Payout Forecaster

See estimated participant payouts before you launch. The new Payout Forecaster shows:

• Estimated hourly rate based on your screener complexity and study duration
• Budget calculations that automatically sync—change your target response count and the budget updates, or adjust your budget and see how many participants you can afford

No more guessing what to pay. The forecaster analyzes your screeners and interview length to suggest competitive rates.

These improvements are live now. Create a new study to experience the streamlined setup flow.

Your interviews, emails, and participant badges can now carry your brand identity. We've completely rebuilt the branding experience to make it effortless to customize—and instant to preview.

One-Click Brand Import

Paste your website URL and we'll automatically fetch your logo and brand colors. No more hunting for hex codes or uploading assets one by one. Your brand assets are pulled from your website and applied across the platform.

Live Previews

See exactly how your branding will appear before you publish:

• Interview preview: A realistic phone mockup showing your logo and background colors as participants will see them
• Email preview: How your logo and brand colors render in participant communications
• Badge preview: The branded chip that identifies your organization in the platform

Toggle between light and dark mode previews to ensure your brand looks great in both.

Flexible Color Controls

Fine-tune every color with our new color picker:

• Pick colors visually or enter exact hex values
• Use the eyedropper to sample colors from anywhere on screen
• Changes save automatically—no save button needed

Set different background colors for light and dark mode interviews to match your brand guidelines.

Automatic Asset Sync

Upload your organization avatar once and it automatically populates your interview avatar, email logo, and badge logo. Override any of these individually when you need different assets for different contexts.

Head to Settings → Branding to customize your organization's look.

We've shipped a collection of platform improvements focused on making Terac faster to navigate, more intuitive to configure, and smarter in how the AI interviewer understands your participants.

Organization-Scoped URLs

Your workspace URLs now include your organization slug. Instead of navigating to /org/my-project/studies, you'll see /acme-research/my-project/studies.

This change brings several benefits:

• Clear context: The URL always tells you which organization you're working in
• Easy sharing: Copy a URL and your teammate lands in the right place
• Multi-org support: Switch between organizations without confusion
• Bookmarkable: Save links to specific projects or studies that just work

All existing links redirect automatically—no action required on your part.

Study Timeline Configuration

You can now set an intended study duration when configuring your study. This helps both you and your participants understand expectations:

• Pacing guidance: The AI interviewer uses this to pace conversations appropriately
• Participant expectations: Display estimated time commitments upfront
• Planning: Better forecast when your study will complete based on recruitment velocity

Set your timeline in hours, from quick 1-hour sprints to month-long longitudinal studies.

Study-Level Media Requirements

Video and screen sharing requirements are now configured at the study level rather than the research guide level. This means:

• Reusable guides: Use the same research guide across multiple studies with different media requirements
• Clearer configuration: All study settings live in one place
• Better defaults: Each study starts with sensible defaults that you can adjust

These improvements are live now. If you have questions or feedback, reach out to our team.

We know that launching a research study can feel high-stakes. You've spent time crafting your research guide, defining your target audience, and setting up your study—but how do you know the interview will flow naturally before your first real participant joins?

That's why we built Demo Interviews. Now you can test any research guide with a single click, experiencing exactly what your participants will see and hear.

How it works

When you're editing a research guide, you'll see a new "Demo" button in the editor toolbar. Clicking it instantly launches a demo interview session where you play the role of a participant. The AI interviewer will ask questions from your guide, respond to your answers, and follow up naturally—just like it would with a real participant.

Key benefits

• Iterate faster: No need to publish your guide or recruit a test participant. Just click and start testing.
• Draft support: Demo mode works with your unpublished draft, so you can test changes before committing to them.
• Share with teammates: Generate preview links to share with colleagues for feedback before launch.
• Clear distinction: A "Demo Mode" badge appears throughout the session, and demo interviews are stored separately from real participant data.

Perfect for

• First-time guide creators who want to see how the AI interviewer interprets their questions
• Researchers iterating on question phrasing or flow
• Teams reviewing research methodology before a study goes live
• Training new team members on how interviews work

This feature removes the friction between writing a guide and understanding how it will perform in practice. We think you'll find it becomes an essential part of your research workflow.

The best research interviews don't follow a script—they follow the conversation. When a participant reveals something unexpected, a skilled researcher knows to dig deeper, ask follow-up questions, and explore where the conversation naturally leads.

We've completely rebuilt our AI interviewer to do exactly that. Instead of rigidly moving through questions in sequence, it now uses reasoning to navigate topics dynamically based on what participants actually say.

What's different

Previously, our system used a state-machine architecture: it would ask Question 1, wait for an answer, move to Question 2, and so on. This worked, but it felt robotic and missed opportunities to explore interesting threads.

The new system gives the AI interviewer the freedom to:

• Follow interesting threads: When a participant mentions something relevant to a later topic, the AI can explore it immediately rather than waiting
• Ask natural follow-ups: Instead of moving on after one answer, it probes deeper with contextual questions
• Return to topics: If something comes up later that relates to an earlier question, the AI can circle back naturally
• Adapt to the participant: Some people give long, detailed answers; others are brief. The AI now adjusts its questioning style accordingly

Screen sharing, reimagined

Screen sharing is no longer tied to specific questions. Participants can share their screen at any point during the interview, and the AI can reference what it sees throughout the conversation. This makes product testing and usability research much more natural.

Data collection without interruption

Need to collect specific data points during the interview? The AI can now gather this information conversationally, weaving questions into the natural flow rather than presenting them as separate form fields that break the rhythm.

Research rigor maintained

We didn't sacrifice structure for flexibility. Your research guide still defines the topics to cover and the depth required. The AI uses intelligent probing to ensure all your research objectives are met—it just does so through natural conversation rather than rigid sequencing.

This update represents a fundamental shift in how AI-moderated interviews work. We think you'll notice the difference in participant engagement and the richness of insights you collect.

Great research depends on great participants. Finding the right people, inviting them efficiently, and tracking who's completed what—this operational work is often the most time-consuming part of running studies.

We've rebuilt our panel and invitation management system to make this work faster and more transparent.

Redesigned invitation system

Every participant invitation is now tied to a specific recruitment run. This means you can see exactly where each participant came from, when they were invited, and how they progressed through your study.

The new invitation flow:

1. Source participants: Run a search to find people matching your criteria
2. Review and select: Browse profiles and select who to invite
3. Track automatically: The system tracks invitation → acceptance → scheduling → completion

You'll never lose track of where a participant came from or why they were selected.

Interactive map view

For studies where geography matters—local services, regional products, market-specific research—you can now visualize your panel on an interactive map.

See at a glance where your participants are located, identify coverage gaps, and ensure you're getting diverse geographic representation.

Richer participant profiles

Participant profiles now include more demographic and qualification data. When reviewing potential participants, you can see:

• Basic demographics (where provided)
• Qualification answers from screeners
• Past study participation and performance
• Communication preferences

This context helps you make better decisions about who to invite and how to approach them.

Internal dashboard

A new internal dashboard gives you a bird's-eye view of study health:

• How many invitations are pending vs. accepted
• Conversion rates at each stage of the funnel
• Participant flow over time
• Bottlenecks that need attention

This is particularly valuable for large studies where you need to monitor progress and adjust recruitment strategy on the fly.

Streamlined workflows

We've reduced the number of clicks required to go from recruitments to interview. Common actions like "invite selected participants" or "send reminder to pending" are now accessible directly from the panel view.

The goal is simple: spend less time on logistics and more time on research.

Research generates insights. But insights scattered across transcripts, notes, and team conversations don't drive action. You need a way to organize, prioritize, and share what you've learned.

We've built a complete reports and insights system to solve this.

Create and share reports

Reports let you organize your research findings into a structured narrative. Each report can include:

• Key insights with supporting evidence
• Direct quotes from participants
• Visualizations and data
• Your analysis and recommendations

Reports are shareable via link, making it easy to distribute findings to stakeholders who don't need access to your full research workspace. You control who can view each report.

Real-time insight generation

When you run insight extraction on your interviews, you can now watch the process in real-time. A progress indicator shows which interviews are being analyzed, and new insights appear as they're generated.

This transparency helps you understand what the AI is finding and lets you course-correct if needed. You can also pause, adjust parameters, and restart if the initial results aren't quite right.

Insights mapped to your guide

Insights are automatically tagged to the relevant questions in your research guide. This makes it easy to see which areas of your research have rich findings and which might need more data.

For example, if you have five questions about onboarding and three about retention, you can quickly see how many insights you've generated for each area. This helps identify gaps and prioritize follow-up research.

Advanced filtering

As your insight library grows, finding specific findings becomes critical. The new filtering system lets you:

• Search across all insights by keyword
• Filter by study, date range, or participant segment
• Sort by relevance, recency, or strength of evidence
• Tag and categorize insights for your own organization

Shareable insight links

Individual insights can be shared via link, complete with the original context and supporting quotes. This is useful for sharing specific findings in Slack, email, or presentations without requiring recipients to navigate the full platform.

The goal of this system is simple: make it easier to go from raw interview data to actionable insights that your team can use.

AI is only as good as the context it has. When you're researching a complex product, exploring a specific market segment, or building on previous research findings, the AI needs that background to ask smart questions and generate relevant insights.

Project and Organization Context lets you provide exactly that.

What is context?

Context is background information that helps the AI understand your research environment. This might include:

• Product documentation: Feature descriptions, user flows, terminology definitions
• Company background: Industry, customer segments, competitive positioning
• Research objectives: What you're trying to learn and why
• Previous findings: Insights from past studies that should inform new research
• Domain knowledge: Industry-specific information the AI should know

How to add context

In your organization or project settings, you'll find a new "Context" section. Here you can add markdown or JSON content that will be used by the AI across all research activities in that scope.

Organization-level context applies to everything in your organization. Project-level context applies to all studies within that project. This hierarchy lets you set broad context once and refine it per project.

How context improves research

When context is set, the AI uses it to:

• Ask better questions: Understanding your product helps the AI probe relevant areas
• Generate relevant insights: Context helps distinguish meaningful findings from noise
• Use correct terminology: The AI mirrors your language rather than generic terms
• Connect findings: Previous research context helps the AI identify patterns across studies

Example use cases

SaaS product research: Add documentation about your features, user personas, and known pain points. The AI will ask questions that map directly to your product experience.

Healthcare research: Include relevant regulatory context, patient journey information, and terminology guidelines. The AI will navigate sensitive topics appropriately.

Competitive research: Add information about competitor products and positioning. The AI will probe for comparative insights naturally.

Context persists

Once set, context persists across all studies in that scope. You don't need to re-explain your product for every new research guide—the AI already knows.

This feature is especially valuable for teams running ongoing research programs. Set up your context once, and every study benefits from that foundation.

Creating a great research guide is both an art and a science. You need the right questions to uncover insights, the right structure to maintain flow, and the right stimuli to ground abstract discussions in reality.

Our new AI-powered editor makes this process faster and more collaborative than ever before.

Chat with AI to build your guide

The editor now includes an AI chat interface that helps you develop your research guide conversationally. Describe what you're trying to learn, and the AI will suggest questions, recommend probe points, and help you structure topics logically.

This isn't just auto-complete—it's a collaborative partner that understands research methodology. Tell it you're studying onboarding friction, and it will suggest both broad exploratory questions and specific behavioral probes. Ask it to make a section more conversational, and it rewrites questions to feel natural.

Upload files and images directly

Research often requires context. Maybe you want to discuss a specific screen in your app, a competitor's feature, or a concept mockup. Now you can upload files and images directly into your guide. These become available to show participants during interviews, grounding discussions in concrete artifacts.

Supported formats include images, PDFs, and documents. The AI interviewer can reference these materials during the conversation, asking participants to react to what they see.

Full version history

Research guides evolve. You might test a version, realize a question needs rewording, and iterate. But what if you want to go back to an earlier version?

Every change to your guide is now tracked with full version history. You can:

• See exactly what changed between versions
• Restore any previous version with one click
• Compare how your guide has evolved over time
• Never lose work, even if you accidentally delete something

Streaming AI responses

When you ask the AI for suggestions, you'll see responses stream in real-time rather than waiting for a complete response. This makes the collaborative experience feel more natural and lets you interrupt or redirect if the AI starts going in the wrong direction.

Rich editing experience

The editor supports slash commands for fast formatting, markdown rendering, and a clean interface that keeps you focused on your research objectives. It's designed to get out of your way while giving you the tools you need.

We built this editor because we believe creating great research shouldn't require switching between multiple tools. Everything you need is now in one place.

You've conducted dozens of interviews. The transcripts are full of valuable insights. But turning all that qualitative data into actionable findings? That's where most research projects stall.

Deep Research changes this. It's an AI-powered system that analyzes your interview data at scale and produces structured insights and reports automatically.

How Deep Research works

When you trigger a Deep Research run, the system executes a multi-step pipeline:

1. Context validation: The system reviews your research guide, project context, and any background materials to understand what you're trying to learn
2. Requirement analysis: It identifies the key questions your research should answer based on your guide's objectives
3. Data retrieval: All relevant interview transcripts, notes, and participant data are gathered and prepared for analysis
4. Insight generation: The AI identifies patterns, themes, and notable quotes across your interviews
5. Report creation: Everything is synthesized into a structured report with visualizations

Real-time progress tracking

Deep Research runs can take time, especially for large studies. You can watch the process unfold in real-time, seeing which step is currently executing and reviewing partial results as they're generated.

Automated visualizations

Reports include auto-generated charts and visualizations based on your data. The system can run Python code to create custom charts, analyze distributions, and visualize patterns that would take hours to produce manually.

Built for scale

Whether you have 10 interviews or 100, Deep Research handles the volume. It's designed for the kind of large-scale qualitative research that would be impossible to analyze manually in a reasonable timeframe.

What you get

A complete report including:

• Executive summary of key findings
• Thematic analysis with supporting quotes
• Participant-level insights and notable outliers
• Data visualizations and charts
• Recommendations based on the findings

This isn't a replacement for researcher judgment—it's a force multiplier. Deep Research does the heavy lifting of pattern recognition and synthesis, freeing you to focus on interpretation and action.