Short answer: Choose Voicepanel if you need consumer reach — a large built-in panel, many modalities in one tool, and the lowest entry price in the category. Choose Conveo if your work is classic market research with recruited video panels and you want multimodal analysis that reads tone and expression, with budget to match. Choose Koji if you want structured depth rather than breadth: AI-moderated voice and chat interviews with six question types in a single study, from EUR 29/month.
These two are often shortlisted together because both are AI-native, both are Y Combinator companies, and both promise interviews-plus-analysis in one loop. In practice they are close to opposites. One optimises for how many people you can reach and how many ways you can reach them. The other optimises for how much signal you can extract from a single video conversation.
Why the category shifted
The move away from static surveys is no longer a forecast. 72% of insights buyers now use generative AI in at least one stage of a research project, up from 23% in 2023, and AI-moderated interviews specifically are used by roughly 81% of dedicated research teams, 73% of UX teams (up from 38% in 2024), and 67% of product teams.
Cost is why. AI-moderated qualitative interviews run about $8-$15 per completed interview against $150-$300 for human-moderated sessions — and industry data suggests AI-moderated interviews produce meaningfully richer responses than surveys, because a static form cannot ask the follow-up question that actually matters. Within a market research industry projected near $150 billion in 2026, AI-native qualitative is the only double-digit-growth segment.
Voicepanel and Conveo are two very different bets on that shift.
Voicepanel vs Conveo: side-by-side
| Voicepanel | Conveo | Koji | |
|---|---|---|---|
| Backing | Y Combinator | Y Combinator, $5.3M raised | Bootstrapped, self-serve |
| Core modality | Voice, video, phone calls, surveys, usability tasks | AI-led async video interviews | Voice + chat interviews |
| Analysis approach | AI summaries and themes with quotes, audio, and video clips | Multimodal — tone, facial expression, and behaviour tied back to verbatim clips | Automatic thematic analysis + transcript chat |
| Languages | 35+ | 50+ | 40+ |
| Built-in panel | 30M+ consumers and professionals, 150+ countries, 10+ panel partners | 8 panel partners (Respondent, User Interviews, Norstat, Bilendi, Sago, Rakuten, Forsta, Rally) | Bring your own + recruit in-product |
| Structured question types | Survey-style questions, separate from interviews | Primarily open-ended video | 6 types inside one study |
| Entry price | Reported from ~$15-$99/month; free tier available | From ~$990/month; enterprise reported near $45K/year | EUR 29/month (29 credits) |
| Enterprise billing | Custom | Prepaid credits by interview minutes | EUR 79/mo tier + credit packs |
| Quality-gated billing | No | No | Yes — only interviews scoring 3+ bill |
| Compliance | SOC 2 Type II, GDPR, HIPAA; no training on customer data | GDPR | GDPR-compliant, EU-transparent pricing |
| Self-serve signup | Yes | Sales-led | Yes |
What Voicepanel actually does
Voicepanel positions itself as "AI agents for customer research," and its bet is breadth.
Where it is strong:
- Modality coverage. Surveys, evals, voice, video, phone calls, and usability tasks in one platform, spanning prototypes, websites, and mobile apps. Few competitors cover phone calls at all — genuinely useful for reaching demographics that will not join a video session.
- Panel scale. Access to 30M+ consumers and professionals across 150+ countries, plus 10+ panel partners including User Interviews, Respondent, and Prolific. For B2C concept testing at volume, this is the strongest reach in the comparison.
- Compliance depth. SOC 2 Type II, GDPR, and HIPAA, with a no-training-on-customer-data policy. HIPAA in particular opens healthcare work that most AI research tools cannot touch.
- Accessible entry. A free tier and reported entry pricing between roughly $15 and $99/month, plus a pay-as-you-go Agent plan for one-off projects. You can evaluate it without a contract.
The trade-off is focus. Spanning surveys, tasks, phone, voice, and video means the interview engine is one capability among many rather than the deep specialism of the product. Teams wanting a rigorous, structured interview instrument sometimes find the breadth comes at the cost of depth in any single mode.
What Conveo actually does
Conveo (Y Combinator, $5.3M raised) markets "confident decisions in days with AI-led interviews," and its bet is depth in video.
Where it is strong:
- Multimodal analysis. This is the real differentiator. Conveo reads tone, facial expression, and behaviour alongside the transcript, then ties what it finds back to verbatim quotes and the original clips. For ad testing and concept reaction work — where how someone says it matters as much as what they say — nothing else in this comparison does it.
- Language breadth. 50+ languages, the widest of the three.
- Panel-partner network. Integrations with eight providers (Respondent, User Interviews, Norstat, Bilendi, Sago, Rakuten, Forsta, Rally) plus bring-your-own, which suits agencies running recruited studies across markets.
- Study assistant and Ask-Anything repository. Helps structure design and quotas, and lets you query past studies as a collective memory rather than a folder of decks.
The trade-off is cost and shape. Conveo starts around $990/month, and third-party buyer guides put enterprise near $45,000/year billed as prepaid credits by interview minutes. Minute-based billing means a discursive participant costs more than a concise one — fine for scoped agency projects, awkward for continuous discovery. And video-first is heavier than many product teams need for fast iteration.
The honest head-to-head
Reach: Voicepanel wins decisively. 30M+ panel across 150+ countries against Conveo's eight partner integrations is not a close comparison for consumer volume.
Analytical depth per conversation: Conveo wins decisively. Multimodal reading of tone and expression extracts signal Voicepanel's transcript-and-clips approach does not attempt.
Languages: Conveo wins (50+ vs 35+), though both cover far more than most teams use.
Compliance: Voicepanel wins. SOC 2 Type II plus HIPAA is a materially stronger posture than GDPR alone, and it is often the deciding factor in regulated industries.
Cost: Voicepanel wins decisively. A free tier and sub-$100 entry against a ~$990/month floor is a different order of commitment. Conveo's minute-based enterprise billing also makes forecasting harder.
Accessibility: Voicepanel wins. Self-serve signup versus a sales-led motion.
The fair summary: Conveo is the better instrument, Voicepanel is the better platform. If you are an agency doing ad and concept testing on recruited video panels, Conveo's depth justifies its price. For nearly everyone else, Voicepanel's economics are hard to argue with.
What both leave on the table
Voicepanel and Conveo sit at opposite ends of a spectrum — maximum breadth versus maximum video depth — but they share a structural limitation: neither gives you qualitative depth and quantifiable structure inside the same interview.
With Voicepanel, surveys and interviews are separate instruments. With Conveo, video interviews are open-ended by design. Either way, when a stakeholder asks "what percentage of them said that?", you are running a second study or hand-counting transcripts.
Koji is built around closing exactly that gap:
- Six structured question types in a single study — open-ended, scale, single choice, multiple choice, ranking, and yes/no. One instrument produces the narrative and the number. See open-ended questions and multiple-choice questions, or the broader case in AI interviews vs surveys.
- AI-moderated voice and chat interviews on an ElevenLabs voice stack, probing adaptively in real time, available 24/7 with no scheduling and no moderator bias. See AI voice interviews and voice vs text interviews.
- Automatic thematic analysis with AI auto-tagging across every transcript, plus chat with your transcripts — ask one question across 100 interviews, get a cited answer in seconds.
- Customizable AI consultants tuned to your domain, so the analysis argues from your context rather than generic frameworks.
- Quality-gated credits. Only conversations scoring 3+ on the quality rubric consume credits. Neither Voicepanel nor Conveo protects you from paying for drop-offs and low-effort responses.
- Transparent EU pricing. EUR 29/month (Insights, 29 credits) or EUR 79/month (Interviews, 79 credits); text interviews cost 1 credit, voice 3. Ten free credits at signup, cancel monthly, GDPR-compliant by design.
- Recruit from your own product — for B2B and product teams, your best participants are already using your software.
The concrete math: 30 voice interviews on Koji is 90 credits — roughly EUR 90 on top of a EUR 29 plan. The same 30 interviews on Conveo require a plan starting near $990/month. Against Voicepanel the price gap narrows considerably, and the real question becomes structure: do you want one instrument that returns both themes and percentages, or two tools that each return half?
How to choose
Pick Voicepanel if you need a large consumer panel, want phone and usability tasks alongside interviews, work in a HIPAA-regulated context, and want the lowest possible entry cost.
Pick Conveo if you run agency-grade market research on recruited video panels, need tone and expression analysis for ad or concept testing, work across many languages, and can justify a ~$990/month floor.
Pick Koji if you want mixed-method depth in one study, transparent pricing with no contract, quality-gated billing, and to be running real interviews with your own customers this week. For the wider field, see our best AI interview software guide and the user research cost calculator.
Run a real study this week
You do not need a $990/month floor or a second tool to get numbers out of your qualitative research.
Koji gives you 10 free credits at signup — enough to run genuine AI-moderated voice or chat interviews with your own customers, blend open-ended and structured questions in one study, get automatic thematic analysis, and generate a one-click report. From research question to synthesized insight in hours, not weeks. No research expertise required, no sales call, no annual commitment.
