Frequently asked questions

AI chatbots (web, app, WhatsApp), voice assistants and IVRs, AI agents and workflows, multilingual AI, websites and mobile apps.

Verified adults (18+) across India, screened and assessed for their languages. Each profile includes city, languages, age band, devices and skills so we can match your customers.

Only those you explicitly authorise in writing or through your customer portal. We prefer staging or test environments.

No. Customer data is never used for AI model training without explicit written authorisation.

From reviewed test evidence only, across six weighted categories. Incomplete evaluations are marked provisional and every report documents the weights.

Yes — platform plans give you a customer portal with AI-suggested scenarios (approved by a human), automated runs on authorised endpoints, dashboards and API access.

Most human tests are delivered within 72 hours of your brief being approved and paid. Growth Retainer and monitoring clients get priority 48-hour slots.

All feedback comes from real people, recorded with their consent. AI only suggests scenarios, transcribes, translates and groups issues to save time. A human reviews every scenario, result and report, and every finding links to its evidence.

No. Testers never post public reviews or ratings. All feedback stays private between AI Yuva and you, in line with Indian consumer-protection rules on incentivised reviews.

Yes. Testers accept confidentiality terms and see only what they need for the task. Raw recordings are deleted 90 days after delivery unless you ask us in writing to keep them longer.

Buy a package in your client portal (an invoice is issued straight away) or request a quote. You receive a GST invoice and can pay by UPI, card or net banking via Razorpay when enabled, or by bank transfer. Testing starts once payment is confirmed.

No. Prices are in Indian rupees plus 18% GST.

Yes. Brands and AI companies worldwide use AI Yuva to test products with real Indian users before launching in India. Contact us for international invoicing.

Yes, free of charge before testing starts. See our Cancellation & Refund Policy for details.

Yes — larger panels, specific cities, languages, enterprise programmes or recurring monitoring. Contact us for a quote.

No. AI Yuva never charges testers any fee.

₹250–₹300 for most 15–20 minute tests, paid weekly by UPI. Longer tests pay more. Every offer shows its payout before you accept.

Anyone 18 or older living in India with a smartphone and a UPI ID. Students, homemakers, shopkeepers and working professionals are all welcome.

No. Just use the product normally and speak honestly in your own language while recording your screen.

It depends on client projects that match your languages, location and device. We do not guarantee work or income.

Approved submissions are paid by UPI in the next weekly payout. You can track status in your portal.

Never your phone number, email or UPI ID. At most, brands see your first name, city, age band, languages and phone type.

Only when a task requires it and only after you give explicit consent for that task.

A link or test access to your chatbot (web, WhatsApp or app) and a short brief on what the bot should and should not do. We only test systems you explicitly authorise.

Both. Real Indian testers hold the conversations and judge quality. AI helps generate scenario ideas and summarise results; a human reviewer approves every scenario and every score.

Yes, when you provide a test number or sandbox. Our testers converse with it from their own phones.

Each test case is scored across six categories using configurable weights. Scores are only produced from recorded evidence, and incomplete evaluations are clearly marked provisional.

A smaller entry test: 10 real users try your AI product — an assistant, a generator, a support bot — in Hindi or Hinglish, and we score accuracy, trust and confusion. It works on its own or as an add-on to another test.

No. Testers never post public reviews or ratings. All feedback stays private between AI Yuva and you, in line with Indian consumer-protection rules on incentivised reviews.

Testers record only with their explicit consent and only for systems you authorise. Recordings are stored securely and used solely for your evaluation.

Word error rate — the share of words your speech-to-text got wrong compared with what the caller actually said. Lower is better.

Hindi, Hinglish and Indian English by default — including Awadhi- and Bhojpuri-influenced Hindi from callers across Lucknow, Kanpur, Varanasi, Patna and nearby towns. Additional regional accents are offered only when qualified testers are available.

No. Share a phone number, WhatsApp number or chat link. Testers use it exactly like your customers would.

Yes. Tell us your customer mix and we match testers by state, city type, language and age.

For every turn we check whether the bot correctly understood what the caller meant, using the recording and transcript. A human reviews every score.

No. We strongly prefer a staging or sandbox environment with test data. We never act on systems without written authorisation.

You share tool-call logs or traces for each test; we compare them with the expected actions for that case.

Verified adults (18+), mainly from Tier-2 and Tier-3 towns, matched by city, language, age band and phone type.

Yes — share a store link, TestFlight or APK. Testers use their own devices — mostly Android, because that is what most Tier-2/3 users have. Tell us if you need iPhone users.

One journey with a clear goal, such as "find a face wash under ₹400 and pay by UPI". Most take 10–20 minutes.

In the Growth package, yes — a 30-minute video call with the person who reviewed your sessions.

For teams that ship often: two Growth tests every month, a trend dashboard and priority 48-hour slots. Minimum term three months, then month to month with 30 days' notice. Unused tests roll over for one month.

Languages are activated for paid testing only when we have qualified, assessed testers. The current list is shown on the enquiry form.

Automated re-runs happen only for systems you have connected and authorised (for example a chatbot API endpoint). Human evaluation is scheduled for the rest. We never claim to monitor a system we cannot access.