Yash Vijayvargiya

Everyone in India thinks AI robocalling means a robotic voice saying "Sir, would you like a personal loan?" or maybe even "Main Arvind Kejriwal bol raha hoon" if you live in Delhi. And then you hanging up in 3 seconds. That was in the past. It is not what is happening in 2026. Let me tell you what happened when we tried it. March 2025. We decided to test AI voice calling at Skill Arbitrage. We had a sales team making calls. Good people. Trained well. But we were capped. 80 to 100 calls per person per day. We needed to reach 30,000 leads a month. The math did not work with humans alone. So we called one of the top AI calling companies. They set it up in a week. We gave them the script. The objection handling. The FAQs. The customer database. They said "leave it to us." First batch of calls went out. Disaster. The AI sounded perfect. Too perfect. Crystal clear voice. Flawless Hindi. No pauses. No breathing. No background noise. Like talking to a newsreader on Doordarshan. People hung up. Not because they thought it was a robot. Because something felt off. They could not explain it. They just did not trust the voice. Our conversion rate was worse than our worst human caller. We almost killed the project. Then someone on our team had an idea. What if we made the voice worse on purpose? We added a tiny bit of background noise. The kind you hear when someone is calling from an office with other people around. We added small pauses before answers, the way a real person takes a second to think. We made the voice slightly less polished. Not robotic. Just human. Conversion went up 40%. That was the first lesson. Humans do not trust perfection on a phone call. A voice that is too smooth triggers the same instinct as a salesperson who is too polished. You want to leave the showroom. A little imperfection signals "real person." Even when the listener probably knows it is not. Then the second surprise. We expected massive hangup rates. Everyone told us "Indians will not talk to robots." We braced for 30, maybe 40% dropping the call immediately. 6% hung up. 94% engaged normally. They answered questions. Confirmed details. Booked appointments. Made decisions. 94 out of 100 people did not care that the voice was artificial. They cared that the call was relevant and respected their time. A bored human reading the same script for the 80th time that day was actually less engaging than a well-designed AI call. Then the third discovery. This is the one that changed how I think about AI calling entirely. Our human QA team could review maybe 30 calls a day out of the thousands being made. They would catch a problem, coach a caller, and hope the fix would spread to the rest of the team by next week. With the AI, we could audit every single call. Every word. Every response. Every point where the conversation broke down. We would find a pattern. "When the lead says 'I already looked into this,' the AI gives a generic response and loses them." We would rewrite that one response. Deploy it. Within an hour it was live on every call. Five improvement cycles in a day. Our human team used to do five in a quarter. By the second month our AI caller was outperforming our best human salesperson on the metrics that mattered. Not because it started better. Because it improved 100x faster. We started with a system that was honestly embarrassing. We iterated it 50 times in 30 days. Nobody who heard it in month two would believe it was the same system. Now here is the part I wish someone had told us before we started. The technology is cheap. Bolna, Vapi, Bland, Exotel. Rs 1 to Rs 5 per minute. A 2-minute call costs less than Rs 10. Compare that to a human caller at Rs 20,000 a month making 80 calls a day. Any vendor can set it up in a week. That is not where the money is won or lost. We went through three vendors before we figured out the real problem. Every time we gave a vendor our process and said "build it," we got a technically functional system that produced mediocre results. The calls connected. The voice worked. The script played out. But nothing converted. Because the vendor did not know our business. What does the AI say when someone asks "how is this different from that other course I saw on Instagram?" That is not in any FAQ document. That is business judgment. When does the AI push and when does it back off? When someone says "call me later," do you call them later or is that a polite rejection? If they say "I need to ask my husband," do you offer to call back when he is available or do you handle the objection now? When the lead switches from Hindi to English mid-sentence, how does the AI respond? In Hindi? In English? In Hinglish? The answer depends on what that switch signals about the caller's comfort level. No vendor can figure this out for you. These are not technology problems. They are sales judgment calls that only someone inside your business can make. Every company I have seen get extraordinary results from AI calling has one thing in common. Not a better vendor. Not a more expensive platform. They have one person on their own team who owns the prompt. This person listens to 50 calls a day. Spots where conversations break. Rewrites the response. Tests it. Listens again. They are not an AI engineer. They are someone who understands the customer and knows what a good sales conversation sounds like. This person is the difference between AI calling that produces mediocre results and AI calling that makes your competitors wonder what you are doing differently. You would never hand a telemarketing agency a one-page brief and expect them to figure out your pitch. You would train them. Listen to their calls. Coach them weekly. AI calling is the same. Except the coaching is editing a prompt and the improvement deploys in seconds instead of weeks. We call over 30,000 leads a month now. We deployed AI for onboarding too. It moved our key metrics in ways I did not think were possible 18 months ago. But the reason it works is not the AI. It is the person on our team who has been shaping it every single day since we started.

@phantomthread_d You can reach out to @VedantValia. We are working with them. Great team and product.

@Yash912 @law_ninja Thanks, please DM me the details. I’m trying to build one but not getting under 400ms latency.

Everyone in India thinks AI robocalling means a robotic voice saying "Sir, would you like a personal loan?" or maybe even "Main Arvind Kejriwal bol raha hoon" if you live in Delhi. And then you hanging up in 3 seconds. That was in the past. It is not what is happening in 2026. Let me tell you what happened when we tried it. March 2025. We decided to test AI voice calling at Skill Arbitrage. We had a sales team making calls. Good people. Trained well. But we were capped. 80 to 100 calls per person per day. We needed to reach 30,000 leads a month. The math did not work with humans alone. So we called one of the top AI calling companies. They set it up in a week. We gave them the script. The objection handling. The FAQs. The customer database. They said "leave it to us." First batch of calls went out. Disaster. The AI sounded perfect. Too perfect. Crystal clear voice. Flawless Hindi. No pauses. No breathing. No background noise. Like talking to a newsreader on Doordarshan. People hung up. Not because they thought it was a robot. Because something felt off. They could not explain it. They just did not trust the voice. Our conversion rate was worse than our worst human caller. We almost killed the project. Then someone on our team had an idea. What if we made the voice worse on purpose? We added a tiny bit of background noise. The kind you hear when someone is calling from an office with other people around. We added small pauses before answers, the way a real person takes a second to think. We made the voice slightly less polished. Not robotic. Just human. Conversion went up 40%. That was the first lesson. Humans do not trust perfection on a phone call. A voice that is too smooth triggers the same instinct as a salesperson who is too polished. You want to leave the showroom. A little imperfection signals "real person." Even when the listener probably knows it is not. Then the second surprise. We expected massive hangup rates. Everyone told us "Indians will not talk to robots." We braced for 30, maybe 40% dropping the call immediately. 6% hung up. 94% engaged normally. They answered questions. Confirmed details. Booked appointments. Made decisions. 94 out of 100 people did not care that the voice was artificial. They cared that the call was relevant and respected their time. A bored human reading the same script for the 80th time that day was actually less engaging than a well-designed AI call. Then the third discovery. This is the one that changed how I think about AI calling entirely. Our human QA team could review maybe 30 calls a day out of the thousands being made. They would catch a problem, coach a caller, and hope the fix would spread to the rest of the team by next week. With the AI, we could audit every single call. Every word. Every response. Every point where the conversation broke down. We would find a pattern. "When the lead says 'I already looked into this,' the AI gives a generic response and loses them." We would rewrite that one response. Deploy it. Within an hour it was live on every call. Five improvement cycles in a day. Our human team used to do five in a quarter. By the second month our AI caller was outperforming our best human salesperson on the metrics that mattered. Not because it started better. Because it improved 100x faster. We started with a system that was honestly embarrassing. We iterated it 50 times in 30 days. Nobody who heard it in month two would believe it was the same system. Now here is the part I wish someone had told us before we started. The technology is cheap. Bolna, Vapi, Bland, Exotel. Rs 1 to Rs 5 per minute. A 2-minute call costs less than Rs 10. Compare that to a human caller at Rs 20,000 a month making 80 calls a day. Any vendor can set it up in a week. That is not where the money is won or lost. We went through three vendors before we figured out the real problem. Every time we gave a vendor our process and said "build it," we got a technically functional system that produced mediocre results. The calls connected. The voice worked. The script played out. But nothing converted. Because the vendor did not know our business. What does the AI say when someone asks "how is this different from that other course I saw on Instagram?" That is not in any FAQ document. That is business judgment. When does the AI push and when does it back off? When someone says "call me later," do you call them later or is that a polite rejection? If they say "I need to ask my husband," do you offer to call back when he is available or do you handle the objection now? When the lead switches from Hindi to English mid-sentence, how does the AI respond? In Hindi? In English? In Hinglish? The answer depends on what that switch signals about the caller's comfort level. No vendor can figure this out for you. These are not technology problems. They are sales judgment calls that only someone inside your business can make. Every company I have seen get extraordinary results from AI calling has one thing in common. Not a better vendor. Not a more expensive platform. They have one person on their own team who owns the prompt. This person listens to 50 calls a day. Spots where conversations break. Rewrites the response. Tests it. Listens again. They are not an AI engineer. They are someone who understands the customer and knows what a good sales conversation sounds like. This person is the difference between AI calling that produces mediocre results and AI calling that makes your competitors wonder what you are doing differently. You would never hand a telemarketing agency a one-page brief and expect them to figure out your pitch. You would train them. Listen to their calls. Coach them weekly. AI calling is the same. Except the coaching is editing a prompt and the improvement deploys in seconds instead of weeks. We call over 30,000 leads a month now. We deployed AI for onboarding too. It moved our key metrics in ways I did not think were possible 18 months ago. But the reason it works is not the AI. It is the person on our team who has been shaping it every single day since we started.

@SuryaRajendhran @law_ninja We are using Timepay by @VedantValia ! Not the biggest but easily one of the best teams and product in the industry.

@Yash912 @law_ninja So what platform did you actually stick with?

@DealzAmazin Well, how does it make a difference if its AI or human making the call? and lovely jumping to conclusion that we violate privacy rules, trai, etc. All our calls are to inbound leads with all checks and points visible. Please read my post again, maybe you will understand how.

@Yash912 So, you are the one responsible for those pesky AI robocalls. Believe me people detest such brands or people who are perfectly OK with automating privacy violations, disregarding TRAI DND rules, with no respect for the other person's time. Shame.

We tried eleven labs too. Didnt work for us. The voices are decent. It's a combination of telephony and vendor that you are using. Dming you details. Also, the trick is not to have calling trained on extensive knowledge base, but limit it to what is important without hardcoding too much.

@Yash912 @law_ninja 300-450ms is too good, which tool are you using. For eleven labs, i was getting 800ms

@gansub Thats the entire point of the post. How to do high quality, high value calls, not collections or transactional calls, like most people believe with AI. How it is very much possible now.

@Yash912 The context and nuances matter a lot. Circa 2001, in a voice BPO, the customers were reporting poor quality. It turned out that the context of the call was collection with a threat of disconnection. Why would I rate the call quality with magnanimity?

@BishtjiSanjay These are registered users or those who have shown interests. 30% of those calls are successful counselling (>10 mins), far higher than what we achieved with human callers. We are not selling anything on these calls.

@Yash912 So you were converting 30% of the supposed 30k cold calls? Yeah right.

@venkatarangan Glad that you found this useful :)

@Yash912 Hands down one of the best guide on how to deploy AI and get real results. What you have outlined is applicable beyond voice calls, the basics are valid for every AI deployments. Thanks for sharing.

@Fursatbhai Would you feel that even if you think that call was valuable and purposeful, not just colelcting data but giving insights?

@Yash912 I got such a call. And I hung up. I got a feeling that I am a real customer not a resource for making money for corporates. This element needs to be carefully addressed by the product.

@yashpresso Correct, same goes here. The hangups will always be there, no matter a human is calling or AI is calling.

@Yash912 Lot of insights in this!! When I got such AI call, I hung up because I wasn't interested, but I could only think how effective it would be for actual customers with intent

@El_saaaad Glad you found it valuable.

@Yash912 Amazing insights. Thank you.

@SurajSalimath Yes, I'll post about that very soon :)

@Yash912 Yash nicely written post! What if you had system AI that listen to your Human Agents calls and learns in real time? Tried this ?

Yes, and it offers them to shift to a human caller. Surprisingly only 4% of our users request for human caller. It all depends on how valuable your content is and if you are able to handle those objections effectively. More often than not, AI is better trained to answer technical questions than most human callers can.

@Yash912 At the least, do you inform the users that they’re talking to a AI chat bot? It’s unethical for an AI system to pose as humans — that should be a ground rule

@sheriffly They already do and we get humans, picking up in between in quite a lot of calls. Thats the best part, if you have good effective valuable script, it goes through screening and leads pick up and engage.

@Yash912 Soon you will have AI answer those AI marketing calls too on behalf of customer, and the receiver's AI screener will reject offers of these robocall AI and put them in spam list. This is what iOS and Android will do in future.

@brofessormaddy @law_ninja Not just that. Most people cannot even tell that its AI, but they know there is something spooky about the call when its too perfect.

@Yash912 @law_ninja brilliant share. my takeaway/ ‘the voice is too perfect’ = something’s off = no trust it’s like these humaniser AI tools working these days to increase perplexity and burstiness of AI generated text. AI products moving towards ‘how to not make it seem like an AI’. Interesting.

Great question. There is a fine balance between latency and how much database you can train your voice calls on. Mostly latency has been 300-450ms varying on telecom providers. But we have found most of these to be good. In fact, when we tried really low latency, the calls became extremely superficial.

@Yash912 @law_ninja 1. How was the latency, was it under 400ms? 2. Were customers aware they were talking to AI or they continue conversation like they do with humans?

@manialok We are edtech, a lot of these calls go for free counselling and career roadmaps, and converting them to show up in our webinars.

@Yash912 Nicely written, Number seems off…didn’t mention what do you sell by calling 30K per month… may be its different than what we sell.

@neoviky Yes, humans who have good understanding of sales, processes and effective communication. Most importantly, who are able to create scripts that respect time and are able to give very high value to those they are speaking to.

@Yash912 This brings back memories of Dell and American Express call centers in GGN. This key lesson i now understand is that AI alone won't be impactful. The tech is good on paper, but real results come when it is fine tuned with humans in the loop in every step.

@CeylaAi Yes, and it surprisingly is able to effectively communicate the purpose in those voicemails also which get picked up. And for those of us who dont answer, we would not have answered to human callers also :)

@Yash912 I never answer my phone anymore.

the impact ai summit in delhi was a perfect demonstration of why india keeps losing in tech and i’m tired of pretending it wasn’t a disaster. let me paint the actual picture: > cash-only payments at a “digital india” upi ?? > pm visit → main hall cleared for hours, everyone else just stood around doing nothing > exhibitors locked out of their own stalls > 3-hour queue just to enter > a founder’s product got stolen during the summit > no wifi at an ai event. > can’t take your keys if you came via car/bike > no laptop/camera at tech event > people were asked to sit on the ground > speaker lineup with consultants/bureaucrats who’ve never shipped a real product > the registration system crashed multiple times. people who registered weeks in advance couldn’t get in. vips walked past massive queues while founders and builders stood outside in the heat. 🤡 and 27 countries witnessed all of this live networking areas? no space to stand. many demos didn’t work because there was no stable internet. 5g?? this is what happens when optics matter more than execution. when innovation becomes photo-op the sad part is india has insane talent. founders building world class products. engineers and researchers doing real work. leave India for a sec, im at network school and the youngest crowd is all Indians. but we keep shooting ourselves in the foot with performative nonsense. the west isn’t winning because they’re smarter. they’re winning because they care about details. because they respect builders. because their tech summits actually work. same story when @sama came to india last time. boomer uncles asked the dumbest questions. and when he said it’s hard for india to build foundational models, we took it on our ego. rn, every founder who attended left embarrassed. imagine international delegate left with stories about our “infrastructure.” many friends and young builder lost a little more faith. this wasn’t just bad planning. it was a signal of what we value. and clearly, it’s security theater and photo-ops over builders. we can do better. we have the talent. we have the market. we have the potential. what we don’t have is execution and respect for the people building the future. maybe one day we will do better. till then if you’re a founder, ignore the noise. keep building.