Founder, AyushPay. Building AI in India without writing a line of code. 600+ hrs, 5 products, 41 agents. Sharing what actually works and how to leverage AI.
Update. After a week, multiple reminder and wrong commitment the car was finally taken for cleaning the message they did and guess what it came back worse.
@18002096006@anandmahindra
Update: After 1000s of follow-up and waiting for 5 months with excuses, my car finally went for PPF coating, I was told it will take 3-4 days. It's the 8th day and no one even knows where the car is. People are not responding to calls/messages.
What a continued miraculous service @18002096006@anandmahindra I will urge you to please seriously looking at the service issue, it has and is damaging the reputation of the brand and people will think thrice even before thinking of buying a @Mahindra_Auto cars.
Very pathetic service by @MahindraRise & @KonceptMahindra
@18002096006 Received now but in a spoiled condition. All seats which were white in color are no more white. Completely spoiled seats and experience along with it.
Extremely disappointed with the post-sales & service experience of my Mahindra XEV 9e. From fake commitments to vehicle mishandling, the experience has been a nightmare. 4 months of follow-ups and still no resolution. @anandmahindra@Mahindra_Auto@MahindraRise#XEV9e#Mahindra
What are you apologetic of? Your team is debating, just giving me dates, saying twitter is useless, you can keep complaining. This is a clear case of service delivery. I want the country to know how pathetic and unapologetic your services are so that future Customers should be aware of what they are getting into. I will ask my friends to cancel the booking, they should just go for @TataMotors or @MGMotorIn, they understand customer service and not just sell the car without any service.
#MahindraSucks
Just watched ElevenLabs drop a slick sponsored ad telling everyone how their AI agents will turn WhatsApp into the ultimate support channel......while immediately below it, a post screams in blood-red letters: "ELEVENLABS IS DONE" The algorithm just served us the most brutal corporate funeral procession in real https://t.co/MVqWm5NX7X tab: "We're revolutionizing customer support!"
Next tab: "Open-source voice model VoxCPM2 just smoked us on silicon."Nothing says "disruption" quite like your own marketing material getting ratio'd by your obituary in the same scroll.Welcome to 2026, where the hype cycle and the replacement cycle are now running on the same https://t.co/tOXVQbxNXe ElevenLabs' moat.
It died of irony.#AI #VoiceAI #OpenSource
Using AI to replace 5 call centre agents is like using a JCB to pour milk in your coffee.
Extraordinary capability. Wrong problem.
Real question is not how do I do this faster.
It is what can I now build that I could not before.
Two different questions.
Your tech team is the last to learn AI. Not the first.
Engineers hired for AI and most did not follow what I was building.
People who adapted fastest were fresh graduates.
Least to unlearn.
Learn it yourself first. Then hire young people who report directly to you.
@vijayshekhar Totally agree ๐ฏ
Productivity from AI will push companies to grow faster, not shrink. History shows this every time, new tools create more opportunities than they take away. Looking forward to the AI-driven expansion in India.
Interesting chart from FT, Anthropic's surge in business adoption is impressive, especially for coding and deep reasoning tasks. Personally, I use a mix:Claude (paid) as my go-to for complex research, writing, and structured thinking โ it feels the most reliable right now.
Grok via X Premium bundle for quick, unfiltered insights, real-time knowledge, and a bit of wit (the bundle makes it seamless).
ChatGPT occasionally for creative brainstorming or when I need broad accessibility.
Gemini for deep research.
Minimax 2.7 for labour coding.
No single model wins everything yet. The real power comes from switching between them based on the job. Multi-model usage seems to be the practical reality for most power users. Whatโs your primary one,@vijayshekhar
Team had a product in backlog for two months
never came out of the sprint
built it in 6 hours with Claude Code
system that listened to every call and scored it automatically
real problem was not the team. it was the sprint model.
that was the moment i saw AI differently
1/
For three days, I used AI in the laziest possible way.
Claude generated code. I pasted it into n8n. Something broke. It gave a fix. I pasted again.
30 to 40 hours disappeared like that. ๐งต
6/
AI is not at its best when it replaces your thinking.
It is at its best when it sharpens it.
Used well, it compresses thinking.
Used poorly, it accelerates confusion.