Shadab Khan

Founder @EligentAI . Building production AI systems end to end—agents, support automation, AI front desks. You own the code. Live in weeks. 🇮🇳 → Worldwide.

[Shadab Khan @ShadabBuilds]

Over the past few months I’ve been building AI systems around support workflows, analytics, RAG infrastructure, and enterprise copilots.

Decided to bring everything together under @EligentAI.

Still early, but excited to keep building publicly.

📌 https://t.co/UTEW3ydaqy

[Eligent AI @EligentAI]

Building intelligent systems for modern businesses.

Currently working on:
• AI customer support infrastructure
• enterprise copilots
• RAG systems
• multi-agent AI workflows
• AI-powered analytics

More updates soon.
📍 https://t.co/j4pIWYsEjm

[Shadab Khan @ShadabBuilds]

Improved my front desk demo.

This time both sides on one screen:
left = patient booking on a dental clinic's site
right = the clinic's dashboard updating live

Patient books, it instantly shows up for staff to confirm. No tab switching this time.
54 sec 👇

#AI #Dentistry

[Shadab Khan @ShadabBuilds]

@kelvinbuildss Cortisol

[Shadab Khan @ShadabBuilds]

@emonmeena @organism_sh @Meta @WhatsApp Smart constraint... replies-within-existing-threads is basically the whole ban-mitigation story. Curious what your session drop rate looks like once a number crosses ~500 active threads? That's where linked-device usually gets flaky in my experience.

[Shadab Khan @ShadabBuilds]

@omkarships More competitive at the bottom, less at the top.

AI flooded the market with people who can produce code. It did nothing to increase the number who can actually ship a reliable system end to end, understand the problem, build it, get it to production.

That gap got wider.

[Shadab Khan @ShadabBuilds]

@sattyyouneed Agree, mostly. But in production this stops being a debate - you route each task to whichever model does it best. Claude for writing and agents, GPT for reasoning over messy data.

Picking one team is a consumer habit. Builders just route.

[Shadab Khan @ShadabBuilds]

@omkarships Everyone has the same tools. Almost nobody ships them into something that actually works.

That gap... a demo that impresses vs a system running reliably in production - is the entire game. Execution is the moat. It always has been.

[Shadab Khan @ShadabBuilds]

@Venkydotdev DeepSeek

[Shadab Khan @ShadabBuilds]

Spent today cleaning up a demo I've been building. Small detail, but getting the "appointment confirmed" flow to feel instant was oddly satisfying.

What's a small thing in your work that's way more satisfying than it should be?

[Shadab Khan @ShadabBuilds]

@pilvand Less pure n8n, more custom AI agents + APIs - but I glue tools when workflows fit.
What’s the beta?

[Shadab Khan @ShadabBuilds]

@Polymarket $10B for compute access is the quiet truth of AI right now:

Models matter.
But GPUs + power + contracts decide who ships.

[Shadab Khan @ShadabBuilds]

The 5 that must be grounded:
1) Timings
2) Insurance
3) Fees
4) Doctor / specialty
5) What to bring / prep

No source → no answer.

[Shadab Khan @ShadabBuilds]

Clinic AI that invents fees or insurance answers is worse than no bot.

Patients remember the wrong answer.
Staff clean up the mess.

I wrote how production systems ground FAQs in *your* docs (RAG) - and the 5 questions that must never be guessed.

https://t.co/2eyyJsCYax

[Shadab Khan @ShadabBuilds]

Every hospital works differently.
That's why this isn't just another chatbot.

Each AI Front Desk is configured around the hospital's own doctors, timings, fees, booking workflow, and notification preferences.
Excited to keep improving it.

[Shadab Khan @ShadabBuilds]

Built an AI front desk for hospitals and clinics.
Patient asks a question → instant answer, pulled from the hospital's own data.

Wants to book appointments→ chat naturally agent will book it , or use a guided wizard: pick doctor, day, slot, fill details. Either way, done in under 30 seconds.

Staff gets notified, confirms in one tap, patient gets an email.
One tap-- that's it.

Watch it end to end 👇
DM me for a live demo.

Building AI systems at @EligentAI

#AI #Healthtech

[Shadab Khan @ShadabBuilds]

Built an AI front desk for hospitals and clinics.
Patient asks a question → instant answer, pulled from the hospital's own data.

Wants to book appointments→ chat naturally agent will book it , or use a guided wizard: pick doctor, day, slot, fill details. Either way, done in under 30 seconds.

Staff gets notified, confirms in one tap, patient gets an email.
One tap-- that's it.

Watch it end to end 👇
DM me for a live demo.

Building AI systems at @EligentAI

#AI #Healthtech

[Shadab Khan @ShadabBuilds]

Spent the last few weeks building an AI receptionist for clinics and hospitals.
The part everyone underestimates: taking a booking is the easy bit. Knowing when to stop and get a human is the hard part.

Patient wants a Tuesday slot? Bot handles it start to finish.
Patient describing pain? That goes to a human. Instantly.

Recording the full flow today -- web widget and Telegram booking, live human handoff, all of it.
Demo dropping in a few hours.

[Shadab Khan @ShadabBuilds]

@ClaudeDevs Classy move on the reset. I'll keep my fallback config either way -- this week was the best free lesson in why you never depend on one model. Good to have Fable back though.

[Shadab Khan @ShadabBuilds]

@Polymarket Good news if you depend on it. Irrelevant if you built right -- your product never went down in the first place.
That's the whole point.

[Shadab Khan @ShadabBuilds]

Been building an AI receptionist for clinics, and the same problem shows up everywhere:
A patient messages after hours to book. Nobody's at the desk. By morning they've booked with another clinic instead.
Not a staffing failure -- your front desk is busy with the patients already in the room. It's a coverage gap.
What it does today:

books appointments through a web widget on your site
books through Telegram too
answers the routine stuff -- timings, location, what to bring
hands off to a human the moment it's an actual medical question

WhatsApp booking is what I'm building next -- that's where most patients already are.
Building the first pilot with a hospital now. If you run a clinic and this sounds like your mornings, let's talk.

[Shadab Khan @ShadabBuilds]

Congrats team. "Sovereign control on the AI stack is no longer optionality" might be the line of the year... and it's the same lesson at every scale.
The Fable ban just proved it:
don't control your stack, and someone else decides when you get access.
Building from India and rooting for this.

[Shadab Khan @ShadabBuilds]

@Polymarket $3.6B confirms what anyone building in this space already knows: AI customer service is the category.
But Salesforce-scale tools price out the businesses that need it most.
The clinic, the local shop, the small agency. That gap is the real opportunity, and it's wide open.