PRANEETH REDDY

Portrait of Praneeth Reddy AI automation & integrations

AI automation for everyday work

I build systems that collect money, update records, and handle calls.

Follow-up calls, browser work, GoHighLevel, and AI receptionists that connect to the tools a business already uses.

See the work Book a 30-minute call

Voice AI award

1st place for building an AI voice agent.

Vapi Build Contest, a contest for AI voice agents. December 2025. Judged on:

  • Clarity
  • Reliability
  • Real-world usefulness
  • Conversation quality

$500 in Vapi credits

Vapi Build Contest certificate

Open full size
Certificate of achievement: Praneeth Reddy, 1st place, Vapi Build Contest, December 2025, $500 in Vapi credits

Selected automation work

Every case shows the problem, what I built, and the proof.

Fee follow-up

$14,000

collected in unpaid fees, in under a week

Education fee follow-up

AI follow-up automation

Problem
An education company needed to follow up with families who had unpaid fees.
What I built
Automated reminder calls that followed up with families and recorded their responses.
Proof
The client confirmed that the system helped collect $14,000 in unpaid fees in under a week.

Browser automation

649

student records on a government portal, in under a day

Government portal deadline

AI browser agents

Problem
The government gave a school a very short deadline to get hundreds of student records onto its portal.
What I built
AI browser agents that read the school’s sheets, checked every student against the portal, then entered, imported, and fixed records.
Proof
649 student records handled in under 24 hours. Every new entry was confirmed by the portal.

Missed-call recovery

5,664

completed calls, counted by Vapi Live

AI receptionist

Voice AI and GoHighLevel

Problem
A business owner received unwanted calls across a large GHL setup and still missed useful calls.
What I built
Call routing across 419 accounts, a Vapi receptionist, and Telegram delivery for each recording and summary.
Proof
Vapi counts 5,664 completed calls since March 2026, and the count keeps climbing.

More systems

Tap a row to open it.

Problem
An agency with hundreds of GoHighLevel sub-accounts was flooded with junk calls and still missed real leads.
What I built
One GHL IVR with keypad routing, retries, and a hand-off to the Vapi AI receptionist, loaded into every sub-account.
Proof
One published workflow runs across 419 sub-accounts.

Problem
One snapshot had to go into more than 400 GoHighLevel sub-accounts, and GoHighLevel has no way to load a snapshot into many accounts at once.
What I built
An AI browser agent that opened each sub-account and loaded the snapshot itself, with retries and a resume-safe log, instead of doing it all by hand.
Proof
The snapshot now runs in 400+ sub-accounts, every one loaded by the agent, none by hand.

Problem
Every unanswered call was a lead that might call a competitor next.
What I built
A GHL workflow that tags the caller, opens an opportunity, texts back, and sends a booking link.
Proof
Six workflow steps take a missed call to a booked appointment, using the existing pipeline and calendar.

Problem
A school with three campuses had to reach every parent about reopening, fast.
What I built
An outbound AI voice campaign that called each family, shared the update, and logged every call.
Proof
484 calls in 133 minutes with zero dial errors and about 262 live conversations.

Problem
A sales team needed one place to call leads, save outcomes, and review calls.
What I built
A browser dialer with lead lists, recordings, SMS follow-up, and on-demand AI call review.
Proof
A live platform where every reviewed call gets an AI transcript, summary, and sentiment.

Problem
A voice agent needed one reliable way to manage Google Calendar meetings.
What I built
A 14-node n8n workflow to book, reschedule, cancel, and check availability.
Proof
The working canvas documents all three scheduling paths.

Problem
Visitors leave when they can’t find an answer or a time to talk.
What I built
A Gemini AI agent in the bottom-right corner that visitors can talk to anytime. It answers questions, guides them around the site, and books calls.
Proof
Try it now: the agent in the corner of this page works the same way.

Problem
Sprout School had three fragmented Wix sites with inconsistent information.
What I built
One responsive site for four campuses, then moved the domain to Vercel without interrupting email.
Proof
Seven pages across four campuses, live on desktop, tablet, and mobile.

Problem
An autoresponder must not reply unsafely to opt-outs, unclear messages, or angry replies.
What I built
Controlled replies, suppression, no-send decisions, owner alerts, and human review.
Proof
A live reply took about 13 seconds. A separate local run passed 158 of 158 safety checks.

Problem
Inbound Gmail needed consistent routing without breaking the original thread.
What I built
A six-node n8n flow that classifies, labels, replies in-thread, and logs the result.
Proof
The canvas documents four email routes across the complete inbox flow.

AI automation across voice, browser work, CRM, email, SMS, APIs, and webhooks.

Praneeth Reddy smiling, wearing glasses and a dark t-shirt

About

One builder, start to finish

I’m Praneeth Reddy, an AI automation and integrations builder. I run Vision Growth and handle each job myself from planning through testing and handoff.

Contact

Tell me what you want to automate.

Send me the process you want to automate and the tools you use. I’ll tell you how I’d build it.

Book a 30-minute call

Book a call

Pick a time that works for you.

30 minutes on video. Choose a day and time right here.

Loading the calendar… Open it on Cal.com