WebinarUpcoming

How to automate support without losing the human touch

Name: How to automate support without losing the human touch
Start: 2026-09-17T18:00:00+05:30

Building an AI agent for ticket routing & escalation

About this session

Most AI projects fail at the systems layer, not the model layer, designing reliable agents, building orchestration patterns, and deploying systems that actually work at scale. This session is that systems layer, applied to one concrete problem: support ticket triage.

A session on "Support Triage That Actually Triages," building an AI agent for ticket routing and escalation. Most support-automation projects stop at classifying a ticket; this one goes further, into how a triage agent actually routes, prioritises, and escalates in a live system.

Drawn from two production systems already running: the Symprosis Support Ticket System and the Agent Tech Support System, real routing and escalation logic, not a theoretical framework, including where the automation deliberately keeps a human in the loop.

Why attend

A real, production-run case for AI-driven ticket routing and escalation, not a theoretical framework.

How the Symprosis Support Ticket System and Agent Tech Support System route, prioritise, and escalate tickets in a live environment.

What breaks when you move a triage agent from a demo to production, and how to design for it.

Where human escalation should stay in the loop, and how to design that handoff deliberately.

What we'll cover

The triage problem

Why most support-ticket automation stops at classification and never gets to real routing and escalation.

Inside a production triage agent

How ticket routing and escalation actually work in a live system — architecture, decisions, and where it deliberately keeps a human in the loop.

Live Q&A

Open floor at the end, bring questions about your own support or ticketing setup.

Systems Jigar has built

WanBuffer (Odoo)

150+ agents live in production, automating document processing, scheduling, coordination, and reporting — the transformation that took operational headcount from 55 to 20 in four months.

Symprosis

Support ticket triage — the routing and escalation logic this webinar is drawn from.

Agent Tech Support

Debugging tools for a live support system — the other production system this webinar draws from.

Backend SYNC

An internal documentation system, keeping engineering docs in sync with the codebase.

Core expertise

ReAct patternsTool designMulti-agent orchestrationMemory systems (RAG/CAG)MCPClaude APILangChainObservability & cost managementFastAPIPostgreSQLDocker

Delivered live as a walkthrough of a real production system, not a slide deck. It closes with an open Q&A, so bring questions about your own support or ticketing setup.

Speaker

Jigar Joshi

Agentic AI Architect, Corporate Trainer & Systems Engineer

Ahmedabad, Gujarat, India

Jigar Joshi is an Agentic AI Architect, corporate trainer, and systems engineer based in Ahmedabad, Gujarat. He specialises in the systems layer of AI implementation — building systems that actually work in production, rather than just LLM selection or prompt engineering.

He has architected and deployed 150+ agents running live in production (an Odoo-based system for Wan Buffer), a transformation that reduced operational headcount from 55 to 20 people over four months. He has built three proprietary production systems — the Symprosis Support Ticket System, the Agent Tech Support System, and Backend SYNC, an internal documentation system — with the first two directly in the ticket-routing and support-automation space this webinar covers.

He has trained 8+ companies, delivering 200+ hours of corporate training with a 95%+ participant satisfaction rate, and currently leads the EY Agentic AI Foundations Program (a 14-session, 40-hour course with a capstone project) alongside a fractional Agentic AI Architect engagement at Bacancy Technology and WhatsApp automation work for Plantax/GoKwik.

Connect on LinkedIn

Share this webinar

LinkedIn Facebook