top of page

Agentic AI in Mumbai: A 30-Day Pilot Plan for BFSI, Pharma and Media Teams

Jul 4
5 min read

Updated: 19 hours ago

Illustration: Agentic AI in Mumbai
Agentic AI in Mumbai. Illustration.

By Parikshit Khanna, corporate AI trainer, TEDx speaker and founder of Digital Training Jet | Updated 22 October 2026


Quick answer

A safe way for a Mumbai company to start with agentic AI is a 30-day pilot: pick one repeat workflow, train the team in a one-day workshop, run the supervised agent for three weeks with a named reviewer, then measure time saved and errors caught before deciding to scale. Parikshit Khanna trains teams to run this pilot for BFSI, pharma and media workflows.

Trainer-reported reach: 3,50,000+ professionals since 2020.


Book in three steps

Step

What you share

1. WhatsApp or email

Organisation, city, team size and tools in use

2. 20-minute call

The repeat task you want an agent for and your data rules

3. Written proposal

Agenda, format, dates and one fee; confirm in writing to book



AI-generated illustration: a laptop showing workflow automation nodes
AI-generated illustration. Not a photograph.

What is new in 2026

  • Agents at scale in India: Wipro employees have built more than 29,000 of their own agents, as Infosys, TCS and Wipro scaled Microsoft 365 Copilot past 3,00,000 employees. Microsoft Source Asia, 3 June 2026

  • Maharashtra AI Policy 2026: approved with a target of INR 10,000 crore in investment, 1.5 lakh jobs by 2031 and AI training for 2 lakh youth and professionals. Inc42, 30 April 2026

  • Data protection: the DPDP Rules, 2025 are notified, so every agent design must cover what personal data it may read or store. PIB


Who books this training

  • BFSI operations and compliance support teams

  • Pharma regulatory and medical affairs support

  • Media and agency production teams

  • Finance and shared services

  • Leadership teams wanting proof before scaling


30-day pilot plan

Pilot week

What happens

Week 1

Pick one workflow, record baseline time and quality

Week 1

One-day hands-on workshop on approved tools

Weeks 2 and 3

Run the supervised agent with a named reviewer

Week 3

Fix instructions, templates and limits

Week 4

Measure time saved and errors caught

Day 30

Scale, change or stop


Agent workflows teams build

Workflow

What the agent does

Human checkpoint

Enquiry handling

Reads the enquiry, drafts a reply and logs it

Staff approve before sending

Weekly report

Collects figures and drafts commentary

Manager confirms numbers

Document intake

Extracts fields from invoices or forms

Accounts checks exceptions

Meeting follow-up

Creates actions and reminders

Owner confirms tasks

Research brief

Gathers sources and drafts a summary

Analyst verifies sources


Sample agenda

Session

What the team builds

1. Agents explained

Chatbot versus assistant versus agent

2. Pick one workflow

A repeat task with inputs, steps and an owner

3. Guardrails

Approval points, data limits and failure replies

4. Build

A supervised agent on Copilot Studio, Claude, ChatGPT or n8n

5. Test

Edge cases, wrong inputs and logging

6. Launch plan

Owner, review cadence and 30-day measures


What you get

What you get

Detail

Written scope

One agreed workflow, inputs, steps, owner and success measures

Hands-on workshop

Your team builds and tests a supervised agent design

Agent design template

Approval points, data limits, failure replies and logging

Prompt and instruction library

Saved instructions your team can reuse

30-day review

Results, fixes and the next workflow


AI-generated illustration: Claude and agentic AI training for Mumbai teams with Parikshit Khanna
AI-generated illustration. Not a photograph.

Parikshit Khanna's achievements

Achievement

Detail

TEDx speaker

Talk Redesigning Work with Artificial Intelligence, TEDxEicher School Faridabad Youth, 1 August 2026

Reach

Trainer-reported reach of 3,50,000+ professionals since 2020 across Digital Training Jet programmes

Claude credentials

Anthropic Claude Academy badges for Claude 101, Claude Cowork, and AI Capabilities and Limitations

Government and public media

Two-day Generative AI programme for Akashvani and Doordarshan teams at Prasar Bharati's NABM

IIT campuses

ChatGPT workshop at IIT Roorkee (April 2025) and Google Gemini workshop at IIT Guwahati (November 2025)

Recent corporate work, 2026

Copilot Chat workshop for GMR Airports' Delhi Duty Free (September); ChatGPT and Copilot workshop for vivo India, Gurugram (September); AI training for CP PLUS (October)

Author

Co-author of Rejection to Redirection and Digital Black


Portfolio

Sector

Selected work

Corporate and FMCG

Emami, Arvind Fashions, Sheela Foam (Sleepwell), Hetero, Tinna, CP PLUS

Real estate

Godrej Properties, Gaursons, RMZ, Saheel Properties (Pune, Phase 1 in September 2026)

Airports and electronics

GMR Delhi Duty Free, vivo India

Healthcare

A four-session AI for Doctors programme for a senior neurologist, September to October 2026

Government and media

Prasar Bharati NABM, for Akashvani and Doordarshan teams

Academia

IIT Roorkee, IIT Guwahati, Chitkara University faculty programmes, visiting faculty at GL Bajaj

International

Malabar Group, Dubai (Phase 1, live online, July 2026)


Photos from real sessions

Parikshit Khanna leading a hands-on corporate AI training session in India
Real session photo: Parikshit Khanna leading a hands-on corporate AI session.
Parikshit Khanna delivering an AI training session for the Pansari Group team
Real session photo: AI training for the Pansari Group.

These photographs are from Parikshit's real sessions; each caption states the session.


How to compare agentic AI training options

Option

Best for

What to check

Online certificate courses

Individuals learning at their own pace

Whether you build something on your own work

Automation agencies

Done-for-you builds

Whether your staff can run and change the agent

Vendor demos

Product features

Whether guardrails and governance are covered

Practitioner-led workshops, such as Parikshit Khanna

Teams that want a supervised agent for their own workflow

Ask for a sample agent design and a written agenda

No independent body ranks agentic AI trainers. Compare on relevant work, a sample agent design and a written scope.


Book agentic AI training in Mumbai

Share your organisation, city, team size, the workflow you want an agent for, tools in use and preferred dates. You get a written proposal with one recommended format and fee; dates are blocked once you confirm in writing.



Related pages


Frequently asked questions

Which workflow should we pilot first?

One that repeats weekly, has clear inputs and low risk, such as report preparation or enquiry triage.

What is agentic AI, in plain words?

An AI assistant that can take steps, such as reading an email, updating a sheet and drafting a reply, within limits you set, with a person approving important actions.

Do participants need coding skills?

No for Copilot Studio, Claude, ChatGPT and n8n basics; technical teams can go further with APIs.

What has Parikshit Khanna delivered?

Programmes for Emami, Arvind Fashions, Sheela Foam, Godrej Properties, GMR Delhi Duty Free, vivo India and CP PLUS, a two-day programme at Prasar Bharati's NABM, IIT workshops, and a TEDx talk on AI and work.

Who builds custom agents after training?

Custom builds and integrations are delivered with partners Ashesh Shah (Fusion Informatics) and Rushabh Mehta, under a separate written scope.

How do we book?

WhatsApp +91 99972 13177 or email pkhanna123@gmail.com. You get a written proposal after a short call; dates are blocked only after written confirmation.


Note: research items are from the linked sources as of October 2026 and can change. Reach is trainer-reported. Training is independent; no partnership with OpenAI, Anthropic, Microsoft, Meta or Google is implied.

bottom of page