This IGI Story follows young innovator Mohammad Waseem, whose work focuses on enhancing workplace productivity through AI. His goal is to eliminate time wasted on routine tasks such as reviewing emails.
Inspired by the 2023 OpenAI wave, Waseem believes that human language will become the new form of coding. He developed a customizable, no-code AI workflow built on N8N and Docker.
The system integrates AI agents and Google APIs to filter incoming Gmail messages, categorize tasks, draft automatic replies, and log new tasks in Airtable - achieving complete communication automation with one aim: to save time.
Introduction (Personal story)
Hello Shubhi! My name is Mohammad Waseem, and currently, I am working at Newton School (an EdTech Startup). It’s truly exciting to share my journey and AI use case with you. You know, ever since I graduated, I’ve been living and breathing the tech space. It all kicked off back in 2019 when I plunged headfirst into Python programming, machine learning, and data science. But honestly, my background is a bit hatke (unique).Posting to Linkedin Automation
My father was in the Air Force, so travelling was our constant - schooling happened everywhere from Jammu - Kashmir to Ghaziabad. That nomadic life was a boon, though; it meant I never had trouble collaborating or interacting with people from different cultures.
Funnily enough, I dreamt of being a pilot! But kismat (fate) had other plans. Right after graduation, I tore my ACL. I thought, "Okay, let's give the computer world a rest," but then I realised, chalo, let's pursue this. I completed my Computer Science Engineering in 2017. Back then, dedicated AI and data science courses weren't really around.
Initially, I focused on content marketing - writing blogs and creating content pieces. Slowly, that evolved into leadership roles where I managed content plans and journeys. Later, I deep dived into Natural Language Processing (NLP) and then Computer Vision. What really fascinates me is this: I was never a very serious engineer in college, but I wasn't mediocre either. When I landed a small freelancing stint creating technical content, and I saw how much people benefited, that's when I knew: I want to help people learn through good content creation. That’s how my path led me through academic freelancing, IDA, Intensive Learning, Intellipaat, and finally, here at Newton School.
Problem Overview
My focus area is all about AI productivity at the workplace. Think about it: how much time do people waste on repetitive, mundane tasks? Take leadership roles, they spend ages just reviewing every email and drafting replies. Or consider the small business owner who needs to find potential buyers using Google Maps API and draft cold outreach emails. They often can't afford to hire a lead generation team.
The core problem I wanted to tackle was straightforward: How do we automate these time-consuming tasks? We are trying to solve the issue of excessive time spent on predictable, repeated tasks that prevent individuals and teams from focusing on strategic work. My ultimate goal was to save time as much time as possible.
Data Analysis and Reporting Workflow
Email Automation Workflow
Linkedin Post Creation
Posting to Linkedin Automation
Motivation
The big wave started in 2023 with OpenAI. I saw the power, but I also saw the barrier. If someone wanted to build AI agents, they needed to learn NLP, how transformers work, and how to create and reference APIs. For a generalist, that’s a massive hurdle; they’d lose motivation halfway.
That’s when I started exploring no-code AI tools. Tools like Bolt, Lovable, orchids showed me that you can create a working app or website doing exactly what you want, without writing complex code. This removes the barriers for anyone who wants to consult on ads, draft pitches, or create solutions, even if they aren't associated with a big organisation.
This realisation was huge: Human language, specifically English, will be the new coding in the future. They are making coding so easy that people who don't know the trade - even someone in a tribal area - can create beautiful applications using their language. My motivation stemmed from empowering those people to save time, using simple prompts.
Solution Overview - The AI Use Case
My solution is a specialised workflow designed to enhance productivity using AI agents. I call it AI productivity at the workplace.
The workflow targets anyone who handles a team or works extensively with Google Sheets, which is basically everyone in the corporate world. It allows organisations to automate their mundane tasks, saving time which can then be redeployed for other valuable work, not necessarily cutting staff.
This solution is built around a tool called N8N, which runs the workflow. While it can be hosted on the cloud for a subscription, it can also be used for free if you host it on Docker. The beauty is that it leverages large language models (like GPT, a transformer). Anyone can follow the workflow, customise it, and just differentiate the prompts based on what they want from the system. Essentially, it's a customizable automation engine for office life.
Technical Details - How AI is Actually Utilised
Let's get into the technical jugaad (fix). I used Docker to host the N8N component, which is the core workflow engine. My workflow is driven by Google Suite APIs, covering tools like Google Sheets and Gmail.
The entire automation is set up using triggers: I started with manual triggers, but then moved to a Gmail trigger. When an email hits my inbox, it immediately goes through a filter component. Here, I decide which emails are important based on predefined filters.
Next, the email hits another node, which analyses its purpose - is it a follow-up? A new task? Or an irrelevant/other task?. Based on this categorisation, the process splits into three different "gates," or nodes.
In these nodes, the AI analyses the subject line and body of the email. For example, if it's a new task, the system registers it in a database, which in my case was Airtable. If it’s a follow-up, it automatically drafts a prompt reply saying something like, "I will reach out to you". By the end of the day, my entire Gmail communication is automated. I know exactly what my new tasks are, what follow-ups are needed, and what replies have been sent - it's fully automated.
Challenges & Failures
I believe challenges are where the real learning happens!
In the email automation task, the major headache was the AI agent taking unnecessary context. The agent would assume things and draft replies based on those assumptions, which is dangerous. I had to strictly coach the agent - like a student - to reply only based on the context provided in the email or the specific knowledge base, and absolutely not to assume context. This failure taught me how to create great prompt templates.
Outside of email, I faced a mammoth challenge in video generation: character consistency. If you are creating ads or stories, you need the character to look the same across all frames, right? This is still tough. Plus, the credit system is a killer. Trying to use powerful tools like Sora 2, even for an 8-second video, demands way too many credits. For an individual like me, that's very challenging.
My solution for consistency was using tools like Freepik, which has a proprietary model. I create the character there first, and then use that consistent character as a reference across subsequent frames in different models.
AI Readiness Skills
This whole experience reinforced the core idea that technical knowledge is useless if you don't know how to talk to the AI. You have to articulate instructions beautifully. It's like having a brilliant student; if you give them the right instructions, they will do exactly what you want.
Beyond basic articulation, we need niche-based communication—understanding technical terms relevant to your domain (like tokens, inference, and bylines). This is why foundational skills—like Online Learning, Critical Thinking, and Collaboration—are vital before jumping into AI tools. We need that output clarity that comes from manual experience.
Ques- Let’s understand how good your Collaboration skills are – Name the persons or organisations you collaborated with for making this project successful.
This project wasn’t built in a vacuum. I’ve had amazing mentors who acted as my big motivating and guiding forces throughout my career.
The key people I collaborated with and who supported this journey are: Sourabh, Hemant Sharma, Nishant Chandra and Gyan Arora.
They truly enabled me to think differently.
Ques - Let us know how good your ‘Critical Thinking & Creativity’ is – Do you think your solution is the best one for the stated problem? If yes, then why? If not, what is the best solution?
Arre nahi, bilkul nahi! (Absolutely not!). I definitely don't think my solution is the best one for the stated problem.
My philosophy is to break every problem down into small milestones first and then achieve the larger picture. My original, ideal problem statement was just to save time. That’s it. Automating the email workflow was just one-use case I found to address that goal.
The "best" solution is always evolving, but it must start with that fundamental question: how do we save time? And that requires constant re-evaluation and finding better, faster automation methods than the ones I deployed today.
Ques- Let’s test your Online Learning skills – What self-learning did your team do to understand the problem better, or design the solution, or for any other aspect of the project? Also, please specify the sources (if possible) from which you learnt.
Self-learning is the oxygen of innovation! I consumed tons of literature, especially focusing on the documentation of the specific tools I was using, like understanding how to create APIs.
A major moment was seeing an example of automated AI interviews online at Mercor
(I think it was around 2023). That gave me the idea that this kind of complex interaction could be automated efficiently.
For regular learning, I rely heavily on online sources from YouTube (specifically following people who have credibility). Andrew NG (via Coursera and Deep Learning AI), Hugging Face, For video and ad creation, I often refer to Freepik; they have an excellent AI suite and launch videos showing you how to build things.
Ques - Let’s test your Online Reputation skills – If you have made any effort to popularise your solution on the Internet, please share the relevant links.
Yes, I have definitely leveraged this for personal branding! I created an automation for personal branding on LinkedIn and also developed a personal tracker. I can share those workflow links. (Note: Specific URLs are not provided in the sources, but the effort is documented).
Impact so far & vision for the future
The immediate impact of this workflow has been internal: It was developed specifically to run a workshop that is part of an upcoming online offering. The primary measure of success right now is the readiness of the system, using evaluation metrics like success rate and the number of true positives.
For the future, the vision is clear: We are taking this offering forward. In a few months, the entire online curriculum and offering will be ready, and we plan to run dedicated cohorts. This is about equipping generalists with the power of no-code AI.
Your advice to fellow innovators
My advice to aspiring AI innovators in India is this: Don’t just talk about it. It’s very easy to say 'AI' right now.
Just start something. Pick up any use case - I don't care if you call it a use case, just try it out. Be constantly on the lookout for something new and better. Leverage the models that make it easy, like the agent kits of OpenAI (KITS) and tools like N8N. You only need time, patience, and practice. Just do experiments.
Mentor, Support & Inspiration
As mentioned, I had incredible mentors and guiding forces.
But in terms of pure inspiration, I am deeply influenced by the work of Varun Mayya. I follow him very closely, especially what he's doing in the AI segment. I also admire the work of Vishnu and Nikhil, who specialise in voice AI agents. Vishnu, in particular, gets first access to Google models and constantly posts really cool stuff online, which is very motivating. It’s inspiring to see people constantly innovating and pushing the boundaries of what AI can do.
