A Book Club Changed How I Think About My Work

Most of my working life as a developer happens in a tight orbit. My team, the codebase, the next sprint. I am aware other departments exist - Finance, Tax, Audit - but they show up mostly as requirements documents and Jira tickets. Not as people with actual problems.

SSN changed that slowly, without announcing it.

SSN is our firm's social network - not LinkedIn, not a tool. It is a set of hobby groups that GT Bharat built to give people reasons to show up for something other than work. Singing, cricket, dancing, movies, books. I joined books and movies because of course I did.

The book club is about a year old now. This marks its second financial year. It is, on paper, just people meeting to talk about what they read. That is mostly what it is. But the side effect is the interesting part.

What happens when you put a developer in a room with a transfer pricing specialist

At one of the meetups, I ended up in a conversation with Dhaval Sir. He is passionate about transfer pricing in the way people only get passionate about things that have genuinely cost them sleep. He was explaining something complex - multi-entity pricing structures, documentation requirements, the gap between what regulations say and what systems support.

I was not following all of it. But something shifted anyway.

I was not hearing a requirement. I was hearing a person describe a problem that mattered to them. There is a different quality to that. Requirements arrive pre-shaped, edge cases trimmed. Real problems are messy and come with frustration attached.

Ashrut talked about consolidation - data spread across multiple spreadsheets, manual reconciliation, the kind of work that builds up like sediment and becomes invisible because it has always been there.

Gargi Patwardhan and Dinky Lakhani Ma'am - people I would usually see heads-down across the floor - had opinions, stories, specific frustrations. Tea in hand, no project context framing it. Just conversation.

The shift it made

I have spent the last couple of years thinking heavily about AI - specifically about RAG pipelines. Retrieval-augmented generation: build a knowledge base, embed it, answer questions from it. Clean pattern. I have shipped it in production.

But those conversations at SSN started pulling me in a different direction. A consolidation workflow is not a question-answering problem. A transfer pricing documentation process is not a search problem. These are multi-step, context-carrying, judgment-requiring workflows - the kind of thing an agent handles better than a retrieval system.

I started reading more about agentic architectures. Not because of a project requirement. Because someone described a problem over chai and my brain would not let it go.

That is what SSN did. It did not create a project. It created an itch.

The smallest reward

For active participation through the year, I got a diary. Physically the smallest thing on the reward spectrum. I use it.

There is something specific about a physical object as recognition. A shoutout fades. A certificate gets filed. A diary sits on your desk and you open it. Every time you do, you remember what you showed up for. It is a strange loop.

What the firm actually built

SSN is optional. It has no performance review weight, no career trajectory attached. You can entirely ignore it and nothing changes on paper.

Which is exactly why it works.

The firm created a condition - a reason for people across departments to be in the same room without an agenda. And because there was no agenda, people talked differently. Transfer pricing does not come up in a sprint planning meeting. Consolidation nightmares do not make it into an email thread. But they come up when someone has had a cup of tea and there is no deliverable on the table.

This is harder to build than any internal tool. It requires trusting that informal things have value even when you cannot measure them. GT Bharat got that right.

I am in the book club next year. And I will be better at listening this time.