Why We Built ZeroForget: A WhatsApp Message That Started It All

Last year, I got laid off.

It was not dramatic. No warning signs I had ignored, no politics I could have navigated better. Just a restructuring email, a thirty-minute call, and suddenly a decade of context — architecture decisions, incident playbooks, the reasoning behind every infrastructure choice I had made — belonged to a company I no longer worked at.

I moved on. Started exploring what was next. And then, a few weeks later, my phone buzzed.

The Message

It was from my former manager. More than a manager, really — a mentor and a friend. Someone I had worked alongside for years, building systems together, debugging production incidents at 2 AM, arguing over architecture decisions in whiteboard sessions that ran two hours longer than anyone planned.

His WhatsApp message was simple. He needed to know about an infrastructure decision we had made together. Something about a specific configuration choice — why we had set things up a particular way, what the tradeoffs were, what would break if someone changed it.

The thing is, we had discussed this extensively. There was a whole Slack thread about it. Multiple conversations, actually, spread across months. Engineers had weighed in. We had considered alternatives. The final decision had context and reasoning that mattered.

But that was over a year ago. And now he could not find it.

The Dead Thread Problem

If you have ever tried to find a specific Slack conversation from six months ago, you know the feeling. You remember the discussion happened. You might even remember who was involved. But Slack search returns hundreds of results, most of them irrelevant. The thread you need is buried somewhere between standup updates and lunch plans.

My manager tried. He searched Slack. He checked Confluence. He asked the team. Nobody remembered the specifics. The knowledge was gone — not because it was never documented, but because it was documented in a place that made it effectively invisible.

And I realized: this was not a unique problem. This was happening everywhere, in every company, every single day.

Knowledge Does Not Leave When People Leave — It Leaves Gradually

Here is what most people get wrong about knowledge loss. They think it happens the day someone walks out the door. It does not. It starts months or years earlier, when conversations happen that never get turned into documentation. When decisions get made in Slack threads that scroll into oblivion. When the only person who knows why something works a certain way never writes it down — because they are too busy doing the actual work.

By the time someone leaves, the knowledge has been effectively lost for a long time. The departure just makes it obvious.

My manager's message made this painfully clear. I was still reachable. I could answer his question. But what about the hundreds of other decisions scattered across thousands of Slack messages? What about the next person who leaves, and the one after that? What about the knowledge that nobody even knows is missing until something breaks?

What If an Agent Could Answer Instead?

That evening, I could not stop thinking about it. Not about the specific question — that was easy to answer. About the pattern.

What if there was an AI agent that had already ingested every conversation, every document, every decision thread? What if, when my manager searched for that infrastructure decision, the agent could surface the exact Slack thread, explain the reasoning, and cite the people involved — without anyone having to remember where the conversation happened?

Not a search engine. Not another documentation tool nobody would use. An intelligent agent that captures knowledge passively from the tools teams already use, and makes it findable when someone needs it.

An agent that means when someone leaves — or gets laid off, or retires, or transfers to another team — their knowledge stays behind. Not because they spent their last two weeks in knowledge transfer meetings. But because the knowledge was being captured all along, from the everyday conversations where real decisions actually get made.

Building ZeroForget

That is why we built ZeroForget. Not from a business plan or a market analysis. From a WhatsApp message from a friend who could not find an answer that was already discussed at work.

The name says it. Zero forget. When experts leave, their knowledge stays.

We connect to the tools where knowledge actually lives — Slack, Notion, GitHub, Confluence, Google Drive, and more. We do not ask anyone to change how they work or write extra documentation. We capture knowledge as it is created, make it searchable through AI that understands context and intent, and surface it when someone needs it.

We detect bus factor risks before they become crises — identifying which critical knowledge depends on a single person. We find documentation drift, where written procedures no longer match actual practice. We connect people to the experts who can help, even when they do not know who to ask.

And yes, we built it so that the next time someone gets a message from a former colleague asking "why did we set this up this way?" — the answer is already there, waiting.

The Knowledge Is Already There

Every company I talk to has the same problem. The knowledge exists. It is in Slack threads, in Confluence pages, in Google Docs, in GitHub pull request descriptions, in email chains. It is there. But it is scattered across dozens of tools and buried under months of noise.

The challenge was never creating knowledge. People create knowledge every day just by doing their jobs. The challenge is making that knowledge findable, permanent, and independent of any single person.

That is what ZeroForget does.

Because nobody should have to text their former manager to find an answer that was discussed, debated, and decided — but never made findable.