Why Dossier exists
I built Dossier because I watched the memory walk out the door.
Dossier did not start as a product idea. It started on a volunteer strata council, where I spent four years watching a building's hard-won decisions disappear every time someone left.
It started on a council, not in a pitch deck.
I sit on the council of a 430-unit, two-tower residential community with a multi-million dollar operating budget. Over four years I became the longest-serving member, which is another way of saying I became the only person who still remembered why most things were the way they were.
In that time we cycled through two management firms, several managers, and more than one full turnover of the council itself. Almost all of it ran on email. Outside of monthly meetings, governance happened in the inbox: decisions, vendor quotes, escalated owner issues, legal questions, all of it scattered across threads and people.
The problem was not search. It was memory.
It stopped being a background annoyance the year a new management firm started at the same moment the council almost completely turned over. Suddenly every question came to me, and answering even simple ones meant an hour of digging through my own inbox, stitching a timeline back together from fragments.
That was when it became clear this was not a search problem. Search finds messages. It cannot tell you what was decided about a project across three threads and four months, who agreed to own what, or what the board understood at the time it chose one path over another. And half the time the answer was not buried, it was gone, because the person who held it had left and their inbox left with them.
Institutional memory, the thing that makes a board's decisions defensible and a manager's advice worth more than a cheaper competitor's, turned out to be the one asset nobody had a system to keep.
So I built the system I wished we had.
Dossier is a system of record for the decisions, commitments, and history that organizations run on. You share the threads and documents worth remembering, and Dossier turns them into a sourced record you can ask in plain language, with every answer traced back to where it came from.
It was built around one hard constraint from the start: it cannot ask people to change how they work. Volunteer boards and busy managers will not adopt a new tool or migrate their history into a new system. If you can forward a thread or share a document, you can build a record. Anything heavier dies on contact with how these organizations actually operate.
Your record belongs to you.
A memory you cannot keep is not much of a memory. Your threads are never used to train AI models, only the people you invite can see your record, and you can export or delete everything at any time. The point of Dossier is to make your organization's memory something it owns and can hand off, not something locked in one more place you depend on.
Michael Zsigmond Founder, Dossier