I sit on the volunteer strata council (condo board / HOA for American readers) where I live. It is a 430-unit, two-tower residential complex with a $2 million annual operating budget. I am the longest-serving member, which after four years means I am the only one left who remembers most of what happened.
In that time we have gone through 2 different property management firms, 3 property managers, and turned over most of the council more than once. And the entire thing runs on email. Outside of monthly council meetings and the yearly AGM, governance happens in the inbox. Decisions, vendor quotes, escalated resident issues, legal questions. All of it.
If you manage buildings for a living, none of this is news to you. You already live in it. What I want to describe is what it looks like from the board side, because that is the side that quietly makes your job harder, and it is the side most software has ignored.
The week it stopped being an annoyance
For a long time the email sprawl was just friction we accepted. I have a tech background, and I used to wonder why we did not just adopt something better. A council member once set up a Slack workspace and invited all of us, including the manager. Weeks went by. Not one message. Everyone went back to email, because email is not the communication tool for these organizations. It is the operating system. The inertia behind it is total, and the lowest common denominator always wins.
It stopped being an annoyance the year we onboarded a new management firm at the same moment the council almost completely turned over. Suddenly I was the only person left with any historical context, and everyone came to me with questions I mostly could not answer from memory. Each one meant twenty minutes to an hour digging through my own inbox, stitching together a timeline, and reconstructing context I had half-forgotten myself, just to send a usable answer back.
That is when it clicked that this was not a search problem. Search finds messages. It cannot tell you what was decided about the roof across three threads and four months, or who agreed to coordinate site access, or what the board understood at the time it chose one quote over another. That is synthesis, and no inbox does it.
Worse, half the time the answer was not buried. It was gone. The person who held it had left, and their inbox left with them. This is the part that should matter to you, because it is not history for its own sake. It is the history that shapes the decision in front of the board right now. When it disappears, boards re-make decisions that were already settled, miss obligations they committed to, and expose the corporation to real legal and financial risk. And you are the one who gets asked to explain why.
Why this lands on the manager
Turnover on volunteer boards is not a bug. It is how they are designed. Members rotate out, new ones rotate in, and every cycle the institutional memory thins. The new treasurer asks why the reserve study recommended what it did. The new president wants to know what was decided about the parking dispute two years ago. The owner at the AGM asks why a vendor was chosen.
The board does not hold those answers. So the questions come to you. You are managing one building among many, each with its own years of vendor threads, its own settled decisions, its own cast of people who have since moved on. When an owner asks why something was done, you need the answer, not an afternoon spent digging through three inboxes to reconstruct it.
The cost is rarely a single catastrophe. It is the steady tax of re-litigated decisions, repeated questions, and context that has to be rebuilt from scratch every time the board changes. Multiply that across a portfolio and it is a meaningful drain, and it falls on the firm.
What I built, and what it actually does
I built Dossier for exactly this problem, starting from my own council.
It is a system of record for organizations that run on email. You add a single email address unique to the organization, and Dossier captures the threads and documents shared with it. It does not ask anyone to change how they work, move to a new tool, or learn anything. If the board can forward a thread or Cc an address, it can start building a record. That constraint was not optional. Anything that asks an email-run board to adopt a new workflow dies on contact with the inertia I described above.
What you get is the ability to ask questions across years of scattered history and get an answer with its sources attached. A few examples:
How many smoking-related complaints were there in 2026, and what happened with each?
Two separate disputes involving different units. Units 215 and 115: Unit 215 complained about Unit 115’s balcony smoking in February. A bylaw notice was issued to Unit 115 on February 20. Unit 115 denied smoking, no further incidents were reported, and the file closed by April 9. Units 312 and 311: Carmen Reyes in Unit 312 complained about Unit 311 in late August. A bylaw notice was issued on August 28. Reyes disclosed her involvement and recused herself at the September 10 council meeting. The Unit 311 owner disputed the claim. After mediation, Unit 311 agreed not to smoke within 3 metres of windows or balcony openings, and the file stayed open for 60-day monitoring through February 2027.
Were any liens filed against units for non-payment of the special levy?
No liens were filed. Council directed the manager to issue final demand letters to 6 outstanding units on May 16, noting that failure to pay would result in a lien, but all units eventually paid. Unit 512 was granted a hardship payment plan, and the final payment was received June 30, 2026. No lien filing is recorded.
Every answer links back to the threads and documents it came from, so the board can inspect the source and confirm the context rather than trust a summary in isolation. Verified sources, not vibes. Synthesis, not search.
But will the inbox tools not just do this?
It is a fair question. Google and others are building AI into email, and it is genuinely useful.
But it does a different job. Inbox AI helps one person move faster through their own mailbox in a single session. It can summarize a thread or pull a fact from an old email. What it does not do is passively build a shared, multi-party record that reconstructs a decision across separate threads over months, treats decisions and commitments as their own searchable objects, and carries that record forward when the person who held the seat leaves. That is an organizational memory model, not a personal-assistant feature. It is the difference between helping someone read their mail and giving the organization a memory that outlives its members.
That last part is the whole point. The boards you manage are built to turn over. The memory should not turn over with them.
If your buildings run on email, and the history keeps walking out the door every time a board changes, this is the problem I set out to solve. You can see it at askdossier.ai.