Your CRM ends at “Committed.” That’s exactly where the work starts.

Watch what happens at a GP the day an LP finally commits.

Eighteen months of relationship live in the CRM: who introduced them, the three objections raised in the second meeting, which partner they actually trust, the fact that their head of alternatives is on leave until March.

Then the commitment gets signed — and the record starts again. Somebody re-keys the investor into the fund system. Legal name, entity type, contacts, addresses, bank details. The eighteen months stay behind in a system the operations team doesn't open.

Two years later a new IR hire is preparing for that LP's re-up. They read a record that begins on the day the money arrived.

The problem isn't that funds lack a CRM

Every fund has one. Some run a horizontal CRM, some run a relationship-intelligence tool, plenty run a spreadsheet with conditional formatting that one person maintains beautifully.

They all work fine for the part they're built for — tracking a pipeline.

The failure is at the seam. A generic CRM models a deal that closes. A fund needs a record that keeps going for twelve years after it closes — through calls, distributions, K-1s, side letters, an AGM, a re-up, and eventually a transfer to somebody else.

So the CRM's job ends at the exact moment the relationship becomes valuable.

What we built: one record, five stages, no re-keying

The LP record starts life as a prospect and never gets replaced. It carries a relationship status — Prospect → In Discussion → Due Diligence → Committed → Passed — and a relationship owner, the person on your team accountable for it.

When the LP commits, nothing migrates. The commitment attaches to the record that already holds eighteen months of notes, the contacts, the open follow-ups, the email history. Their contacts get approved into portal users — the same people, now with logins — and everything behind them carries over.

LP	Stage	Owner	Primary contact	Last interaction	Open follow-ups
Northgate Pension	Due Diligence	Priya	J. Almeida (Head of Alts)	4 days ago	2
Ashford Family Office	In Discussion	Marcus	R. Ashford (Principal)	11 days ago	1
Beacon Endowment	Prospect	Priya	T. Okonjo (Investment Dir.)	94 days ago ⚠	0
Clearwater Insurance	Committed	Marcus	S. Lindqvist (CIO)	6 days ago	3

Look at the Beacon row. No open follow-ups, no contact in 94 days, still sitting in Prospect. That combination is the single most useful signal in a fundraise and almost no team can query for it, because the two halves live in different systems — or in nobody's head at all.

The people, not just the institution

Here's a modelling decision that sounds small and changes how the log gets used.

Every note and every follow-up can point at a touch point — the specific human at the LP it involves. Not "we spoke to Northgate." "Call with J. Almeida, Head of Alternatives."

The difference shows up when someone leaves. Your relationship isn't with Northgate Pension; it's with three people there, and when the one who championed you moves to another institution, you need to know instantly what they knew, what they asked, and what you promised them — and that they're now a warm contact somewhere else.

Contacts sit under the LP with a title, what they handle (decision maker, finance, documentation), a line of context, and a primary flag. Crucially, adding someone as a contact doesn't give them portal access or send them anything. A prospect's CFO can be in your CRM for a year before they ever get a login — and on the day the commitment closes, approving them into a portal user is one action, not a re-entry.

⚙️ Under the hood: the logging problem, and how we attacked it

Every CRM deployment at every firm in history has failed for the same reason. Not features. Nobody logs anything.

An IR team that logs every interaction has a CRM. One that logs when they remember has a directory of names and stale dates. So the engineering effort went where the failure actually is: making the log fill itself.

Email capture, three ways. File an email onto an LP from inside Gmail or Outlook without leaving the mailbox. Or use the per-record email address the platform generates and simply BCC it — the email files itself against the right record on arrival.

And when the sender isn't an exact match, a three-tier cascade runs before anyone is asked to intervene:

1. Exact contact match — the sender (or anyone cc'd) matches a known contact email under an LP.
2. Sender domain match — no exact hit, so match the sender's domain against LPs' primary email domains. @northgatepension.com finds Northgate.
3. CC domain match — still nothing, so try the domains of everyone cc'd. Catches the intro email where the LP is copied but the sender is a placement agent.

With one guard that matters more than the three tiers combined: nine generic consumer domains are excluded from domain matching. gmail.com, outlook.com, yahoo.com, icloud.com and the rest. Without that exclusion, one LP using a personal address would silently claim every consumer-domain email that ever hit the system — and a CRM that files things to the wrong LP is worse than one that files nothing, because now you distrust the whole log.

If the cascade produces exactly one LP, it files. Several, it asks. None, it asks. The system guesses only when it can be sure.

Then the follow-ups become a board. Tasks carry an assignee, a due date and a touch point, and appear on a three-column drag-and-drop board — To do, In progress, Done — filterable by assignee, LP or tag. Drag a card and the underlying task updates; tick the task complete anywhere else in the product and the card moves. One state, two views, kept in sync in both directions so the board never lies.

And an agent reads the log back to you. The Note Summarizer Agent runs across every LP, summarises the last month or quarter of notes, and writes the summary back as a note on the record — so it sits in the same timeline as everything else. It can tag untagged notes first, which quietly cleans up the log that busy people made a mess of.

📊 The impact

Before: a horizontal CRM for the raise, a fund platform for everyone who committed, and a manual re-key between them. Relationship history stops at the commitment date. "When did we last speak to them?" requires opening someone's inbox.

After: one record across all five stages, with notes and tasks attached to named humans, emails filing themselves through a 3-tier match, and a list that answers who's gone quiet in one filter instead of a partner's memory.

The number worth measuring in your own firm: what percentage of your LP interactions last quarter left a trace anyone else can find? Not in an inbox — in a place a new hire could read.

For most funds that number is under half. Everything above is aimed at it.

	Horizontal CRM	LP CRM inside the fund platform
What the record becomes at close	A closed-won deal	The investor — commitments, calls, distributions attach to it
Relationship history after commitment	Stays behind in the CRM	Continues on the same record
Contacts	CRM contacts	The same people, one action away from portal users
Email capture	Inbox sync, all of it	Filed deliberately, or matched by a cascade that refuses to guess
Knows what an LP has actually paid	No	Yes — same system as the capital account
Cost	A second licence per seat	Already in the platform

What to take from this

See it on your own structure

If you're running a raise in one system and your investors in another, the useful exercise is to pick three LPs who committed last year and try to read their pre-commitment history in the system your ops team uses today.

If it isn't there, DM me — I'll show you what it looks like when the record doesn't restart at the close.