After the round closes
Every VC tool on the market wants to talk to you about deal flow. Screening, scoring, pipeline dashboards, all built around the moment before you write the check. Almost none of them have anything to say about the moment after.
That's backwards. At Sapphire Capital Partners, where we manage over 40 sector-focused venture funds, the screening decision takes weeks. The monitoring relationship runs for years. And it's the part that quietly falls apart first.
The monitoring gap
Here's the pattern we kept seeing before we built our own answer to it. A fund invests. For the first quarter or two, someone on the team emails the founder for an update, gets a reply, drops the numbers into a spreadsheet. By quarter four, the founder is heads down, the analyst who owned that relationship has moved on to new deal flow, and the update either doesn't happen or happens six weeks late with no one chasing it.
None of this is a discipline problem. It's a design problem. Manual monitoring has no default cadence, no owner once the deal team's attention moves on, and no single place that shows a partner "here's every portfolio company's actual current status" without someone assembling it by hand first.
Why ad hoc doesn't scale past a handful of companies
A fund with six portfolio companies can run this on memory and goodwill. A fund with sixty can't. Every additional company adds its own reporting rhythm, its own founder who's more or less responsive, its own thread in someone's inbox. There's no fund-wide view, because there's no fund-wide system, just a pile of individual relationships that each depend on someone remembering to follow up.
The fix isn't "remind people more." It's removing the dependency on any one person remembering at all.
What a real cadence actually requires
The way we built this into DealsFlow starts with the fund setting a schedule once: how often (monthly, quarterly, whatever fits the fund), a start date, and a grace window for late updates. From there, it runs as batch rounds. Every deployed company that hasn't reported for the current round gets invited together, on the same day, rather than each company drifting onto its own personal timeline based on when it happened to be invested.
That matters more than it sounds like it should. Per-company drift is how monitoring becomes unmanageable, because eventually you have sixty companies all due on sixty different days and no one has a clean answer to "what's the state of the portfolio right now."
Making it easy enough that founders actually do it
A monitoring system only works if the founder on the other end actually fills it in. So the invitation goes out as a signed link, no account to create, no password to remember, valid for months rather than days so a founder who's traveling or mid-raise isn't locked out. They submit once, it's recorded, and if they submit again it doesn't create a duplicate mess for anyone to clean up.
This sounds like a small detail. It's the difference between a 90% response rate and a 40% one.
Turning an update into a signal, not just a data point
Once a founder submits an update, the point isn't just to file it away. A single pass reads the new data alongside the last few periods of history and produces a short read: a survival estimate, a trend direction, a plain-language summary, and the key risks worth a partner's attention. It's not a black box score dropped on a portfolio screen. It's meant to be read in thirty seconds and understood.
And it's not a verdict. It's a narrated read of the data, sitting next to the numbers, for a human to weigh alongside everything else they know about that founder and that market. The system doesn't decide anything. It just makes sure someone doesn't have to reconstruct the picture from scratch every quarter.
Keeping a human on every valuation
An admin reviews each update as it comes in and can record a valuation against it. From there, the portfolio view rolls everything up in one place: latest valuation, survival read, next due date, and who's overdue, sorted so the companies that need attention actually surface instead of getting buried between the ones that are fine.
Catching what falls through the cracks
The other half of monitoring is knowing when something didn't happen. If a company goes past its grace window without reporting, the fund owner gets notified, once, not every day, not buried in noise, but not silently dropped either. If a scheduled check happens to miss a day, the notice still goes out later rather than vanishing. Overdue doesn't mean forgotten.
Where this feeds back into how you screen the next deal
Every company that eventually exits monitoring, whether that's a strong exit, an acquisition, or a company that didn't make it, becomes part of the fund's own outcome history. That's not just recordkeeping. It's the raw material that lets a fund eventually compare a brand new applicant against its own past decisions instead of a generic benchmark, which is a piece we'll go into on its own another time. Monitoring done properly isn't the end of the pipeline. It's what makes the next round of screening smarter.
The practical test
If you're evaluating whether a monitoring setup will actually hold up, ask one question: what happens on the day nobody remembers to check? If the honest answer is "nothing, until someone notices three months later," that's the gap worth closing before you scale the portfolio any further.
See how this runs without a spreadsheet. Book a walkthrough and we'll show you a live portfolio round end to end.
Frequently asked questions
Why do VC portfolio monitoring programs fall apart after the first couple of quarters?
Because most run on individual memory rather than a system. There's no default cadence, no single owner once the deal team's attention shifts to new pipeline, and no fund-wide view, so updates depend entirely on someone remembering to chase them.
How often should a VC fund request updates from portfolio companies?
It depends on the fund, but the specific interval matters less than having one fund-wide cadence that every company is held to, rather than each company drifting onto its own personal schedule based on when it happened to close.
Do founders need to create an account to submit a monitoring update?
No. A well-designed process sends a signed, time-limited link by email that lets the founder submit without registering anywhere, which is a meaningful factor in whether updates actually come back on time.
What should an AI-generated portfolio read actually contain?
A short, plain-language summary: a survival estimate, a trend direction, and the key risks, generated from the new update in the context of the company's recent history. It should be something a partner can read in under a minute, not a score with no explanation behind it.
Does AI decide anything in a monitoring workflow like this?
No, and it shouldn't. The AI narrates what the data shows. A human reviews the update, records the valuation, and makes any judgment calls. The system's job is making sure that human has the full picture without reconstructing it by hand.
What happens when a portfolio company misses its update deadline?
After a defined grace period, the fund should get exactly one notification, not a flood of repeated reminders, and not silence. If a scheduled check is missed for any reason, the notice should still go out later rather than the miss going unrecorded.
How does portfolio monitoring connect back to deal screening?
Every company that reaches a final outcome, whether it survives, exits, or fails, adds to the fund's own track record. Over time that history is what lets a fund evaluate new applicants against its own past decisions rather than generic, market-average scoring.