Checks your portfolio and watchlist every morning, writes down anything real that happened, and pings Slack - but only if there's actually news.
I/O
Hand it a company and your thesis, get back a score out of ten, how sure it is, and the sources it used. The same agent runs once or a hundred times side by side - one call per company.
Downloads the whole UK company register and pulls out the ones just incorporated in your space - R&D code, university town, this year.
Two offers on the table. It works out what each one pays you at every exit size, and hands you a spreadsheet you can keep editing.
REG
When you scout founders from research papers, corporate labs sneak in - nothing in the word "Fujitsu" says company. This checks the official registry instead of guessing from the name.
You own 12%, two rounds are coming. It shows where you end up if you sit them out, and what holding your share costs in each one.
Your TVPI, DPI and IRR - plus the question nobody enjoys: what happens to them if your best company goes to zero.
Finds the week's rounds in your sector and checks each against your thesis. Shows you the rejects too, with a reason for each.
Ask what actually happened in a market. It looks the sector up inside a paid data provider, pulls the news for it, and sorts what comes back by the kind of event it is.
Reads every company in your CRM and tells you what's wrong with the data - the same firm entered twice, records with no website, records that are just a name. It proposes; it never edits anything.
Reads your calendar, decides which meetings actually need a follow-up, and leaves the drafts waiting in Gmail.
Before you automate anything: what does a thousand runs a month actually cost? Priced on real published rates, not a guess.
Who's in this market, how good their data is, what they charge - and who won't put a price on their website at all.
What three company-data vendors would cost you at a thousand lookups a month - and, more importantly, which fields each one actually gives you.
Three convertible notes turning into shares at the next round. Who got the better deal, and what the founders are left holding.
Every company in the portfolio on one screen, sorted by who runs out of money first. Profitable ones are marked as such, not given a made-up date.
Scores early-career researchers as founder material - career stage, grants, who they publish with. The score is calculated, not an opinion.
Describe it in plain text - an agent ships in 30 seconds. The library grows every week.
Build yours →