Portfolio Management.

The default mode for most investors is reactive: founders email, investors respond. The AI-native version flips that. Three lightweight surfaces — news monitoring, full thread ingestion, and AI-digested investor updates — give the partnership a constant, low-effort read on every portfolio company between board meetings, so any partner can walk into a conversation already knowing what's going on.

1. News monitoring with team-wide visibility

The firm already runs a pipeline news feed — Feedly's Pipeline category piped through Zapier and Clay into a live Notion database, with article contents extracted and summarized so the team can scan signal instead of headlines. The portfolio version is the same pattern with the dial turned up on visibility.

• Dedicated Feedly category for portfolio companies, piped into a separate Portfolio News Notion database with the same extract-and-summarize step. One live feed the whole firm can sort, filter, and act on.
• Auto-broadcast big moments. Funding rounds, hires, launches, press hits, M&A — when the summary trips a materiality filter, the article is posted into the #portfolio-companies Slack channel automatically. No one has to be the human RSS reader.
• Weekly AI rollup. A scheduled job pulls the week's rows, asks an LLM to surface the most interesting ones, drafts a short internal brief ("here's what your companies did this week"), and sends it to the team. Each item links back to the Notion row and the source.
• Reasons to reach out. The whole point: every flagged item is a prompt for a partner to send a two-line congratulations, offer help, or just show up. The system creates the occasions; the partner provides the touch.

2. Ingest every email thread with the company

For the highest-conviction names, the firm should be able to walk into any conversation with full context — not just what the lead partner remembers from the last board call. The mechanism is a lightweight ingestion pipeline on the firm's outbound and inbound email with the company, with an AI filter deciding what's actually worth capturing.

• Trigger. A Zapier (or platform-of-choice) watcher on firm inboxes fires on new threads matching a portfolio company — by domain, by mentions of the company name, or by participant list.
• Triage. The thread is handed to an AI step — a Claude routine fired by webhook for richer reasoning, or a simple LLM call inside the automation for the easy cases — that decides: is this meaningful context, an investor update, an intro request, a fundraising signal, or noise?
• Route. Meaningful context becomes a note on the company's CRM record (Attio), tagged by type. Monthly / quarterly investor updates are filed to the company's folder in the file driveand mirrored as a CRM note so they're searchable alongside everything else. Time-sensitive items (asks, fundraising signals, escalations) also ping the deal owner in Slack.
• Why it matters. Investor updates in particular are gold — they're the founder's own narrative of the business, written on a cadence. Capturing them in a structured place means partners can help faster and downstream AI systems (the assistant, memo drafts, reference prep) can actually use them.

3. AI-digested investor updates

Once updates are flowing into a known place, layer a digest on top. Each new update is summarized into a one-screen brief — what changed on the metrics, what's working, what's blocked, what the asks are — and posted to the deal team. Across the portfolio, the same pipeline rolls up into a monthly internal read: which companies are accelerating, which are quiet, which have open asks no one has picked up. The team gets a shared memory of how each company is actually trending, without anyone having to re-read twenty PDFs.