Routines run by Marc

Set once. Trigger automatically. Review the work.

Tell Marc once what to prepare, where to look, and when to start. When an approved email, file, meeting, or schedule calls for it, Marc launches the Routine, staffs the right agents, and returns the work for review.

What routines are

Recurring work, prepared without asking every time.

A Routine is a repeatable AI workflow that turns approved signals into work your team can review.

Instead of asking the same question again and again, you define the work once: which sources Marc can use, what should start it, what output you expect, and which actions need approval.

Routines are built for recurring finance work: deal screening, diligence intake, IC preparation, portfolio monitoring, market updates, reporting cycles, CRM updates, legal follow-up, and inbox triage.

How it works

Define the path once. Let the run start when the signal appears.

01

Choose what Marc can use

Select the inboxes, folders, calendars, files, or workflow sources the routine is allowed to use.

02

Define the routine once

Set the objective, expected output, trigger type, frequency, source context, and approval rules.

03

Marc staffs the right agents

When the Routine starts, Marc coordinates the agents needed for research, analysis, modeling, writing, review, or deck preparation.

04

Marc returns the work

Marc returns concise work products: briefings, summaries, drafts, one-pagers, model updates, action cards, or slide materials.

05

Review, approve, or continue

You stay in control. Edit the output, reply to Marc, or continue the work in the linked run or chat.

What can start a Routine?

Start from the signals that already drive professional work.

Routines can start from the signals that already drive professional work. Triggered routines are being rolled out progressively, with user-controlled permissions and approval gates.

New email or thread activity

Turn new asks, CIMs, follow-ups, and revised documents into prepared work.

Schedules

Run daily briefings, weekly reviews, reporting prep, and market monitoring on time.

Upcoming meetings

Prepare context before ICs, board meetings, management calls, and diligence sessions.

New or updated files

React when CIMs, VDDs, models, legal docs, board packs, KPIs, or templates change.

Connected signals

Approved apps feed repeatable routines.

Outlook
Teams
Gmail
Slack
etc.
1

New signal

Approved email, file, meeting, or schedule reaches Marc

2

Marc starts the Routine

The trigger and approved context are matched

3

Agents are assigned

Marc staffs the right specialists

4

Output created

Research, model, draft, or deck work runs

5

Sent for review

Marc returns the result to your team

Routine examples

Ready-to-run work Marc can manage for you.

Professional Inbox Manager profile

Professional Inbox Manager

Prepares reply drafts with relevant context from your data and Memory, and applies labels to keep your inbox clear.

Every relevant email
Meeting Briefing profile

Meeting Briefing

Prepares a focused brief before every internal or external meeting.

Before each calendar event
CIM Review profile

CIM Review

Reviews every new CIM against your investment thesis and decision criteria.

When a CIM reaches your inbox
One-Pager Agent profile

One-Pager Agent

Turns new investment opportunities into consistent, review-ready one-pagers.

When a new opportunity arrives
Auto Deal CRM Updates profile

Auto Deal CRM Updates

Keeps CRM records current using emails, meetings, and team activity.

After every relevant interaction

Discover more ready-to-go AI agents, or ask Marc to put one to work now.

Control, approval, and auditability

Routines are proactive, but controlled.

Marc only uses the sources you approve. He can prepare work automatically, but external messages, file changes, calendar changes, and sensitive actions require human approval.

Approved context only: choose which inboxes, folders, calendars, and files each routine can use.

Clear trigger rules: define what can start a routine, what should be ignored, and when the routine should ask first.

Human review: review outputs, drafts, model updates, and deck materials. No external use without approval.

Approval gates: external emails, irreversible actions, file changes, calendar changes, or sensitive workflow actions require approval.

Run history: every Routine keeps a history of outputs, delivery status, and follow-up context in your Lampi workspace.

Pause or edit anytime: adjust the trigger, context, output format, delivery channel, or approval rules as the workflow evolves.

Turn recurring work into proactive AI routines.

Set the Routine once, let Marc run it from approved triggers, and review the finished work directly with him.