How FounderFold works

FounderFold runs one method on your portfolio: each idea keeps its evidence and the CEO evaluates it when you ask, people decide what moves, projects pass measurable gates or trigger their kill criteria, the CEO agent brings up each next step, and lessons stay on record for the next search.

1. The portfolio sets the rules

A portfolio is a set of ideas and projects that share one configuration: who you are, what you want, what you rule out, how you score and how much agents may spend.

  • Profile and goals: horizon, budget, risk appetite and who decides.
  • Priorities: sectors, markets, advantages and what does not interest you.
  • Reasons to discard: knock-out questions that are enough on their own to discard an idea, with the one that applied recorded.
  • Scoring rubric: dimensions scored 0, 1 or 2 points, out of 10 in total; every point needs evidence, and a dimension without evidence counts 0. A minimum score flags an idea.
  • Sources and search rhythm.
  • Spending limits.

You don’t fill a blank form. A short conversation with FounderFold proposes each block, quotes the answer it came from, and applies nothing until an owner or editor approves it. Each approved change raises the portfolio version.

An organization can hold several portfolios — one per business unit, for example.

2. Ideas arrive with evidence, and leave with a reason

Every idea is a record with a permanent ID, a status, its evidence, a score per criterion when it has one, and its history.

Idea statuses

New, undecided
Created by an agent or a person. Agents can only create ideas here.
Under study
Being studied: the evaluation answers one concrete question, with sources.
In test
An experiment with a numeric success criterion is running.
Chosen
Turned into a project with “Turn it into a project”, with its own lifecycle.
On hold
Paused with a written condition and a review date.
Discarded
Closed with the reason to discard that applied, its explanation and family. To bring it back, “Study it again” asks for new evidence. Records are archived, never deleted.

Every claim in an evaluation carries a URL and the date it was consulted. Claims that could not be checked are marked NOT VERIFIED rather than dropped or rounded up.

New lines 2027Illustrative data — Fictional company, ideas and figures, shown for illustration only.

I-0217

Usage-based insurance for e-cargo bikes

In test
Score7/ 10Scoring version 1.0

Pay-per-kilometer insurance for delivery fleets using e-cargo bikes, sold through fleet-management software.

Score by dimension

  • Demand2 / 2
  • Willingness to pay1 / 2
  • Channel access2 / 2
  • Competitive room1 / 2
  • Strategic fit1 / 2

Reasons to discard

  • K1Within current licensesPass
  • K2No personal health data neededPass
  • K3Channel reachable within 12 monthsPass

Evidence

  • E-cargo bike fleets in the three target cities grew year on yearexample.org/fleet-registrationsConsulted Sep 14, 2026
  • Two delivery fleets agreed to a 60-day pilotInterview notes, Sep 22, 2026Consulted Sep 22, 2026
  • Comparable coverage sells for about €9 per bike per monthNOT VERIFIED
ExperimentE-0031 · Two fleet pilots. Success if at least 30% of quoted riders buy coverage within 30 days. Ends Nov 15, 2026.

3. Each project gets a lifecycle that fits its type

FounderFold starts each project from a reference lifecycle for its type of business: stages, milestones, exit gates made of measurable milestones, and kill criteria with deadlines. Gates and kill criteria are measured automatically from the project’s metrics; a metric with no data shows “not read” and is never replaced by a similar one.

Example: the six stages proposed for Halden Mutual, an insurer, and the question each one answers

  1. 01ValidationIs there demand someone will pay for?
  2. 02PilotDoes it work with real policyholders, through existing channels?
  3. 03First revenueDo customers outside the pilot buy, repeatably?
  4. 04TractionDoes it grow, and do customers stay?
  5. 05ScaleDoes it reach a size that matters to the group?
  6. 06HandoverDoes it pay its way, ready for a business unit to run, or is it time to sell?

Projects by stage

Lifecycle: new business line in an existing company · Halden Mutual

Illustrative data — Fictional company, ideas and figures, shown for illustration only.
  • Milestone done
  • In progress
  • Pending
  • Gate
  • No risk
  • Watch
  • Triggered

A project sits in the stage where it really operates. If it skipped a gate on the way, that gate is not hidden: it shows as gate debt, in red.

A triggered kill criterion never closes a project on its own. A person decides whether to stop, pivot or keep going, and the decision stays in the record with its reason.

The catalog has reference lifecycles for software, a new line inside an existing company, the search for and purchase of a company, services, commerce and physical products, and a portfolio can define stages of its own. Your portfolio configuration sets thresholds, deadlines and limits, and every change raises its version.

4. Experiments, retrospectives and lessons

Experiments test one assumption against a number agreed in advance. Retrospectives record what went well, badly or so-so. Lessons keep what was learned, with their status and date.

  • Experiment: hypothesis, numeric success criterion, deadline and result.
  • Retrospective: periodic review of a project against its gate and kill criteria.
  • Lesson: a finding from experiments or retrospectives, kept in the portfolio.

5. The next step, closed with the CEO

In a project, the next step is not a sentence lost in a document: the CEO agent brings it up on the day it is due, asks whether it was done and prepares the close for you to confirm.

  1. 1The CEO raises the step in Today: the project’s next step, an overdue experiment, a kill criterion to review or a gate with a deadline.
  2. 2You say whether it was done: fill it in yourself, move it to a new date, or have the CEO prepare it.
  3. 3The CEO states what the records show was done and proposes the next step with a date. If nothing shows, it asks you.
  4. 4You see a preview of exactly what will be written in the project record.
  5. 5You confirm and save. The previous step stays in the history with your name.

Who helps you close it

CEO
Coordinates today’s steps, priorities and closes. Every conversation starts with it.
Analyst
Status, inconsistencies and pending decisions. Answers by default in documents.
Research
Competitors, market and what is still to be verified.
Marketing
Channels, message and funnel.
Finance
Costs, price, margin, CAC, LTV and break-even.
Compliance & legal
Regulation, personal data, contracts and licenses.
Product
What to build first, the minimum product and what to leave out.
FounderFold Expert
How-to questions: where everything is in the panel and how to get more out of it.
Your own agents
Specialists you create, with their keywords and the projects they reach.

Projects in FounderFold don’t need code.

AI use this month

Illustrative data — Fictional company, ideas and figures, shown for illustration only.

38%Left of the monthly AI limit · renews on the 1st

  • Used62%
  • Left38%

Used: 62% · 159 uses

WhatOf the limitUses
P-0107 · Working-dog insurance21%58
P-0112 · Landlord rent guarantee17%41
P-0118 · Event cancellation cover9%23
The portfolio15%37
One monthly AI limit per organization. When it is reached, AI features stop until the 1st, when it renews; the rest of the panel keeps working.

6. Report to the board

Whenever you need it, or every week, month or quarter, FounderFold produces a PDF of the portfolio with its weighted pipeline value, funnel, projects and risks.

How board reports work

See it on your own portfolio.

Early access is by invitation.