Bring your own assistant to Paso

What this is

Your own AI, such as Claude or ChatGPT, can read your Paso pages and Paso’s guides through a connection called MCP. It is not the group’s automatic replies, and Paso does not write the answers.

What it can read

Pages Paso has cleared for your AI, the action board items Paso has cleared, Paso’s guides and the glossary below. It can only read; it cannot change anything.

Paso sees which tools are used and when, not your questions.

Connect it

Your link is on your portal page, under Connect your AI. Keep it to yourself: anyone holding it can read what your AI reads.

Claude: 1. In Claude, open Settings, then Connectors, then Add custom connector. 2. Name it Paso and paste your link. 3. In a chat, turn on Paso from the tools menu and ask your question.

ChatGPT: 1. In ChatGPT, open Settings, then Apps & Connectors, then Advanced settings, and turn on Developer mode. 2. Back in Apps & Connectors, choose Create, name it Paso, paste your link as the MCP server URL and choose No authentication. 3. In a chat, add Paso from the tools menu and ask your question.

Five questions to try

• What does Customer Segments mean, and what does my page say about mine?
• Summarise my latest page from Paso in five lines.
• Which stage of the Idea to PMF guide am I at, and what finishes it?
• What are my open action items and why is each one first?
• Where do Paso's guides explain how to test my riskiest assumption?

Glossary

Customer Segments

Who you're for - the specific target customer.

Early Adopters

The most desperate subset who'll buy first.

Problem

The top problems worth solving.

Existing Alternatives

How customers solve it today - the real competition.

UVP

The single compelling promise - why you, not them.

High-Level Concept

The "X for Y" one-liner.

Solution

The minimal solution to the top problems.

Channels

The paths to reach customers.

Revenue

How you make money.

Cost Structure

What it costs to operate.

Key Metrics

The few numbers that matter (→ Traction).

Unfair Advantage

What can't be easily copied (→ 7 Powers).

Beachhead

The one starting customer segment to win first - big enough to matter, small enough to lead, a fit with your crown jewels, and adjacent to the next.

Kill criteria

A signal, decided in advance, that says it is time to stop or pivot - a fire-able threshold, not a sentiment.

Riskiest assumption

Prioritize what’s riskiest first and ignore the rest - until it becomes what’s riskiest. The corrective to building before you’ve validated.

Appetite (time-box)

Fixed time, variable scope (Shape Up). You time-box the learning, not the build: "we’ll spend at most X days finding out whether Y is true."

Premortem

Imagine the failure from the future and work backward - a tool for developing high-quality kill criteria before you’re committed.

Resulting

Annie Duke’s term for the error of judging a decision by its outcome. A decision can be good even if the outcome is bad, and vice versa.

Diagnosis (Rumelt)

Rumelt’s missing box: name the central challenge before the guessing starts. "Without diagnosis there is no strategy - only wish lists, budgets, or hope."

Pivot

Eric Ries: a structured course correction to test a new fundamental hypothesis - change direction with one foot anchored to what you have learned. Keep the validated learning; change the strategy.

Customer Development

Steve Blank: a startup is a search for a repeatable business model, not the execution of a known one. Customer Discovery → Validation → Creation → Building - test the model in the field before you scale. GOOB is its core tactic.

Get out of the building (GOOB)

Steve Blank: there are no facts inside the building, so go outside and find them. Talk to real customers (book the calls, run the interviews) before you build - the tactic at the heart of Customer Development.

Jobs to be Done (JTBD)

Customers "hire" a product to make progress on a job in their life; the job, not the demographic, is the unit of analysis. Naming the job, and the current thing being fired to do it, sharpens the problem and the existing alternatives.

Traction

The rate at which you capture monetizable value from customers - the only thing that counts as success. Not signups, not vanity.

PMF / the 40% test

Sean Ellis’s field standard for product-market fit: ≥40% of users answering "very disappointed" if they could no longer use the product (at 30+ respondents).

Retention

Whether people keep coming back after the novelty wears off. It is the truest read on whether a product is genuinely wanted. A day-365 curve that flattens above zero says the value is real; one that decays to nothing says the early numbers were borrowed.

North-star metric

The one number that best captures the value customers get, chosen so that moving it moves the business. It is the metric a team steers by day to day, distinct from vanity counts, and downstream of traction (see Key Metrics).

Default alive / default dead

Paul Graham: on current growth and burn, does the money last to profitability before it runs out? Default alive = yes; default dead = no. Ask it early, while there is still time to act - it pairs with a kill criterion.

Feature vs company

The VC term of art "is it a feature or a company?" - could an incumbent absorb this as one feature next release? If yes it has no company-scale moat. The deeper why is the moat: a durable power, not a clever feature.

Where to Play / How to Win

Lafley & Martin: two of the five strategic choices. Where to Play (which segment / channel / axis) and How to Win (the way you beat rivals there) are chosen together - winning on the wrong axis is still losing.

Moat

The plain-language name for durable defensibility: the structural reason an advantage survives contact with competitors and time. The read is never "how deep" but which kind you are building and whether it fits your stage. Hamilton Helmer’s 7 Powers is one framework for thinking about it. A better product is not by itself a moat; the moat is the structural reason others cannot reproduce it. Defensibility was the older name for it.

Positioning

The choice of what your product is, who it is for, and what it is better than, decided before any copy is written. When a landing header or deck opener will not come together, the fault is usually here, one layer up, not in the words.

Above the fold (ATF)

The part of a page visible before the visitor scrolls. It has one job: answer "what is this and is it for me?" fast enough that they keep reading. Everything below it is for visitors who already cleared that gate.

Specificity over cleverness

When nobody knows who you are yet, ambiguity reads as noise, not intrigue. Be specific first and interesting second: a stranger should be able to repeat what you do in one sentence without getting it wrong.

Hook

The one element after the core claim that earns a second read: either a bold specific claim (a number or outcome that makes a reader think "wait, how") or a pre-empted objection. Pick one, not both.

Pre-empted objection

The first skeptical thought a reader has when they see your claim, defused in a handful of words before they can raise it ("No integrations required," "Works with your existing stack"). Use it when you have talked to enough users to know what they push back on first.

Call to action (CTA)

The button that should read like the next sentence in the story your header started, not a request for the visitor's time. If your header promises "ship faster," the CTA is "Start building," not "Request access." If you cannot write one that continues the story, the positioning is not finished.

Conversion

The share of visitors who take the one action a page is built to earn. A page that reads beautifully but does not convert has usually failed one layer up: the thinking underneath is still fuzzy, and the page reflects the fuzziness (see Positioning and Call to action).

Brand equity

The recognition and trust an audience already carries for a name. Most landing-page advice assumes you have it; a founder without it must earn every second of attention with specificity, because a visitor will not decode cleverness on credit.

Pitch-deck opener

A deck's opening slide follows the same rules as a landing above-the-fold: specific claim, one hook, clear persona, proof of mechanism. If the page header and the deck opener say different things, one of them is wrong, and often the positioning underneath both has not been pressure-tested.

Proven-Better-New (PBN)

Compose a build in order: borrow everything Proven, sharpen a Better edge, and reserve the one genuinely New thing you must bet on.

Traction narrative

The story a deck tells about momentum: not just the numbers, but the shape they make and the reason to believe the line keeps going. It stands on real traction; a narrative without the underlying rate is a wish (see Traction and PMF / the 40% test).

Warm intro

An introduction to an investor made by someone they already trust, rather than a cold approach. It is the default currency of a fundraise: the quality of the intro path often gates the meeting more than the deck does.

Flow

How customers move, in five steps, from finding you to paying, staying and bringing others. You work on one constraint at a time.

Key Assumptions

Your business model written as a list of dated claims, each marked as tested or not, with the few that the rest depend on. Running Lean calls the most important of them the riskiest assumptions.