Free framework · CC BY-SA 4.0

The PAST Framework

From random AI experiments to strategic clarity. Four questions that turn vague AI ambition into a brief you can act on.

The same four questions guide billion-dollar implementations and the single prompt you're about to type. That's why it works.

A practitioner with a drafting compass at the centre of four labelled cards arranged at cardinal points — Purpose, Audience, Scope, Tone

Why this framework exists

Walk into any organisation today and you'll see the pattern: marketing has their ChatGPT subscription, sales discovered a different tool for lead scoring, customer service is testing a chatbot, HR is screening resumes. Each team discovered AI independently, chose different tools, measured different things — and none of them are talking to each other.

That's not a technology problem. It's a strategy problem.

The question everyone asks: “How can we use AI?”
The question they should ask: “What business outcomes do we need to achieve?”

PAST is the shift from technology-first to outcome-first thinking. It's a thinking framework, not a tool checklist — so it survives the next wave of AI products, and the one after that.

The four questions

Each letter is a discipline. Run any AI decision — strategic, tactical, or single-prompt — through these four in order.

Purpose

What specific outcome are you trying to achieve?

Most AI work starts with the technology and works backward to business value. Purpose flips it. Before you evaluate any tool, you define what success actually looks like — measurable, with a deadline, owned by a person.

Prevents: Vague goals, technology-first thinking, solution-seeking-problem.

Audience

Who does the AI serve — and who receives the outputs?

There's the person paying for it, the person using it, and the person on the receiving end of what it produces. They aren't the same person. Naming all three changes what you build.

Prevents: Designing for buyers instead of users. Ignoring the people who'll have to live with it.

Scope

What's in this round — and what's explicitly Phase 2?

Scope is the discipline of saying 'not yet' out loud. Five dimensions to size up: data, users, processes, integrations, governance. A small thing done well beats a big thing half-built.

Prevents: Scope creep. Trying to solve everything at once. Pilots that die because they tried to do too much.

Tone

How should the AI align with your culture, voice, and standards?

Tone is where adoption actually lives. If the output sounds like a chatbot, your team will route around it. If it sounds like your business, they'll lean in.

Prevents: Generic AI slop. Cultural mismatch. The thing technically works but no one uses it because it doesn't sound like you.

Framework at a glance

Element Question What it prevents
Purpose What specific outcome are you trying to achieve? Vague goals, technology-first thinking, solution-seeking-problem.
Audience Who does the AI serve — and who receives the outputs? Designing for buyers instead of users. Ignoring the people who'll have to live with it.
Scope What's in this round — and what's explicitly Phase 2? Scope creep. Trying to solve everything at once. Pilots that die because they tried to do too much.
Tone How should the AI align with your culture, voice, and standards? Generic AI slop. Cultural mismatch. The thing technically works but no one uses it because it doesn't sound like you.

Works at every level

The same four questions scale up and down. That's what makes PAST useful — you don't switch frameworks when you switch altitudes.

  • Organisational strategy. Enterprise AI implementation, governance, cross-team alignment.
  • Team workflows. Department-specific process design. Marketing, sales, ops, support.
  • Individual productivity. Personal workflow enhancement. The single operator getting leverage.
  • Prompt engineering. The four questions become the structure of the prompt itself.

Quick start: the Purpose Clarity Exercise

Before any AI initiative — whether it's a six-month enterprise rollout or the prompt you're about to write — fill in this sentence:

Success means [specific measurable outcome] which will [business impact] by [timeline] as measured by [metric].

If any blank is empty or reads “TBD”, you're not ready to evaluate tools. Go back to Purpose.

Where this fits in our work

PAST is the framework behind Move 4 of the Five Moves — the discipline of running any AI request through the four questions before you type. In workshops, we make PAST the scaffolding for prompts. In client engagements, we use it to scope whether an AI build is worth doing at all.

The framework is yours to use, adapt, redistribute, and build on under the CC BY-SA 4.0 licence — same terms as Wikipedia. Attribution to the author is the only ask.

Want PAST applied to your actual work?

The framework is free. Applying it well to your specific business is where we earn our keep. If you'd like us to run your situation through PAST and hand back a brief you can act on, start with our /discover brief — ten minutes in, a Business DNA Brief back.