Polaris Lab

Stress-test a North Star.

Choose a strategic context and an AI signal. These illustrative scenarios offer hypotheses to investigate. Test the suggestions against source evidence and observations in your own context before choosing Stay, Adjust, or Shift.

01 · CHOOSE A CONTEXT
02 · CHOOSE A SIGNAL
CURRENT NORTH STARSite conversion
SIGNALAgents

AI can pursue a goal across multiple steps, tools, and systems.

ASSUMPTION UNDER PRESSURE

The customer journey still happens primarily inside the brand-owned interface.

POSSIBLE SHIFT

Discovery, evaluation, and transaction can compress into an agent-mediated interaction.

STAR QUESTION

If agents increasingly complete the journey, should site conversion remain the primary measure of digital success?

SYSTEM IMPLICATION

Measurement may need to follow customer value across human and agent-mediated journeys, not only visits to the site.

Make this useful for your team: Create your North Star Brief →

A DECISION TO PRACTICE

The numbers look good.
Would you change course?

Step into a leadership meeting. Make a provisional choice, examine new evidence, then decide what deserves to change.

Fictional learning scenarios · No score or prediction · About 3 minutes

01 / YOUR FIRST READ

A retailer with a healthy dashboard.

Your team is improving website conversion. An AI shopping assistant can now compare products and complete purchases elsewhere. A successful demo has the leadership team asking whether the strategy must change.

Current North StarWebsite conversion

You can fund one small experiment this quarter. A full platform rebuild would delay committed work.

What is your provisional response?

A capability demo is a signal. It does not establish a structural shift.

Choose a response to reveal the evidence.

NEXT INVESTIGATION

Agency needs a boundary.

If agents can coordinate a larger share of the outcome, the next question is how much agency the system actually needs and where human accountability should remain explicit.

Explore the Delegation Gradient →
Research prototype · Qualitative diagnostic

How much organization has to move?

A Polaris Shift is not equally difficult for every organization. Reorientation effort depends on organizational mass: the systems, commitments, dependencies, governance, and expectations that must move with the strategy.

REORIENTATION EFFORTLOWER

Direction can often change before systems, incentives, and customer expectations become deeply embedded.

Legacy depth
Governance load
System interdependence
Sunk commitments
Customer expectations
POLARIS PRINCIPLE

The difficulty of changing direction is shaped not only by how far the Star moves, but by how much of the organization is already orbiting the old one.

This is a qualitative research prototype, not a predictive score. Company size is only one input. A large organization can reorient quickly when leadership, architecture, incentives, and decision rights are already aligned for change.

Assumption Register

Turn strategy into something falsifiable.

Every strategic choice rests on beliefs about customers, economics, interfaces, capabilities, and competition. Write the assumption down before the environment writes the answer for you.

Drafts stay in this browser
THE DISCIPLINE

Do not ask whether strategy changed. Ask which assumption expired.

TAKE YOUR THINKING WITH YOU

My North Star Brief.

Leave with a working brief for your team or organization’s next strategy conversation. New here? Start with the three questions below. You can also use a fictional scenario or bring in an assumption register.

Keep a draft in this browser if you want to return later. Saving is optional; your brief is not sent to Polaris Shift. Scenario suggestions are illustrative, not findings about your organization.

See a worked example ↓

Bringing a group together? Open the printable 25-minute discussion guide →

Start with your own question, or use a scenario above.

Help me start: three questions

Start with your own situation. Leave an answer blank if you do not know yet. Your answers fill only blank fields in the brief; nothing is sent to Polaris Shift. Add them to the brief before leaving this page, then save or download the brief to keep them.

Question 1 of 3

What outcome matters?

Describe the result you want for customers or the organization. Start with the outcome, rather than the technology.

Not saved. Download or copy your brief before leaving.

Set the scope

A team may challenge the wider direction without owning the decision to change it. Make the connection and decision boundaries explicit. These prompts are optional; unknowns stay visible.

After the test: return to the evidence

Keep your original reasoning above. Record what you learned here, then decide what should happen next. This is the learning loop after SYSTEM.

A review can support staying the course. “Still uncertain” is also a valid conclusion. This is one working review; download a copy before replacing it with a later review.

Up to 400 characters per field. Blank fields remain “Not yet established” in your brief.

Prepare a meeting summary

A compact reading of your own writing. Start the conversation with the decision, then examine the evidence and what is still unknown. This summary does not assess whether your decision is correct.

Decision scope

Not yet established.

Wider outcome

Not yet established.

People and decision boundaries

Not yet established.

Outcome we are pursuing

Not yet established.

Current working decision

Not yet established.

Assumption under discussion

Not yet established.

Evidence recorded

Not yet established.

Evidence that would change our minds

Not yet established.

Next action

Not yet established.

Owner and review point

Not yet established.

Questions to settle together

    Downloads a printable HTML document. Open it in your browser and choose Print → Save as PDF. Longer briefs may span more than one page.

    Save an editable backup or reopen a brief

    Download a backup to continue on another device or keep an earlier version. Reopen the backup here to edit it. The printable HTML download is for reading and sharing.

    Choose a backup downloaded from this tool. Files are read on your device, not uploaded. They contain your writing, so share them only when you intend to share the brief. Maximum file size: 64 KB.

    A WORKED EXAMPLE · ILLUSTRATIVE, NOT A CASE STUDY

    When the shopper sends an agent.

    Imagine a retailer whose digital team is measured on site conversion. A customer asks an AI assistant to compare products. The team’s question: are we improving the place customers buy, while overlooking how they choose?

    01 / SIGNAL

    Notice a capability. Verify a behavior.

    An assistant can compare product options. That alone does not establish customer adoption. Interview customers and observe whether they actually delegate comparison, which categories they trust, and where they take control back.

    02 / SHIFT

    Test where the journey moves.

    If meaningful evaluation happens before a site visit, a polished product page may influence less of the decision. This is a hypothesis: compare assisted and unassisted journeys before claiming the customer journey has changed.

    03 / STAR

    Question the measure of success.

    Site conversion may still be useful. But could it miss customers choosing the brand through another interface? Consider whether profitable customer outcomes across channels better express the goal. Do not replace the metric without evidence.

    04 / SYSTEM

    Make one small change observable.

    For a limited product set, improve product-data clarity and record whether assistants represent availability, price, and product fit accurately. Assign an owner and a review date. Keep purchase permissions and customer control explicit.

    Then return to the evidence.

    Support: repeated observations of customers delegating meaningful comparisons, with purchase choices influenced before a site visit.

    Challenge: customers still evaluate primarily on the site, or agent recommendations rarely affect purchase choices.

    Decision: improve execution if the journey is stable; redesign the system if access to value changes; reconsider the goal only if the existing measure no longer represents the outcome that matters.

    No outcome is assumed. Record what you learn and revisit the assumption.

    Now write the question for your organization ↑

    The Lab is a thinking environment, not a recommendation engine. Its purpose is to make strategic assumptions and reorientation effort visible before leaders automate, redesign, or reorient around them.

    Continue your course

    Do not stop at the diagnosis.

    Use the result to deepen the reasoning, watch the environment, or return to the canonical framework.