Know what to expect.

Use Feedback in the signed-in app, or the optional prompt after your brief. Your message is saved for the owner with your current step or page, app versions, release phase, and account identifier. Decision content is not attached automatically. You can request a reply at your sign-in email. If you cannot sign in, reply to your pilot invitation for help.

Read the fictional sample brief about hiring a second salesperson. It demonstrates assumptions, responses to challenges, unresolved gaps, tests, and early-warning indicators. It is an editorial example, not a customer case or a live AI response.

The current model is displayed beside the AI review controls. New reviews record the model used and generation date; you can see these in the brief. Earlier reviews without that information are labeled as not recorded. Changing the model does not automatically update saved reviews.

Participation is free and by invitation. You receive up to three analyses and 30 successful AI reviews, including re-runs and final assessments. You get a guided decision analysis, saved findings, and a printable brief with an optional final AI assessment. Reviewing all seven steps once and generating the final assessment uses eight reviews. The allowance applies across the pilot, not each month.

The pilot has no fixed end date. Access is temporary; continued free access or access after the pilot is not guaranteed. We will give advance notice before ending the pilot so you have an opportunity to export your briefs.

You will not be charged automatically. Any future paid access will require a separate purchase. Purchasing the Decision Toolkit does not include ongoing app access.

Use it for consequential professional choices: hiring, launching an offer, entering a market, choosing a supplier, changing a strategy, or committing significant resources. Start with one specific choice that you can still reconsider.

No. The workflow explains each question and gives examples. The complete guide walks through the seven steps and how to judge evidence.

A red team challenges a plan’s assumptions and reasoning. A pre-mortem asks you to imagine the decision has failed, then work backward to plausible causes. Both help expose weaknesses while you can still act on them.

The process is designed to challenge your reasoning and identify useful tests. It leaves the decision to you. Treat its suggestions as analytical hypotheses to examine, and correct or remove findings that do not fit your situation.

The review uses the information you supply. Do not treat a finding as independently verified evidence. Check factual claims against reliable sources, distinguish judgments from facts, and mark unsupported assumptions honestly. AI can make mistakes. See decision support and limitations.

Record what is unknown. Focus on gaps that could change the decision, rather than researching everything. The process helps you identify practical ways to test the most consequential assumptions.

No. “Up to date with your inputs” means the review reflects the inputs it received. Amber highlights a review that needs attention; red marks a failed review. Status tracks the work, not whether you should proceed.

Yes. Move between workflow steps to revise your answers and recheck later findings for consistency. Changes save to your signed-in workspace. Wait for “All changes saved” before leaving, then use Resume Analysis when you return. Keep the editor open if a save or sign-in error occurs.

Use the sign-in page for the available options. The site is currently in private testing, so the hosting platform may require ChatGPT access before you reach the app. The app supports emailed verification codes. General public access has not yet launched.

Yes. Request a verification code, then enter it on the app’s sign-in page using the same email address. You can read the email on any device. To enter an existing code in another browser, enter your email and select “Already have a code?” Use the latest code; if it is invalid or expired, request a new one. Linked accounts continue to share the same saved workspace.

Analyses are saved under your signed-in account. When you run an AI review, the submitted analysis is sent to OpenAI for processing. Avoid entering secrets or personal information that is not needed for the decision, and follow your organization’s rules for using AI services. Read the privacy information for storage, providers, retention, and deletion requests.

Access requires an approved email address. Each participant email receives three analyses and 30 successful AI reviews across the pilot. Failed AI attempts do not use the successful-review allowance, but do count toward temporary request limits: two attempts per minute, twelve per hour, and forty per day, with one review running at a time. While your pilot access remains active, saved analyses remain available to read, edit, and print after the allowance is used. Revoked access prevents app use while preserving saved work.

We record signed-in analysis starts, steps visited, successful AI reviews, brief views, completion marks, print-dialog opens, and toolkit clicks. Usage records are linked to your account without copying decision text. Optional role, decision type, timing, and invitation-code answers help us compare aggregate usage. You can skip, change, or clear those answers in the app. The owner-only report does not display individual decision content.

It reviews all seven steps together for contradictions, consequential uncertainties, missing safeguards, and practical tests. Generate it deliberately from the brief; it does not run just because you open the page. Each successful final review uses one review from your allowance. It preserves your findings and stated confidence, and becomes outdated when relevant analysis changes. Your detailed brief remains available if generation fails.

Use the brief’s print control. Your browser’s print dialog can print it or save it as a PDF. Review the contents first, particularly evidence labels, unresolved gaps, and warning thresholds.

Decision Red Team provides a guided review of a specific decision. The Decision Toolkit supports a broader practice of making and explaining workplace recommendations through its book, workbook, worked examples, and reusable tools. The brief links to the toolkit if you want to go further.