Workflows
AI workflows you can follow, step by step
Each workflow states what you will have at the end, which tools it needs and the evidence that it was run. Pick one that matches the outcome you need.
All workflows
Add an AI Output Quality Gate to an Automation
A task-specific rubric, labelled fixture, threshold policy, operational gate, exception path and record of false passes, false failures and drift.
Advanced · 2–4 hours for initial design and calibration. · Tested Oct 3, 2026
Test an AI Agent’s Permissions and Escalation Boundaries
A permission matrix, approval policy, adversarial suite, observed tool-call results, tightened controls and documented residual risks.
Advanced · 2–4 hours for one bounded agent. · Tested Oct 3, 2026
Extract and Validate Structured JSON From Messy Input
A source-preserving JSON pipeline with explicit null policy, normalized values, validation errors, exception queue and field-level audit record.
Intermediate · 90–180 minutes for schema design and a representative test set. · Tested Oct 3, 2026
Build a Human-Approved Email Triage Agent
A bounded classifier, draft-only response path, escalation rules, approval queue, audit log and documented benign and adversarial test results.
Advanced · 2–4 hours for policy, configuration and a controlled test. · Tested Oct 3, 2026
Map a Manual Process Into a Resilient Automation
A process map, trigger and data contract, exception policy, reliability controls, human checkpoints and a tested recovery plan.
Intermediate · 90–180 minutes before production implementation. · Tested Oct 3, 2026
Prepare an AI-Built App for Human Handoff
A reproducible repository handoff with architecture and environment inventory, build/test evidence, security findings, known limitations and prioritized backlog.
Advanced · 2–4 hours for a small prototype. · Tested Oct 3, 2026
Turn Requirements Into a Unit-Test Matrix and Tests
A requirement-to-test matrix, generated and reviewed tests, red-green execution evidence, negative cases and an explicit uncovered-requirements list.
Intermediate · 60–120 minutes for a small module. · Tested Oct 3, 2026
Run a Risk-Based AI Pull Request Review
A prioritized review record whose findings link to requirements, changed code, evidence, verification results and a final human disposition.
Intermediate · 30–90 minutes depending on diff size and risk. · Tested Oct 3, 2026
Debug an Application From Logs Before Editing Code
A preserved reproduction, fact-only timeline, ranked cause table, minimal verified fix and explicit record of residual uncertainty.
Intermediate · 45–120 minutes depending on reproducibility. · Tested Oct 3, 2026
Turn a Product Idea Into a Reviewable PRD and Prototype
A scoped PRD, user stories, acceptance-criteria matrix, prototype and mismatch log with unresolved product decisions clearly visible.
Intermediate · 90–150 minutes for a small product concept. · Tested Oct 3, 2026
Refresh an Existing Article From Coverage Gaps
A baseline record, keep-update-remove map, refreshed sources and outline, preserved unique sections, QA log and post-update monitoring plan.
Intermediate · 60–120 minutes depending on sourcing and page length. · Tested Oct 3, 2026
Turn Customer Reviews Into Messaging Evidence
A sanitized coded dataset, reconciled theme counts, representative excerpts, segment differences, uncertainty notes and testable messaging hypotheses.
Intermediate · 75–120 minutes for a small, prepared dataset. · Tested Oct 3, 2026
Repurpose One Article Without Factual Drift
An approved source-of-truth fact sheet, channel-native drafts, a claim-difference report and a controlled release decision.
Intermediate · 45–70 minutes for one source and three derivative formats. · Tested Oct 3, 2026
Audit Search Intent and Cannibalization Before Publishing
A dated intent classification, URL overlap matrix, evidence-backed SEO action and change list with rejected alternatives and rollback conditions.
Intermediate · 45–75 minutes, excluding implementation and monitoring. · Tested Oct 3, 2026
Build an Evidence-Led SEO Content Brief
An evidence sheet and differentiated content brief with a defined audience, intent, outline, sources, internal links and acceptance criteria.
Intermediate · 60–90 minutes for research, drafting and editorial review. · Tested Oct 3, 2026
Build a Weekly Leadership Brief From Multiple Meetings
A concise leadership brief whose material claims can be traced to individual meetings and whose contradictions remain visible for resolution.
Intermediate · 30–45 minutes for four to six meeting records. · Tested Oct 3, 2026
Create Structured In-Person Meeting Notes Without a Bot
A consented, reviewed meeting record with decisions, actions, uncertainties and corrections linked to the source transcript.
Beginner · 15–25 minutes after the meeting. · Tested Oct 3, 2026
Turn a Sales Call Into a CRM-Ready Follow-Up Draft
A CRM-ready record and follow-up draft that preserve the customer’s actual words, distinguish confirmed facts from unknowns and wait for human approval.
Intermediate · 25–35 minutes after the transcript is available. · Tested Oct 3, 2026
Run a Decision-Focused Recurring Meeting
A reusable agenda and decision log that reserve meeting time for choices, preparation and unresolved work instead of status narration.
Beginner · 25–40 minutes for the first setup, then 10–15 minutes per recurrence. · Tested Oct 3, 2026
Turn a Meeting Transcript Into a Verified Action Tracker
A reviewed action tracker in which every task is tied to the meeting record and every missing owner or deadline remains visibly unresolved.
Beginner · 20–30 minutes after the transcript is available. · Tested Oct 3, 2026
Looking for reusable prompts instead? Browse prompt packs.