Step-by-step Claude workflows from How I AI expert interviews. Each guide includes prompts, code snippets, and practical tips.
Create context-aware guardrails for your enterprise AI agent using a 'Project' governance layer. As a project owner, define rules, scope tools, and require human-in-the-loop approval for sensitive actions, ensuring safe and effective AI use.
Design a weekly meta-workflow that analyzes your interactions with your AI assistant to make it smarter. This system learns your writing style from edits, suggests new automations for repetitive tasks, and helps you filter valuable AI techniques from hype.
Create a daily AI workflow that summarizes yesterday's meetings, identifies urgent tasks, and proactively learns internal company terms it doesn't understand. This keeps your AI assistant's knowledge current and highly relevant to your work.
Build a recurring AI task that automatically gathers updates from your calendar, Slack, and notes to suggest weekly priorities and prepare you for upcoming meetings. Start every Monday with a clear, comprehensive plan without the manual effort.
Use a hybrid AI workflow to bridge design and development. Generate high-level design concepts and tokens with Claude, then use Codex for robust implementation of the technical components in your codebase.
Log your workouts effortlessly without ever opening a spreadsheet. This workflow uses the Claude mobile app to parse a simple voice note and automatically update your workout tracking spreadsheet in the background.
Create a custom email management interface within Claude to read, triage, and draft replies without opening the Gmail tab. This workflow helps you reclaim focus and compound your learnings by training your AI on every interaction.
Build a fully automated system in Claude to generate personalized, interactive HTML proposals and client onboarding materials, saving hours on administrative work and creating a high-touch experience.
Leverage Claude Opus 5's powerful but verbose nature to create detailed front-end prototypes. This workflow focuses on asynchronous generation and blunt feedback to get high-quality outputs without getting bogged down in frustrating chat interactions.
Uncover the underlying personality and tuning of different AI models like Claude and GPT by asking a simple, three-part set of interview questions. This helps you choose the right model for your task based on its tone and directness.
Build an automated system that scours the web for job openings, uses AI to score them for relevance against a personal profile, and delivers a curated list of the best opportunities to you daily.
Gamify content creation for your entire team by launching a company-wide competition. Provide employees with AI tools to help them post on social media, driving brand awareness and attracting top talent.
A 6-step system, inspired by Alex Lieberman, that uses AI to generate ideas, extract your unique insights through an 'interview', and draft high-quality content that sounds like you, effectively eliminating 'AI slop'.
Turn a morning product brief into a queue of isolated builds, independent code review, working previews, and explicit merge decisions in Slack.
Implement a 24/7 security scanning system where a local AI model continuously scans your codebase for issues, and a powerful cloud model periodically reviews and validates the findings for a cost-effective solution.
Create a specialized AI agent to automatically investigate and triage bug reports from Sentry. This harness uses the Claude Agent SDK to gather evidence, identify root causes, and generate actionable engineering reports.
Combine objective LLM-as-judge scores with subjective human 'taste' scores to create a final, definitive model ranking. This workflow shows you how to build a weighted index that reflects your unique priorities.
Create a repeatable, automated benchmark to evaluate AI models on tasks relevant to your work. This workflow uses Claude Code to brainstorm, refine, and generate a complete evaluation system, including a test runner and scoring backend.
Automate the process of verifying and patching security bugs found by an AI agent. This workflow uses a verifier sub-agent and a patching agent to propose and test fixes, with a human expert in the final review loop.
Create an LLM-powered 'judge' to score and prioritize files in a large codebase for security analysis. This workflow helps focus expensive bug-hunting efforts on the most high-risk areas.
Create a custom AI agent that relentlessly hunts for security vulnerabilities in your codebase. This system uses a large language model with specific tools to find, test, and prove the existence of bugs.
Create an AI agent in Claude Code to automatically review aging pull requests, prompt your team for action via Slack, and even fix issues autonomously. This routine helps engineering teams ship code faster and reduce manual follow-up.
Create a printable educational worksheet with enough structure, spacing, and visual clarity for the learner who will actually use it.
Use Claude Cowork to handle the entire product return process, from finding the receipt in your email to drafting a compelling refund request. This workflow saves time on tedious administrative tasks for faulty items.
Create a personal shopping filter in Claude that understands your values, vets brands for quality, and helps you avoid low-quality, drop-shipped goods. This system finds durable, well-made items by researching brand history and materials.
Use Claude as a planning guide while Claude Code and Terminal handle the technical work required to submit a Replit app to the App Store.
Create a playable game by giving the coding agent a clear player, fun target, and test loop.
Use business records to compare time allocation with priorities and produce a strategy grounded in the underlying numbers.
Turn a detailed greenfield prompt into a runnable prototype and judge the result through real interactions.
Analyze business activity against stated priorities, then build a roadmap from the claims that survive source checks.
Build a new prototype in one pass, then test the working surface before making narrow follow-up changes.
Build a Bluetooth desk button that lights up when Claude needs permission, lets you approve the action physically, and celebrates successful runs. Claude Code handles the device protocol and firmware, even if hardware is new to you.
Turn connected Spotify, Gmail, Calendar, and Notion data into a personal dashboard that refreshes on demand. Claude chooses useful widgets, while you control the sources, layout, and visual style.
Turn a dimensionless realtor floor plan into an interactive 3D walkthrough. Claude finds a known measurement in your house documents, scales the plan, builds movable rooms, and adds furniture from purchase receipts.
Extract the real colors, type, spacing, and components from your repositories into a portable HTML design system. Reuse the file as visual context for new interfaces and component variations.
Turn one awkward section of an HTML plan into a temporary editing app. Claude builds controls for the exact data you need to change, then exports the revised rules back into your plan.
Ask Claude Code to brainstorm and plan in a self-contained HTML file with comparable ideas, mockups, risks, code excerpts, and implementation details that people will actually inspect.
Use Claude Code or Codex to build a private, interlinked learning site for a subject you want to understand. Choose the audience and depth, generate a map of the field, and keep revising the structure as your questions become more specific.
Turn an existing article into an on-brand teaching deck with Claude Design, preserving the source’s useful detail while giving it a clear visual sequence.
Import a real design system into Claude Design, generate several high-fidelity landing-page directions, and refine the strongest one through direct visual choices.
Give an agent current context, brand rules, and Memelord access so it can propose timely meme concepts for review.
Design a CLI that lets an agent complete a real SaaS setup, discover the next action from useful command output, and cross an email verification step when the user has granted the necessary tool access.
Build a flaky test skill from your own failure history, let it add genuinely new fixes to its reference material, and use each solved pattern to find related tech debt across the codebase.
Create a PR skill that explains intent from the coding session, then use a hook to route every agent generated pull request through the same team standard.
Create a powerful review agent in Claude Cowork that simulates a roundtable of advisors with different personas to pressure-test your work and get multi-faceted feedback instantly.
Use Claude Cowork connectors to analyze your sent Gmail messages and create a reusable skill that drafts emails in your unique voice, tone, and structure.
Create a persistent, context-aware workspace in Claude Cowork by setting up a project folder and a foundational 'brain' file with your personal preferences and instructions.
Add a recording mode that lets your personal AI use its real context while replacing identifying details before they reach the screen. Consistent aliases keep the demo understandable, and the underlying files remain unchanged.
Turn a recurring frustration into a small Claude Code skill by describing the problem in plain language. Let Claude ask a few questions, push it to remove manual steps, test the script, and package the working flow as a reusable command.
Capture loose tasks by speaking into an iPhone Shortcut, then let Claude Code turn the inbox, your calendar, and learned preferences into a realistic day. Large tasks become one small scheduled action, and a simple activity log helps the plan improve from what you actually do.
Create an AI agent that acts as an economic actor, autonomously paying for third-party API services to accomplish complex goals. This workflow demonstrates an agent planning a birthday party by paying for web browsing, venue search, and mailing services using Stripe.
Give one Claude Code session the builder role and another the reviewer role. The builder writes a plan on a feature branch, the reviewer challenges architecture and security, and the builder incorporates that feedback before the app is tested in Xcode.
Keep feature requests in one Claude chat that estimates build time and scores customer happiness and growth impact. The result is a living table Daniel can scan when he has a few hours to work on Commutely.
Export the states that exist in code but are missing from the design file. One prompt asks the agent to find every state in a flow, place editable versions side by side in Figma, and give designers a current canvas for refinement.
Move a running page into Figma as editable design, let the team refine it together, then give the updated frame URL back to a coding agent so it can apply the design to the local codebase.
Compress the path from live feedback to a testable code change by capturing the user’s words, structuring the issue, creating a Linear ticket, and invoking a context-rich coding agent from Slack.
Turn raw Cursor usage data into defensible adoption cohorts, a reusable analysis, and practical coaching that helps each group try the behavior most likely to deepen its use of AI.
Turn a local prototype into a shareable pull request with one `/deploy` command that handles prerequisites, branch creation, checks, GitHub setup, and CI repair in a visible loop.
Wrap Figma extraction, implementation, and browser comparison in one `/figma` command so a teammate can paste a frame link and get a close first pass that corrects its own obvious misses.
Turn a spoken product idea into a working prototype by pairing plan mode with repository context, a live preview, and browser checks the agent can run without handing every validation step back to you.
Pair Opus 4.6 as the eager builder with GPT-5.3 Codex as the rigorous reviewer: let one create the broad change, then let the other challenge its architecture, performance, and edge cases before shipping.
Use Claude Opus 4.6 for broad creative development by giving it a real codebase and ambitious brief, then correcting generic visual output with direct taste feedback before scaling the approved system.
Build a local Chrome extension that turns a focused Slack link into three to five takeaways in seconds. Joe dictates the PRD, reuses a Chrome extension skill, tests in developer mode, and fixes a raw JSON response with a screenshot.
Use one model to build quickly and a second model to review the diff for plan mismatches, code smells, and better architecture before applying the fixes.
Turn a feature idea into a user flow, timing diagram, and UI mockups that your coding agent can use as the implementation plan.
Turn meeting notes into reliable CRM records by giving Claude a tightly scoped Zapier MCP toolset and project instructions that define lookup order, field meanings, and when to create or update a record.
Give Claude a useful map of your work without loading everything at once. Store focused context in small Markdown files, connect them with profile indexes, and let short prompts pull only the files relevant to the task.
Build a research digest that searches arXiv daily and Google Scholar weekly, filters results against your own topics, and turns selected PDFs into rigorous summaries of methods, findings, effect sizes, and limitations.
Build a task system from plain Markdown files, a small metadata schema, and a `/today` command that assembles overdue work, tasks due today, active ideas, and the research digest in one view.
Speedrun a small multiplayer flight prototype by fixing the minimum scene and control contract, getting one local player working, adding an authoritative WebSocket loop in a separate workstream, and testing synchronization and failure behavior before adding polish.
Build a 3D game as a sequence of playable slices, asking the coding assistant to explain unfamiliar graphics and networking choices, measuring each new behavior, and returning to a clean checkpoint when layered fixes obscure the coordinate or state model.
Build a deliberate code analysis packet by selecting the smallest sufficient set of repository files, stating the question and constraints, and giving a reasoning model exact paths and contents without hidden context or irrelevant generated code.
Connect a local WhatsApp archive to Claude through a narrowly scoped MCP, with explicit participant privacy, retention, tool permissions, and read only defaults so message search does not quietly become unrestricted surveillance or sending.
Use a Claude Project to turn a structured hiring brief into a consistent job description, while keeping responsibilities, level, compensation, location, accessibility, legal language, and inclusive requirements tied to approved company policy and the actual role.
Give Replit Agent a complete product blueprint, build the app in working slices, and steer each feature with real role, integration, and production tests.
Use Claude to turn a real operational problem and user journey into a blueprint with roles, user stories, MVP scope, architecture, and development phases.
Prepare for a difficult stakeholder conversation by giving Claude the situation, people, evidence, constraints, and selected communication references, then use its output to rehearse questions and options without stereotyping colleagues or outsourcing judgment.
Prototype a controlled regulatory document pipeline that ingests approved deidentified clinical data, validates its structure, produces traceable draft sections and standards compliant XML, and gives medical, statistical, privacy, quality, and regulatory specialists inspectable evidence for each output.
Capture a noisy test run, have Claude group warnings by type and worst offending files, save the result as a tiered migration checklist, and let Cursor or Devin take the next bounded task whenever capacity opens.
Design a stable classification system for a collection, test it against representative items and ambiguous cases, then use Airtable AI fields to propose codes while people resolve uncertainty and protect the taxonomy from drift.
Use a Claude Project as a working notebook for a new physical business, turning questions into versioned planning artifacts while keeping financial, legal, spatial, and customer assumptions tied to evidence and expert review.
When one model remains stuck, give a second model the original request, failed output, and desired example, use its critique to identify the mismatch, and test a targeted revision instead of repeating the same prompt.
Write more controllable Midjourney prompts by separating the subject, setting, and style, then vary one photographic choice at a time and keep a prompt log that connects the selected image to its intended use and rights.
Build a more credible product prototype by defining representative structured data before generating the interface, then swap and stretch that data to expose layout and product assumptions that polished placeholder content would hide.
Create a project-specific AI workspace from approved internal context and traceable customer research, then use it to draft product artifacts without treating scraped conversations or model summaries as an unbiased source of truth.
Organize reusable business context, instructions, and small automations in a private GitHub repository so different AI tools can work from the same governed source without mixing client data, secrets, or unrelated tasks.
Design a conversational AI feature from the experience backward: write golden conversations, test them across representative scenarios, define the quality rubric they reveal, and turn the strongest flow into a clearly labeled interactive prototype.
Package a local Claude skill for the web app without carrying private files with it, fix the naming and archive structure, upload it to the intended workspace, and test it with synthetic input.
Use Claude Code inside a local skills repository to generate a task-specific skill, inspect and validate the files it creates, then test the skill on representative input and improve the reusable instructions.
Build a small meta-skill that creates other Claude skills consistently: ground it in the current format, keep the folder lean, validate the structure, and test whether it produces a useful second skill.
Use Spec Kit as a questioning partner for a substantial feature: write the first specification, answer the ambiguities it finds, and turn those decisions into requirements and acceptance criteria an implementation team can use.
Build a field research app that captures the front and back of an archival photo as a linked pair, generates a description and transcription immediately, embeds both into image metadata, and gives every file a predictable name.
Convert static customer research into clearly labeled synthetic personas whose responses stay grounded in cited source material, then use them for exploration without treating them as real customer validation.
Build a personal executive assistant that turns calendar and inbox signals into a focused daily brief, while keeping permissions narrow and every consequential action visible to you.
Join 100,000+ product managers who use ChatPRD to write better docs, align teams faster, and build products users love.