How to Build AI Workflows With Claude Code  🔗 **Source:** [How to Build AI Workflows With Claude Code](https://www.producttalk.org/how-to-build-ai-workflows-with-claude-code/) > [!note]+ Document Notes > ```yaml > tags: [AI, workflows, product-management] > author: Teresa Torres > source: Product Talk > reference: [https://www.producttalk.org/how-to-build-ai-workflows-with-claude-code/] > date created: 2025-12-11 12:00 > ``` > # 🏷️ Title: How to Build AI Workflows with Claude Code (Even If You're Not Technical) <!-- deep-read:sections start --> ## 🧭 Table of Contents - [📝 Summary](#-summary) - [📖 Main Content](#-main-content) - [🗒️ Notes](#️-notes) - [📚 References](#-references) - [🗣️ Paraphrase](#️-paraphrase) - [📰 Newsletter](#-newsletter) - [✅ Action Items](#-action-items) - [🧪 Hypotheses](#-hypotheses) - [📊 Methodology](#-methodology) - [❓ Questions](#-questions) - [🔮 Future Research](#-future-research) - [💡 Implications](#-implications) - [🧠 Learning Styles Assessment](#-learning-styles-assessment) - [📊 Accessibility Assessment](#-accessibility-assessment) - [🌍 DEIB Assessment](#-deib-assessment) - [✅ Logical Validity](#-logical-validity) - [🛡️ DARVO Analysis](#-darvo-analysis) - [🎯 Conclusion](#-conclusion) ### 🔗 References - Include links cited in or relevant to the article. --- ## 📝 Summary - Torres describes how she transformed a “tortured” writing process into structured AI-assisted workflows using Claude Code. - It matters because it shows non-technical professionals how to design practical AI workflows by mapping their own tasks. - Stakeholders: product managers, writers, knowledge workers, and AI tool builders. - Risks: over-reliance on LLMs, weak fact-checking, poorly scoped automation. - Bottom line: treat AI workflows like product solutions—map steps, pick where AI helps, prototype, then iterate. --- ## 📖 Main Content - Torres starts with her messy writing routine: brain-dumps, half-finished outlines, and lots of back-and-forth. Now, she uses Claude as a collaborator to review outlines, spot gaps, challenge ideas, suggest paywall placement, brainstorm headlines, research, and help with SEO, which significantly boosts her output ([opening description](https://https://read.readwise.io/read/01kc4dpchpvkjck9tzq18c9xbt?endLexposition=0000G_00000-000Z&startLexposition=00000&emphasizeText=Claude%20reviews%20my%20outline%20and%20helps%20me%20fill%20in%20gaps)). - She emphasizes that her workflow emerged iteratively by repeatedly asking “How can Claude help with this?” across parts of her work, not in a single grand design session ([iteration point](https://https://read.readwise.io/read/01kc4dpchpvkjck9tzq18c9xbt?endLexposition=0000S_00002-003d&startLexposition=0000I&emphasizeText=I%20did%20it%20iteratively.%20It%20all%20began%20when%20I%20started%20asking%2C%20%22How%20can%20Claude%20help%20with%20this%3F%22)). - She connects this to an ongoing series on Claude Code and personal productivity, clarifying she is an enthusiastic user, not a paid promoter, and commits to FTC-compliant disclosure for any future sponsorships ([disclosure](https://https://read.readwise.io/read/01kc4dpchpvkjck9tzq18c9xbt?endLexposition=0000S_00002-003d&startLexposition=0000I&emphasizeText=I%20have%20not%20received%20any%20compensation%20from%20Anthropic%20for%20writing%20this%20series)). - Core method: first, “Map Out What You Do to Complete the Task.” She demonstrates with her blogging workflow: pick a topic, brain-dump, outline, research gaps, write sections, edit, consider SEO, brainstorm headlines, choose images ([step list](https://https://read.readwise.io/read/01kc4dpchpvkjck9tzq18c9xbt?endLexposition=0000k_00003_00000-000Y&startLexposition=0000U&emphasizeText=Here%27s%20what%20I%20do%20when%20I%20write%20a%20blog%20post%3A)). - She likens this to story mapping: instead of mapping a customer journey, you map your own steps to see where AI may augment or automate; this clarity lets you ask “Where can AI help?” in a focused way. - From that map, she outlines a multi-step design process: choose one step to target, choose an approach (traditional code vs LLM), prototype a workflow with detailed instructions, then test and iterate toward something reliable ([forward plan](https://https://read.readwise.io/read/01kc4dpchpvkjck9tzq18c9xbt?endLexposition=0000k_00003_00000-000Y&startLexposition=0000U&emphasizeText=From%20here%2C%20we%20are%20going%20to%3A)). - The free section ends by promising enough guidance for readers to leave with a first-draft AI workflow and then transitions to a paywalled, step-by-step build-out ([transition](https://https://read.readwise.io/read/01kc4dpchpvkjck9tzq18c9xbt?endLexposition=0000o_00000-002M&startLexposition=0000m&emphasizeText=My%20goal%20is%20to%20give%20you%20enough%20guidance%20that%20you%20can%20follow%20along)). Uncertainties: the paywalled portion likely contains specific Claude Code prompts, implementation screenshots, and examples that aren’t visible here. --- ### 🗒️ Notes - Torres contrasts her old “brain dump → half-outline → stuck → rewrite” loop with a new conversational approach where she “discusses [her] way to what [she] think[s]” with Claude as a thinking partner. - Claude is used as sounding board, critic, researcher, SEO coach, thesaurus/dictionary, and editorial assistant, even advising on newsletter paywall placement for paid vs free content. - Workflow design was emergent: she repeatedly identified friction in her process, then asked how Claude could help that specific friction point. - This article sits in a larger series on using Claude Code to improve personal workflows; the prior installment covered picking which workflows to automate or augment. - She is explicit about not being compensated by Anthropic and commits to disclose any future material relationships, aligning with ethical and regulatory expectations. - Key technique: thoroughly map your current manual process step-by-step (akin to a personal story map) before introducing any AI—this exposes natural seams where AI can support or replace sub-steps. - Her blogging process includes ideation, outlining, research, drafting, editing, SEO, title exploration, and image decisions; each of these becomes a candidate zone for an AI workflow. - She frames AI workflow design as product discovery: understand the “job to be done,” map the journey, pick a focused slice, prototype minimally, then refine based on real use. - Next steps promised: select a single step to focus on, decide whether simple code or LLM tooling is more appropriate, prototype in Claude Code, and iteratively test/adjust prompts and structure. - The article is positioned as a guided exercise: by following her example with your own task, you should end up with a first-draft AI workflow tailored to your work. --- ## 📚 References - Product Talk – “How to Build AI Workflows with Claude Code (Even If You're Not Technical)” – https://www.producttalk.org/how-to-build-ai-workflows-with-claude-code/ - Claude & Claude Code by Anthropic – https://www.anthropic.com - Teresa Torres on Continuous Discovery Habits – https://www.producttalk.org/continuous-discovery-habits/ --- ## 🗣️ Paraphrase - Torres explains that instead of wrestling with messy drafts and outlines, she now collaborates with Claude to refine her ideas, structure content, and handle surrounding tasks like research and SEO, which significantly increases her writing throughput. - She argues that the best way for non-technical people to build AI workflows is to first map their existing process step-by-step, then progressively introduce AI support to specific steps, treating workflow design like product discovery and iterating toward a dependable system. --- ## 📰 Newsletter In this week’s productivity spotlight, we’re looking at how product discovery expert Teresa Torres designs AI workflows with Claude Code—no CS degree required. Her approach is highly relevant for WordPress pros: map your content or development process, then let AI assist with outlines, research, SEO, and even paywall strategy. If you’re building with WordPress, this maps neatly to: - Content workflows in the Block Editor (draft → refine → optimize). - Developer workflows for plugin docs, changelogs, and release notes. Helpful resources: - WordPress + AI tooling overview: https://wordpress.org/plugins/tags/ai/ - Claude & Claude Code details: https://www.anthropic.com - Content strategy wit <!-- deep-read:sections end --> [[3. Resources/Readwise/Full Document Contents/Articles/How to Build AI Workflows With Claude Code|Full Document Contents]] ## Document Notes ```yaml tags: #AI #workflows #product-management author: Teresa Torres source: Product Talk reference: [https://www.producttalk.org/how-to-build-ai-workflows-with-claude-code/] date created: 2025-12-11 12:00 ``` # 🏷️ Title: How to Build AI Workflows with Claude Code (Even If You're Not Technical) ## 🧭 Table of Contents - [[#📝 Summary]] - [[#📖 Main Content]] - [[#🗒️ Notes]] - [[#📚 References]] - [[#🗣️ Paraphrase]] - [[#📰 Newsletter]] - [[#✅ Action Items]] - [[#🧪 Hypotheses]] - [[#📊 Methodology]] - [[#❓ Questions]] - [[#🔮 Future Research]] - [[#💡 Implications]] - [[#🧠 Learning Styles Assessment]] - [[#📊 Accessibility Assessment]] - [[#🌍 DEIB Assessment]] - [[#✅ Logical Validity]] - [[#🛡️ DARVO Analysis]] - [[#🎯 Conclusion]] ### 🔗 References - Include links cited in or relevant to the article. --- ## 📝 Summary - Torres describes how she transformed a “tortured” writing process into structured AI-assisted workflows using Claude Code. - It matters because it shows non-technical professionals how to design practical AI workflows by mapping their own tasks. - Stakeholders: product managers, writers, knowledge workers, and AI tool builders. - Risks: over-reliance on LLMs, weak fact-checking, poorly scoped automation. - Bottom line: treat AI workflows like product solutions—map steps, pick where AI helps, prototype, then iterate. --- ## 📖 Main Content - Torres starts with her messy writing routine: brain-dumps, half-finished outlines, and lots of back-and-forth. Now, she uses Claude as a collaborator to review outlines, spot gaps, challenge ideas, suggest paywall placement, brainstorm headlines, research, and help with SEO, which significantly boosts her output ([opening description](https://https://read.readwise.io/read/01kc4dpchpvkjck9tzq18c9xbt?endLexposition=0000G_00000-000Z&startLexposition=00000&emphasizeText=Claude%20reviews%20my%20outline%20and%20helps%20me%20fill%20in%20gaps)). - She emphasizes that her workflow emerged iteratively by repeatedly asking “How can Claude help with this?” across parts of her work, not in a single grand design session ([iteration point](https://https://read.readwise.io/read/01kc4dpchpvkjck9tzq18c9xbt?endLexposition=0000S_00002-003d&startLexposition=0000I&emphasizeText=I%20did%20it%20iteratively.%20It%20all%20began%20when%20I%20started%20asking%2C%20%22How%20can%20Claude%20help%20with%20this%3F%22)). - She connects this to an ongoing series on Claude Code and personal productivity, clarifying she is an enthusiastic user, not a paid promoter, and commits to FTC-compliant disclosure for any future sponsorships ([disclosure](https://https://read.readwise.io/read/01kc4dpchpvkjck9tzq18c9xbt?endLexposition=0000S_00002-003d&startLexposition=0000I&emphasizeText=I%20have%20not%20received%20any%20compensation%20from%20Anthropic%20for%20writing%20this%20series)). - Core method: first, “Map Out What You Do to Complete the Task.” She demonstrates with her blogging workflow: pick a topic, brain-dump, outline, research gaps, write sections, edit, consider SEO, brainstorm headlines, choose images ([step list](https://https://read.readwise.io/read/01kc4dpchpvkjck9tzq18c9xbt?endLexposition=0000k_00003_00000-000Y&startLexposition=0000U&emphasizeText=Here%27s%20what%20I%20do%20when%20I%20write%20a%20blog%20post%3A)). - She likens this to story mapping: instead of mapping a customer journey, you map your own steps to see where AI may augment or automate; this clarity lets you ask “Where can AI help?” in a focused way. - From that map, she outlines a multi-step design process: choose one step to target, choose an approach (traditional code vs LLM), prototype a workflow with detailed instructions, then test and iterate toward something reliable ([forward plan](https://https://read.readwise.io/read/01kc4dpchpvkjck9tzq18c9xbt?endLexposition=0000k_00003_00000-000Y&startLexposition=0000U&emphasizeText=From%20here%2C%20we%20are%20going%20to%3A)). - The free section ends by promising enough guidance for readers to leave with a first-draft AI workflow and then transitions to a paywalled, step-by-step build-out ([transition](https://https://read.readwise.io/read/01kc4dpchpvkjck9tzq18c9xbt?endLexposition=0000o_00000-002M&startLexposition=0000m&emphasizeText=My%20goal%20is%20to%20give%20you%20enough%20guidance%20that%20you%20can%20follow%20along)). Uncertainties: the paywalled portion likely contains specific Claude Code prompts, implementation screenshots, and examples that aren’t visible here. --- ### 🗒️ Notes - Torres contrasts her old “brain dump → half-outline → stuck → rewrite” loop with a new conversational approach where she “discusses [her] way to what [she] think[s]” with Claude as a thinking partner. - Claude is used as sounding board, critic, researcher, SEO coach, thesaurus/dictionary, and editorial assistant, even advising on newsletter paywall placement for paid vs free content. - Workflow design was emergent: she repeatedly identified friction in her process, then asked how Claude could help that specific friction point. - This article sits in a larger series on using Claude Code to improve personal workflows; the prior installment covered picking which workflows to automate or augment. - She is explicit about not being compensated by Anthropic and commits to disclose any future material relationships, aligning with ethical and regulatory expectations. - Key technique: thoroughly map your current manual process step-by-step (akin to a personal story map) before introducing any AI—this exposes natural seams where AI can support or replace sub-steps. - Her blogging process includes ideation, outlining, research, drafting, editing, SEO, title exploration, and image decisions; each of these becomes a candidate zone for an AI workflow. - She frames AI workflow design as product discovery: understand the “job to be done,” map the journey, pick a focused slice, prototype minimally, then refine based on real use. - Next steps promised: select a single step to focus on, decide whether simple code or LLM tooling is more appropriate, prototype in Claude Code, and iteratively test/adjust prompts and structure. - The article is positioned as a guided exercise: by following her example with your own task, you should end up with a first-draft AI workflow tailored to your work. --- ## 📚 References - Product Talk – “How to Build AI Workflows with Claude Code (Even If You're Not Technical)” – https://www.producttalk.org/how-to-build-ai-workflows-with-claude-code/ - Claude & Claude Code by Anthropic – https://www.anthropic.com - Teresa Torres on Continuous Discovery Habits – https://www.producttalk.org/continuous-discovery-habits/ --- ## 🗣️ Paraphrase - Torres explains that instead of wrestling with messy drafts and outlines, she now collaborates with Claude to refine her ideas, structure content, and handle surrounding tasks like research and SEO, which significantly increases her writing throughput. - She argues that the best way for non-technical people to build AI workflows is to first map their existing process step-by-step, then progressively introduce AI support to specific steps, treating workflow design like product discovery and iterating toward a dependable system. --- ## 📰 Newsletter In this week’s productivity spotlight, we’re looking at how product discovery expert Teresa Torres designs AI workflows with Claude Code—no CS degree required. Her approach is highly relevant for WordPress pros: map your content or development process, then let AI assist with outlines, research, SEO, and even paywall strategy. If you’re building with WordPress, this maps neatly to: - Content workflows in the Block Editor (draft → refine → optimize). - Developer workflows for plugin docs, changelogs, and release notes. Helpful resources: - WordPress + AI tooling overview: https://wordpress.org/plugins/tags/ai/ - Claude & Claude Code details: https://www.anthropic.com - Content strategy with WordPress: https://learn.wordpress.org/ Try documenting one of your WordPress workflows (e.g., publishing a tutorial or shipping a plugin update), then experiment with an AI assistant to improve just one step—such as outlining or changelog generation. --- ## ✅ Action Items #### DevRel - Identify 1–2 high-friction content workflows (blog posts, tutorials, release notes) and map steps end-to-end. - Build Claude/LLM example prompts for: outline review, gap-finding, and “developer critic” mode on draft docs. - Add AI-assisted quickstarts or code examples to SDK docs showing how to wire Claude Code into existing workflows. - Launch a content series (“From messy draft to workflow”) showcasing real dev workflows upgraded with AI. - Run community office hours focused on designing safe, reliable AI workflows for docs and demos. - Establish internal feedback loops: capture successful community prompts/workflows and fold them back into docs. #### OSPO - Draft policy guidelines on acceptable AI usage for open source contributions (e.g., AI-assisted vs AI-generated code). - Integrate license scanning and attribution checks for AI-generated snippets in repos. - Define a security review path for any automated workflow that can commit code or change configs. - Maintain an inventory of AI-assisted tooling used across projects, with owners and review cadence. - Offer consult hours on using AI to triage issues, summarize design docs, and prepare governance materials. - Align AI workflow initiatives with broader open source strategy and reporting metrics. #### Community - Support maintainers in documenting their “real” workflows (release, triage, roadmap) as a base for AI augmentation. - Create contributor-friendly, AI-assisted onboarding flows (e.g., “explain this issue in plain language” prompts). - Share example prompts/workflows for community moderation, newsletter creation, and event recaps. - Give sponsors visibility into time saved and quality gains from AI workflows via transparent reports. - Practice cross-project incident drills that include AI-assisted summarization and communication templates. #### WordPress - Core: Explore AI-assisted proposals summaries and changelog drafting; document safe patterns and limits. - Core Performance: Use AI to highlight performance regressions in commit messages and summarize profiling logs. - Core AI: Define guardrails, evaluation sets, and UX patterns for integrating AI in wp-admin and Gutenberg. - Core Program Team: Create ritualized “workflow mapping” sessions for release leads and focus teams. - Mobile: Experiment with AI-assisted release notes and user-facing update copy. - Playground: Provide AI-ready sandboxes for trying and sharing WordPress workflows (content, plugin, theme). - Community: Add AI workflow examples to WordCamp/meetup talk kits; encourage live-mapping sessions. - Training: Build lesson plans around mapping and improving WordPress workflows with AI. - Marketing: Pilot AI-assisted campaign brief generation and newsletter outlines. - Support: Use AI to cluster tickets and generate draft responses, with clear human review steps. - Design: Try AI for UX copy exploration and alternative flows, while keeping decisions human-led. - Documentation: Develop prompts and patterns for turning dev notes into user-focused docs. - Photos / TV / Polyglots / Accessibility / Meta / CLI / Hosting / Tide / Openverse / Themes / Plugins: - Map key workflows (submission, review, translation, accessibility checks, ingestion). - Identify 1–2 candidate steps per team for AI augmentation (summaries, classification, draft text). #### The WPCC - Design fellowship tracks around “workflow mapping and AI augmentation” for maintainers. - Provide microgrants for teams who document and publish AI-assisted workflows that respect open source values. - Run governance clinics showing how to document AI policies and decision rights transparently. - Pilot funding models that reward demonstrable improvements in maintainer time savings and burnout reduction. ### FAIR - Encourage technical independence by promoting locally stored metadata about AI workflows and their provenance. - Publish funding transparency reports for any AI-related sponsorships or grants. - Maintain and share reusable governance artifacts (charters, agreements, scorecards) for AI use in open ecosystems. --- ## 🧪 Hypotheses - Mapping an existing task step-by-step significantly increases the quality and reliability of resulting AI workflows. - Treating AI workflow design like product discovery leads to higher adoption and sustained use. - Non-technical users can successfully build useful AI workflows when guided by concrete, task-based examples. - Using AI as a critic and sounding board improves the clarity and structure of written work. - Narrow, targeted AI augmentation (one step at a time) outperforms trying to fully automate entire job workflows. --- ### 📊 Methodology - Run controlled A/B tests: one group builds AI workflows directly; another first maps their process, then designs workflows. Compare effectiveness and satisfaction. - Survey knowledge workers before and after adopting the “map-then-design” approach, measuring confidence, time saved, and perceived quality. - Collect writing samples created with and without AI-as-critic workflows, then have blinded reviewers rate clarity and structure. - Track retention and repeated use of AI workflows over several weeks to see which design patterns stick. - Conduct qualitative interviews on failure modes and friction when non-technical users attempt to adopt Claude Code or similar tools. --- ### ❓ Questions - Which categories of tasks benefit most from AI-as-collaborator vs fully automated workflows? - How detailed should a task map be before it yields diminishing returns for workflow design? - What skills do non-technical users most need to safely and effectively design AI workflows? - How can we standardize prompts and patterns without making workflows brittle or overfitted? - What governance is required when AI workflows start touching production systems (code, configs, customer data)? --- ### 🔮 Future Research - Develop frameworks for evaluating the robustness and safety of AI workflows designed by non-technical users. - Explore integrations between task-mapping tools (e.g., story maps, user flows) and AI orchestration platforms. - Investigate long-term cognitive effects of heavily relying on AI for idea development and writing. - Study best practices for disclosure and transparency when content is AI-assisted rather than purely human-authored. - Examine how AI workflows can support, rather than replace, collaborative human team processes. --- ### 💡 Implications - Practice: Professionals should start any AI initiative by mapping an existing workflow, then introducing AI in small, testable slices. - Policy: Organizations need clear AI usage and disclosure guidelines, especially around content and open source contributions. - Security: Any workflow that can make real changes (code, infra) requires tight controls, reviews, and auditability. - Education: Training should include workflow-mapping exercises and prompt design as core digital literacy. - Funding: Backing tools and programs that teach “workflow-first” AI design may yield better ROI than generic AI pilots. --- ### 🧠 Learning Styles Assessment Reading level: Approx. Grade 9–10 (Flesch-Kincaid estimate, based on relatively short sentences and minimal jargon). #### 👀 Visual learners - Present: Some structure via headings and bullet lists of steps; occasional screenshots are implied. - Improve: Add workflow diagrams (before/after), Claude prompt–response screenshots, and visual story maps. #### 🔊 Aural learners - Present: Mention of an audio version for subscribers. - Improve: Provide a short free audio summary and optional TTS for key how-to sections. #### 🗣️ Verbal learners - Present: Clear narrative, definitions by example (e.g., story mapping), concrete step lists. - Improve: More analogies (e.g., “AI as junior editor”), comparison tables for “manual vs AI-assisted.” #### 🧩 Kinesthetic learners - Present: Encouragement to “follow along with your own ideas” and map your own task. - Improve: Checklists, fill-in-the-blank workflow templates, and small guided experiments with Claude Code. #### 🧠 Logical learners - Present: Sequential lists of steps, clear cause-effect framing (map → choose step → prototype → iterate). - Improve: Decision trees for “code vs LLM,” matrices for evaluating which steps to automate. #### 🧑‍🤝‍🧑 Social learners - Present: Framed as part of an ongoing series and shared practice among readers. - Improve: Discussion prompts, suggested peer-review sessions for comparing AI workflows. #### 🧘‍♀️ Solitary learners - Present: Reflective framing around personal workflows and habits. - Improve: Journaling prompts (“Where do I get stuck?”) and self-checklists after each design step. #### 🌿 Naturalistic learners - Present: Limited; mostly productivity-focused. - Improve: Analogies to ecosystems (workflows as habitats) or sustainability (reducing cognitive waste). --- ### 📊 Accessibility Assessment #### Screen reader semantics - Present: Clear headings, paragraph structure, and lists should be screen-reader friendly. - Improve: Ensure images/screenshots have descriptive alt text; avoid using images for text. #### Keyboard navigation - Present: Standard article layout likely supports linear navigation. - Improve: If interactive examples exist, ensure full keyboard operability and visible focus states. #### Contrast and readability - Present: Product Talk generally uses decent contrast and readable typography. - Improve: Verify contrast ratios for all text and links; maintain generous line spacing. #### Text resizing and reflow - Present: Modern responsive layout likely supports zoom and mobile reflow. - Improve: Test at 200%+ zoom, avoid content clipping, and honor prefers-reduced-motion. #### Media and captions - Present: Audio version for the article. - Improve: Provide full transcripts for audio, captions for any embedded video, and avoid autoplay. #### Forms and interactions - Present: Basic subscription/paywall forms. - Improve: Ensure labels, error messages, and hit areas are accessible and screen-reader friendly. #### Standards reference - Map identified gaps to WCAG 2.2 AA—especially 1.4.x (contrast), 1.2.x (media), 2.1.x (keyboard), and 3.3.x (forms). --- ### 🌍 DEIB Assessment #### Cultural breadth - Present: Generalizable productivity framing; not strongly region-specific. - Improve: Include diverse examples of workflows from different roles, regions, and backgrounds. #### Inclusive language - Present: Neutral, professional language with minimal jargon. - Improve: Add short glossary boxes for terms like “story mapping” and “workflows.” #### Neurodiversity support - Present: Clear structure, chunked sections, bullet lists. - Improve: More consistent subheadings, summary boxes, and explicit callouts for key steps. #### Representation - Present: Primarily the author’s own example and perspective. - Improve: Include short vignettes from different users (e.g., non-writers, non-English speakers). #### Psychological safety - Present: Non-judgmental tone about messy processes; encourages iterative experimentation. - Improve: Explicit reassurance about mistakes with AI, and guidelines to avoid using AI in high-risk contexts without review. --- ### ✅ Logical Validity #### Strong reasoning - Inductive: From her own writing process improvements, she infers that mapping tasks and iterating AI workflows can help others; this is plausible and transparent. - Causal: She links mapping workflows to better AI design outcomes (clearer points to augment), which follows standard process-improvement logic. - Method analogy: Treating AI workflow design like product discovery is a strong analogy for her audience of product people. #### Fallacies - Potential overgeneralization: Heavy reliance on one-person experience; could be strengthened with more varied case studies. - Fix: Acknowledge limitations explicitly and point to other roles or data where the approach has been tested. #### Tone of voice - Tone: Practical, reflective, and transparent, with a strong ethical stance on disclosure and no hypey AI language. --- ### 🛡️ DARVO Analysis - Deny: No denial or minimization of problems; she openly discusses past struggles and limitations. - Attack: No attacks on critics or skeptics; AI is framed as optional help, not mandatory. - Reverse victim/offender: Not present; she does not position herself as a victim of AI skeptics or tools. - Overall: The piece is explanatory, not defensive, and doesn’t exhibit DARVO patterns. --- ## 🎯 Conclusion Torres argues that the key to building useful AI workflows—especially with tools like Claude Code—is to start from a clear map of your existing process, then iteratively introduce AI where it can best augment you. For readers, the practical next step is to pick one recurring task, document its steps, and experiment with AI on a single, low-risk slice before expanding. --- ## 🧱 Refactor Appendix (append after 🎯 Conclusion) ### 📚 Literature Split Pad - Purpose: Stage Konik CEE units that become per-source Literature notes. - Split level: **H3**. - Include verbatim quotes under 25 words and a [view highlight] link. ### [H1] AI-augmented workflows improve writing output > [!quote] [How to Build AI Workflows with Claude Code](https://www.producttalk.org/how-to-build-ai-workflows-with-claude-code/) by [[Teresa Torres]] via [[Product Talk]]. Published 2025-12-10. > “Instead of writing my way to what I think, I discuss my way to what I think.” Two lines: context and use. Cite data if relevant. [view highlight](https://www.producttalk.org/how-to-build-ai-workflows-with-claude-code/#H1) #xref #articleseed ID:rw_[MISSING] ### [H2] Task mapping reveals where AI can help > [!quote] [How to Build AI Workflows with Claude Code](https://www.producttalk.org/how-to-build-ai-workflows-with-claude-code/) > “Once you've identified the AI workflow you want to create, start by mapping out exactly what you do to complete the task…” Two lines: context and use. [view highlight](https://www.producttalk.org/how-to-build-ai-workflows-with-claude-code/#H2) #xref ID:rw_[MISSING] --- ### 🔖 Atomic Split Pad (Zettel Candidates) - Purpose: Promote reusable claims into Atomic notes. - Split level: **H4**. - Each item: one idea, present tense, specific, testable. #### [Z1] Mapping tasks precedes good AI workflows - Scope/condition in one line. - Pointer to page/time or dataset. - Links: [[AI-Workflows]] [[Task-Mapping]] #### [Z2] AI as critic sharpens written thinking - Scope/condition in one line. - Pointer to page/time or dataset. - Links: [[AI-Workflows]] [[Writing-Process]] --- ### ✏️ Fleeting Queue (Extract by Selection) - Purpose: Capture quick actions or sparks for Fleeting notes. - Extraction: **Select one line** → “Refactor: Extract selection to new note (content only).” - Keep each line short and self-contained. - [F1] Map my own WordPress content workflow step-by-step—/how-to-build-ai-workflows-with-claude-code/ #fleeting #inbox - [F2] Experiment with Claude/LLM as a “critic” on my next article draft—/how-to-build-ai-workflows-with-claude-code/ #fleeting #inbox - [F3] Draft an internal guide on “workflow-first AI design”—/how-to-build-ai-workflows-with-claude-code/ #fleeting #inbox --- ## 🧰 Template Library (paste into Note Refactor → Refactored Note Template) ### 📝 Fleeting — Refactored Note Template (Konik-ready) --- type: fleeting status: inbox created: {{date:YYYY-MM-DD}} {{time:HH:mm}} process_by: {{date:YYYY-MM-DD}} # review within 24–48 hrs source_title: [MISSING] source_link: author: [MISSING] source_type: article|paper|video|podcast|thread reading_stage: inspectional|analytical|syntopical project: tags: #fleeting #inbox --- # 📝 Fleeting Note — {{title}} ## 🧠 One-liner (my words) - State the single idea in one short sentence. ## 🧭 Context & cue - Where did this arise? Page/timecode/speaker. - Trigger: quote, data point, claim, or question. ## 🧪 Adler checks (quick) - What is this about as a whole? - What is said in detail, and how? - Is it true? Evidence for or against? - So what? Impact or next move. ## 🔎 Bias & reliability snapshot - Possible bias or framing? - What evidence would change my view? ## 🧩 Konik bridge - Draft a claim header for promotion. ### [H1] 5–10-word present-tense claim > [!quote] [{{source_title}}]({{source_link}}) by [[{{author}}]]. Accessed {{date:YYYY-MM-DD}}. > “Short quote under ~25 words.” One line on why this matters. [view highlight]({{source_link}}) #xref ID:rw_[MISSING] ## 🧰 Next moves (pick 1–2) - [ ] Promote to *Literature* with citation. - [ ] Draft *Atomic* claim from [H1]. - [ ] Park as testable question. ## ✅ Processing checklist - [ ] Rewrote in my words. - [ ] Decided: trash | keep | convert. - [ ] Linked to at least one related note. - [ ] Tagged and moved out of inbox. --- ### 📚 Literature — Refactored Note Template (Konik-ready) --- type: literature status: draft created: {{date:YYYY-MM-DD}} {{time:HH:mm}} adler_level: inspectional|analytical|syntopical reading_stage: inspectional|analytical|syntopical source_title: [MISSING] author: [MISSING] year: publisher_or_journal: edition_or_volume: doi: url: accessed: {{date:YYYY-MM-DD}} citekey: project: tags: #literature #reading --- # 📚 Literature Note — {{title}} ## 🧰 Action Inbox - [H1] Short action — context/{{url}} #articleseed - [H2] Short action — context/{{url}} #xref - [H3] Short action — context/{{url}} #share ## 🧭 Bib Capsule - Title, author, year, venue, and link. - One-line scope: what the work attempts. ## 🧠 Author’s Thesis (their words) - Core claim briefly. Include page/timecode. ## 🧩 Key Arguments - Argument A — method/evidence — page/time. - Argument B — method/evidence — page/time. - Argument C — method/evidence — page/time. ## 🧪 Evaluation (Adler) - What is this about as a whole? - How is it argued? Fairly and clearly? - Is it true or plausible? Why? - So what? Practical or theoretical impact. ## 🗒️ Notes (Konik CEE units) ### [H1] 5–10-word present-tense claim > [!quote] [{{source_title}}]({{url}}) by [[{{author}}]] via [[{{publisher_or_journal}}]]. Published {{year}}. Accessed {{accessed}}. > “Verbatim highlight under ~25 words.” Two lines: context and use. Link to data if relevant. [view highlight]({{url}}#H1) #xref #articleseed ID:rw_[MISSING] ### [H2] 5–10-word present-tense claim > [!quote] [{{source_title}}]({{url}}) by [[{{author}}]] via [[{{publisher_or_journal}}]]. Published {{year}}. Accessed {{accessed}}. > “Verbatim highlight under ~25 words.” Two lines: context and use. [view highlight]({{url}}#H2) #xref ID:rw_[MISSING] ## 🔖 Zettel Candidates #### [Z1] 5–10-word present-tense claim - One-line context and page/time. #### [Z2] 5–10-word present-tense claim - One-line context and page/time. ## 🏷️ Tag Coverage #articleseed: 0 • #xref: 0 • #share: 0 • #bmf: 0 • #addendum: 0 ## ✅ Processing Checklist - [ ] Thesis captured in my words. - [ ] 1–3 Zettel candidates drafted. - [ ] Action Inbox pruned to high-signal items. - [ ] Cross-links added and filed. --- ### 🔖 Atomic (Zettel) — Refactored Note Template (Konik-ready) --- type: atomic id: {{date:YYYYMMDDHHmm}}-{{title}} status: evergreen created: {{date:YYYY-MM-DD}} {{time:HH:mm}} origin: reading|thinking|experiment proof_state: conjecture|working|supported|refuted source_literature: [[Source-Literature-Note]] citekey: project: tags: #zettel #idea --- # 🔖 {{title}} <!-- 5–10 words, present tense, specific --> ## 🧠 Claim - One idea only. Clear and testable. ## 🧾 Evidence & Pointers (Konik) - [H#] from [[{{source_literature}}]] — page/time — brief rationale. - ID:rw_[MISSING] — link to highlight or dataset. > [!quote] Evidence snippet (≤25 words) if essential. > Cite page/time. Keep it minimal. ## 🔍 Scope & Conditions - When this holds. Boundaries and exceptions. ## 🔗 Links - Broader: [[Parent-Concept]] - Narrower: [[Specific-Case]] - Related: [[Parallel-Idea]] ## 🧪 Tests / Falsifiers - One check that could disconfirm the claim. - Simple study or observation plan. ## 🧰 Implications - Practice, policy, or design moves in one line each. ## 🔙 Backlinks - From: [[{{source_literature}}]] • Context line. ## ✅ Maintenance - [ ] Links in and out. - [ ] Title still precise? - [ ] Proof state updated. ## Highlights - I'd start with a brain dump of everything that was swirling in my head. Then I would try to structure it into an outline. But I often lost patience with that. So I'd just start writing. ([View Highlight](https://read.readwise.io/read/01kc6w8sm1qt1hw03cnzjyhskh)) - Tags: 👻 ai highlighted