Knowledge Management Research & Analysis Personal Productivity

Teresa Torres's AI use case

Product discovery coach and author at Product Talk

Built a personalized Claude Code system for task management, automated academic research, and modular context files that make AI collaboration easier with shorter, simpler prompts.

The problem

What was broken before AI

Teresa’s work spans writing, teaching, research, product thinking, and daily execution. A normal task app could hold to-dos, but it did not give her enough control over portability, structure, search, or AI access. Research was another problem: staying current with new papers takes time, and useful context can get scattered across notes, files, and tools. Even when AI could help, she did not want to rebuild the same context in every prompt.

What changed

What the use case made possible

Teresa moved the center of her workflow into local Markdown files and Claude Code. Tasks live as individual Markdown files in Obsidian with metadata that scripts can read. A custom /today command assembles a daily list of due, overdue, and in-progress work. Research scripts search academic sources and produce daily digests. A modular context library gives Claude maps to her business, writing style, products, and personal context so she can ask shorter questions and still get calibrated answers.

Why this matters

Why this use case is worth studying

Teresa’s workflow is valuable because it treats context as an asset. The task manager, research digest, and context library all make the same point: AI gets more useful when it can work inside a structure that already matches the user’s life and thinking. The system is personal, local, and modular, which means Teresa can keep her data, change the rules, and keep improving the assistant over time.

Use this when

When this pattern applies

Use this pattern when your work depends on tasks, research, notes, and context that need to stay portable and easy for AI to read. It works especially well if normal productivity apps feel too rigid, or if you keep re-explaining your business, writing style, products, or research interests to AI.

Exponential Builder analysis

Context belongs in durable files.

Teresa’s setup shows that AI collaboration improves when the important background lives outside the chat window in Markdown, metadata, and index files that can be reused across sessions.

Small commands beat big rebuilds.

A custom /today command, research digest, and paper-summary flow each solve one repeated friction point, which makes the system maintainable instead of turning it into another sprawling productivity project.

Personalization comes from structure, not longer prompts.

By giving Claude modular maps to her tasks, research interests, business, writing style, and personal context, Teresa can ask shorter questions and still get feedback that fits her work.

Who this is for

Best fit

Writers and educators

Product discovery coaches

Researchers and analysts

Knowledge workers using Obsidian

People comfortable with Markdown or light scripting

Operators who want custom workflows instead of rigid SaaS tools

Anyone building a long-term personal context library for AI

What to avoid

Mistakes and warnings

Where this pattern can go wrong if you copy it too literally.

Do not try to build the whole personal operating system at once.

Avoid one giant context file that loads irrelevant information for every task.

Keep research discovery filtered, or the digest can become another inbox.

Do not trust academic summaries without reading enough to judge the source.

Make the file structure simple enough that you can maintain it without AI.

Public workflow preview

The shape of the workflow

A high-level look at how the use case works, with the reusable pattern made clear.

Move tasks into plain files

Each task becomes a Markdown file with metadata, notes, and checklists that Claude can read.

Generate the day with one command

A custom /today slash command scans task files and creates a daily plan with due, overdue, and in-progress work.

Bring research into the same rhythm

Scheduled scripts search academic sources for Teresa’s topics and add a daily research digest to her task flow.

Summarize only what passes her filter

When she saves a paper, Claude generates a summary focused on methodology, effect size, and research quality.

Build context in small files

Instead of one giant memory file, Teresa keeps modular context files with indexes so Claude can load the right background for the task.

Improve the system after each session

She asks Claude what should be documented, then turns useful discoveries into new context files.

Copy the pattern

The reusable idea

Pattern in one sentence

Put tasks, research, and context into small readable files so AI can help from structure instead of relying on long prompts.

Reusable idea

Teresa’s use case is a reminder that personal AI systems get better when the context is organized outside the chat window. If you keep tasks, research, notes, and preferences in plain files, AI can work with them more flexibly. The power is not in building the perfect app. It is in making your existing work legible enough that the assistant can help without being re-taught everything each time.

Steal this workflow

Build a one-folder AI operating layer:

Create three folders: /tasks, /research, and /context.

For each task, make one Markdown file with YAML fields for status, due date, tags, and project.

Write a short context file called task_rules.md that explains your tags, statuses, and how you want daily planning handled.

Create a daily review command or routine that asks Claude to scan /tasks and return: overdue items, due today, in-progress work, and a suggested order.

Put your research topics and keywords in research_config.md, then use a daily or weekly digest instead of checking sources manually.

When you save a paper, ask for a summary using your criteria: methodology, effect size, limitations, and whether it is strong enough to use.

Split reusable background into small files such as business_profile.md, writing_style.md, product_context.md, and personal_profile.md.

Add an index file that tells Claude which context file to load for which kind of request.

After any useful session, ask what should be documented, then update the relevant context file before the insight disappears.

Suggested prompt

“Claude, use the context index to decide which files matter for this request. Review my task files and create today’s plan with overdue items, due-today items, in-progress work, and a suggested order. If any task is unclear, flag the missing metadata instead of guessing. After you produce the plan, tell me what new rule or context we should document so tomorrow’s plan is easier.”

Field notes

Get new AI use cases in your inbox

A short weekly note on how real people are using AI to save time, make money, build tools, and run their lives.

Related use cases

Keep exploring nearby systems.

Browse all

Business OperationsKnowledge Management+1

Included Health clinical team

Clinical team using an internal ambient scribe workflow at Included Health

Telehealth ambient scribe workflow where visit audio is transcribed with Whisper and processed with GPT-4o into SOAP notes and patient instructions for clinician review.

WhisperChatGPTOpenAI Audio API+5

Use case

View use case

Sales & OutreachBusiness Operations+1

Oracle AI Sales Agents

AI agents for sales administration and account intelligence at Oracle

Sales operations workflow where AI agents help sales professionals update company records and generate reports from varied data sources, including multilingual information.

Oracle Fusion Cloud SalesOracle Fusion Cloud Customer Experience (CX)

Use case

View use case

Business OperationsKnowledge Management+1

Jennifer Smith + Aaron Podolny

Co-founders; Jennifer Smith is CEO and Aaron Podolny is CTO at Scribe

Workflow documentation and optimization workflow where Scribe captures process steps automatically, then analyzes process libraries to find automation and AI adoption opportunities.

Scribe CaptureScribe Optimize

Use case

View use case