# Use Cases

## Current Spec Context

These use cases assume the current runtime I-AM format:

- Prefix: `IAM-v0.2`
- Optional `BASE` segment
- `STATE` always appears first after `BASE` when present
- Optional note anchors use `SEGMENT:anchors[item1;item2;item3]:...`
- `SKILLS` labels may use one or two words joined by underscore for disambiguation

See [I-AM format overview](iam.md) for the latest segment syntax details.

## Software
**Question:**
"I need to plan the next sprints work for a refactor. How should I prioritize it?"

**Personalized response style (example):**
"Given your profile, you do best when structure and clarity come first, but you still benefit from room for creative options. I would split this into three buckets: reliability fixes, architecture simplifications, and developer-experience improvements. Start with reliability to reduce risk, then sequence the work that unlocks the most downstream value. Keep each item small enough to complete in 1-2 days so momentum stays high, and leave room for one exploratory improvement if the sprint capacity allows it."

## Science
**Question:**
"How should I design an experiment to compare two hypotheses without bias?"

**Personalized response style (example):**
"For your communication profile, a clear decision framework will likely feel most useful. Define one primary outcome metric first, then pre-register the stopping rule and exclusion criteria. Use blinded analysis for the final comparison step if possible. If the experiment has multiple possible confounders, isolate the most important one and treat the others as secondary notes rather than trying to control everything at once."

## Finance
**Question:**
"Can you help me create a monthly budget that I will actually stick to?"

**Personalized response style (example):**
"You are likely to follow plans that are practical and low-friction rather than overly rigid. I suggest a three-layer budget: fixed costs, flexible essentials, and intentional discretionary spending. Add one weekly 10-minute check-in and one automatic transfer on payday. That keeps control high without creating an exhausting tracking burden, and it gives you a simple way to adjust when real life changes."

## Healthcare
**Question:**
"How can I prepare for a doctor's appointment so I get the most useful answers?"

**Personalized response style (example):**
"You will likely get better results if you prepare a short list instead of trying to remember everything in the room. Write down your top three symptoms, when they started, and what makes them better or worse. Then list your main question in one sentence so the appointment stays focused and efficient."

## Education
**Question:**
"What's the best way for me to learn a new topic quickly without burning out?"

**Personalized response style (example):**
"A mixed approach will probably work best for you: a short overview first, then hands-on practice, then a quick recap in your own words. Keep sessions short enough that you still feel curious at the end, not drained. If you want, I can turn any topic into a 3-step learning plan with a review schedule."

## Personalized Interests: Local Events
**Question:**
"What local events would I probably enjoy this month?"

**Personalized response style (example):**
"Based on your profile, events with thoughtful atmosphere and moderate stimulation may be a good fit: design exhibits, science talks with Q&A, small live music venues, book discussions, and curated community workshops. I would avoid recommending only high-intensity festival environments unless you ask for high-energy options. If you share your city, I can shortlist events by vibe, social energy level, and how structured they are."

## Personalized Interests: New Music
**Question:**
"Recommend new music for me this week."

**Personalized response style (example):**
"Your music pattern suggests a blend of sophisticated and unpretentious tracks, with enough intensity to stay engaging. I would build a 15-song mix across three lanes: reflective tracks for focus, rhythm-forward tracks for momentum, and a few exploratory picks outside your usual genre. If you want, I can return this as a Monday-to-Friday listening schedule matched to work mode, commute mode, or deep-work mode."

## Personalized Interests: Travel
**Question:**
"What kind of trip would I probably enjoy more: a packed city weekend or a slow nature getaway?"

**Personalized response style (example):**
"You would likely enjoy the version that balances novelty with recovery time. A city trip works best if it includes a few planned highlights and some quiet breaks. A nature getaway works best if there is enough structure to prevent the weekend from feeling vague. If you want, I can compare both options based on your energy level and preferred pace."

# Other Potential Uses

The I-AM string could be shared with a friend, co-worker, or someone you are meeting for the first time. They could use it to ask the LLM questions about how to better understand and communicate with you.

Other useful applications include:

- Writing style and tone coaching for emails, chats, and presentations.
- Interview preparation and role-play for jobs, networking, or client conversations.
- Team collaboration guidance for managers, coworkers, and cross-functional partners.
- Content personalization for recommendations, newsletters, playlists, and event suggestions.
- Product and UX persona creation for marketing, design, and research teams.
- Onboarding support for new assistants, agents, or tools that need to adapt quickly.
- Self-reflection prompts that help you notice patterns in motivation, focus, and communication.

# Risks and Concerns
Personality and preference systems can be useful, but they can also be misused. I-AM string outputs should never be treated as a hiring, promotion, medical, legal, or high-stakes decision tool on their own. I-AM string quality also depends on honest and current user input, so stale or inaccurate profiles can produce poor guidance.

I-AM string sharing introduces privacy and control risks. Users should be cautious about who receives an I-AM string file and what scope of data is shared. Even when privacy-enhancing approaches are used, protections are not absolute and depend on key management, implementation quality, and threat model assumptions.

Some privacy directions are community discussion topics and future research ideas, not current built-in protections. This includes Privacy-Enhancing Technologies (PETs) such as Fully Homomorphic Encryption (FHE), selective disclosure, and key-scoped encrypted I-AM string sharing for model access control.