Building Your People Memory
Zia has a meeting with Omar at 11:00 to discuss the analytics brief for the investor deck. He asks for a message draft.
Without people memory, the output is competent but generic: "Hi Omar, I wanted to reach out about the analytics brief we discussed. Could you let me know your timeline for completion?"
With people memory — with Omar's full profile loaded — the output is different: "Omar — I wanted to follow up on the analytics brief for the investor deck. Deadline is [Date + 5 days]. To confirm scope: Q1 student completion rates vs prior year; revenue per student (last 4Q); Q1 cohort retention at 30/60/90 days. Format: slides-ready data, charts preferred. Can you confirm receipt and flag any data availability issues by EOD today? Happy to reduce scope if needed."
The second message gives Omar exactly what he needs: specific scope, confirmed deadline, clear format, and an explicit offer to adjust — because the entry notes he dislikes last-minute asks and will push back on scope creep. The difference is not intelligence. It is context.
Layer 2 of work.local.md is where that context lives.
What Layer 2 Contains
Layer 2 is the team memory layer. For each key stakeholder, it stores the professional context that makes outputs stakeholder-aware rather than generically professional.
This is not a directory or an org chart — those already exist. Layer 2 stores the working intelligence that does not live anywhere else: how this person actually communicates, what they are focused on right now, what they find difficult, and what a good working relationship with them requires.
When /agentic-office:workplace-context produces a delegation message, a meeting brief, or a handoff communication, it loads the relevant person entries to calibrate the output. Without Layer 2, every output uses default professional tone. With Layer 2, every output is tailored to the specific person receiving it.
The Person Entry Format
Each person entry has seven fields. Here is the format from the workplace-context skill:
- name: "[Full name]"
role: "[Job title]"
reports_to: "[Manager name]"
communication: >
[How they prefer to receive information; what works; what annoys them]
current_focus: "[What they are working on right now]"
priorities: "[Their current top 1–3 priorities]"
note: "[Anything important about working with this person]"
sensitivity: "[Optional — anything that must not be shared broadly]"
The most important field is communication. This is what directly calibrates outputs. A strong communication entry describes specific behaviours — channel preference, response to urgency, how they process requests — rather than general traits. "Detail-oriented" tells the skill nothing useful. "Prefers written briefs to verbal requests; will ask follow-up questions if context is missing; give her more context than you think necessary" is actionable.
The note field is for working observations — not personality assessments, but professional patterns that affect how collaboration works. "Will push back on scope creep — be specific about what you need" is a working observation. "Perfectionist" is a personality assessment. The former helps; the latter adds noise.
The sensitivity field is optional. When used, it carries a RESTRICTED flag. More on this below.
Building the Case Study Entries
Here are the three entries for the Panaversity team members you will be working with throughout this chapter.
Omar Farooq — Head of Analytics
- name: "Omar Farooq"
role: "Head of Analytics, Panaversity"
reports_to: "Zia Khan (CEO)"
communication: >
Prefers data-backed requests. Give him lead time — he dislikes last-minute asks.
Best channel: Slack DM for routine requests; email for formal requests.
Be specific about exactly what you need — he will flag scope ambiguity.
current_focus: "Rebuilding the student performance dashboard; Q1 analytics refresh"
priorities: "Q1 analytics refresh; data pipeline audit"
note:
"Will push back on scope creep — be specific about what you need and why.
Data requests need at least 3 business days' lead time. Responds better
to written briefs than to verbal conversations for analytical tasks."
Ayesha Raza — Senior Data Analyst
- name: "Ayesha Raza"
role: "Senior Data Analyst, Finance & Analytics"
reports_to: "Omar Farooq"
communication: >
Detail-oriented; prefers written briefs to verbal requests. New to edtech —
experienced in fintech. Help her bridge between the two domains where needed.
She processes best with structured context; do not assume she knows internal terms.
current_focus: "Q1 analytics refresh (first project at Panaversity); pipeline audit"
priorities: "Q1 analytics refresh; pipeline audit"
note:
"New starter — joined March 2026, still in onboarding window. Timely feedback
in the first month shapes confidence. A week's delay on work she has submitted
sends the wrong signal. Be specific and encouraging in all feedback. She will
use fintech mental models — help her translate to edtech where needed."
Dr. Sana Mirza — Head of Curriculum
- name: "Dr. Sana Mirza"
role: "Head of Curriculum"
reports_to: "Zia Khan (CEO)"
communication: >
Academic precision; prefers structured proposals with clear evidence base.
Will ask for research backing on pedagogical claims — prepare sources.
Prefers formal written communication for proposals; values thoroughness over speed.
current_focus: "Onboarding; taking formal ownership of the PHM framework"
priorities: "Understanding the curriculum architecture; establishing PHM framework ownership"
note:
"Starts Monday — joining from Aga Khan University (PhD Learning Sciences).
Give her formal ownership of the PHM framework — she will want to build on
it, not just inherit it. Key relationship to manage: Omar Farooq — different
working styles on data questions for the curriculum; may need mediation on
early interactions. Do not make informal curriculum design decisions before
she is settled in."
Sensitivity Handling
Some information belongs in Layer 2 but must not surface broadly. The sensitivity field handles this.
Consider Zara Hussain, a Senior Data Engineer being assessed for a Team Lead role:
- name: "Zara Hussain"
role: "Senior Data Engineer"
reports_to: "Head of Engineering"
communication: >
Technically precise; will flag scope ambiguity immediately. Prefers async;
does not like surprise meeting invites. Likes sufficient context before responding.
current_focus: "Infrastructure roadmap; platform reliability work"
priorities: "Platform stability; engineering velocity"
note: "Direct and efficient — prefers concrete requests with clear success criteria."
sensitivity: "RESTRICTED — Being assessed for Team Lead role. Do not reference
succession planning in group communications or shared briefings."
The sensitivity field carries two functions. First, it stores information that is professionally relevant — Zara's succession process affects how to brief about her in certain contexts. Second, the RESTRICTED flag instructs the skill to never include this information in group outputs or shared briefings where Zara might see it.
The rule: Sensitive entries are applied only when directly relevant to a one-to-one task (e.g. Zia preparing for a private conversation with Zara's manager). They are never surfaced in:
- People briefs shared with a group
- Delegation messages
- Meeting prep documents where the subject is an attendee
- Any output that could be seen by the person the entry concerns
This is not overcaution. It is professional practice. The information has value — and that value disappears the moment the entry surfaces in the wrong context.
The Person Brief in Action
Once entries are in place, you can request a person brief before any significant interaction. The brief synthesises what Layer 2 knows about each person into actionable pre-interaction guidance.
/agentic-office:workplace-context
> Give me a brief on everyone I'll be speaking with today:
Omar, Ayesha, and Dr. Sana Mirza.
Sample output:
PEOPLE BRIEF — 17 March 2026
════════════════════════════════════════════════════════════
─── OMAR FAROOQ — Head of Analytics ───────────────────────
Reports to: Zia Khan (CEO)
Current focus: Q1 analytics refresh; data pipeline audit
Current mood: Analytics brief request was delegated yesterday;
has not confirmed receipt — may be managing the
pipeline audit alongside the new ask
How to approach: Give the analytics request in writing with full
scope confirmed; offer to reduce scope if timeline
is a problem; confirm he has everything he needs
Watch for: Will flag scope ambiguity immediately — be specific
Do not: Make last-minute asks; leave scope open-ended
Today's context: Still waiting for confirmation on analytics brief;
may need a nudge — but give him the benefit of the
doubt on timeline before following up
─── AYESHA RAZA — Senior Data Analyst ─────────────────────
Reports to: Omar Farooq
Joined: March 2026 — Day 17 of onboarding (still in 30-day window)
Background: Fintech → edtech transition; strong analytical background
Current work: Q1 analytics refresh; pipeline audit (first project)
How to approach: She submitted her first analysis this morning —
review and respond today; be specific and encouraging
Note: Timely feedback in the first month shapes confidence;
a week's delay would be a missed opportunity to signal
that her work is seen and valued
Bridge: She will use fintech mental models — help her translate
to edtech metrics and terminology where needed
Do not: Delay feedback; leave her uncertain about quality
─── DR. SANA MIRZA — Head of Curriculum (NEW) ─────────────
Reports to: Zia Khan (CEO)
Starts: Monday 18 March 2026
Background: Education technology; PhD Learning Sciences; Aga Khan University
Style: Academic precision; evidence-based; prefers structured proposals
Strengths: Will strengthen the pedagogical foundation of the curriculum
To give her: Formal ownership of the PHM framework — she will want to
build on it, not just inherit it
To manage: Omar relationship — different working styles on data questions
for the curriculum; introduce them carefully; mediate early
Do not: Make informal curriculum design decisions before she is settled in;
present her with decisions already made on her core remit
════════════════════════════════════════════════════════════
This output did not come from the skill's intelligence — it came from the entries you wrote. The skill structured and surfaced what you already knew. The quality of the brief is a direct function of the quality of the entries.
If the brief for Omar reads as generic — "good communicator, data-focused" — that is a signal to return to his entry and make the communication field more specific.
Building Your Own People Entries
Now build Layer 2 for your own context. Aim for five to eight people: the direct reports, peers, and stakeholders you interact with most frequently and whose working style most affects your outputs.
For each person, work through the seven fields. Two discipline points:
Write how they actually communicate, not how you wish they did. If someone dislikes confrontation but you prefer directness, the entry should reflect their preference — because the brief will calibrate to them, not to you. An entry that describes how you want to work with them produces briefs calibrated to your preferences, not to reality.
Avoid the generic. Before saving any communication entry, ask: "Would this describe 80% of professionals?" If yes, it is too generic. "Prefers clear communication" fails this test. "Responds better to written Slack messages than verbal requests for analytical tasks; needs two business days' lead time on data pulls" passes it.
Exercise: Building Layer 2
Time: 30 minutes
Plugin command: /agentic-office:workplace-context
Step 1 — Add 5–8 stakeholders (20 minutes)
For each person in your Layer 2, use:
/agentic-office:workplace-context
> Add person: [Name]
Role: [Title]
Reports to: [Manager]
Communication: [How they prefer to receive information]
Current focus: [What they are working on now]
Priorities: [Top 1-3 priorities]
Note: [One specific working observation]
Focus on people whose working style directly affects your own outputs: the colleague whose approval you need, the direct report you delegate to most frequently, the stakeholder whose communication style is most different from yours.
Step 2 — Generate person briefs (10 minutes)
Once entries are in place, run a brief for the three people you are most likely to interact with today or this week:
/agentic-office:workplace-context
> Give me a brief on [Name 1], [Name 2], and [Name 3].
Evaluate the output: Is the guidance specific enough to be actionable? For each person, ask: "Would reading this brief change how I approach the interaction?" If the answer is no, the entry needs more specificity.
Deliverable: Layer 2 of your work.local.md with 5+ person entries. At least one person brief that you evaluate as specific enough to change your approach to an interaction.
Your work.local.md is progressive — Lesson 5 adds Layer 3 (Projects). Do not start a new file. Every lesson from here builds on what you have already written.
Try With AI
Use these prompts in Cowork or your preferred AI assistant.
Reproduce: Add the Omar Farooq entry and test it with a delegation scenario.
Add the following person entry to memory:
Name: Omar Farooq
Role: Head of Analytics, Panaversity
Reports to: Zia Khan (CEO)
Communication: Prefers data-backed requests. Give him lead time —
dislikes last-minute asks. Best channel: Slack DM for routine,
email for formal. Will flag scope ambiguity.
Current focus: Rebuilding the student performance dashboard; Q1 analytics refresh
Priorities: Q1 analytics refresh; data pipeline audit
Note: Will push back on scope creep — be specific about what you need.
Data requests need at least 3 business days' lead time.
Now I need to delegate an analytics task to Omar. The task is:
produce a 2-page investor-facing analytics summary covering Q1
student completion rates vs. prior year; revenue per student trend
(last 4 quarters); Q1 cohort retention at 30/60/90 days. Needed
by [Date + 5 days]. Format: slides-ready data, charts preferred.
Draft the delegation message to Omar. What should I know before sending?
What you are learning: The delegation message output should reflect Omar's profile specifically — lead time acknowledgement, specific scope, written format, offer to adjust. If the output is generic, the entry needs more specificity in the communication field. This exercise shows the direct relationship between entry quality and output calibration.
Adapt: Build entries for your own key stakeholders and test them.
I am going to add three people to my workplace memory. For each one,
help me write a complete person entry using this format:
- name, role, reports_to
- communication (specific behaviours, not generic traits)
- current_focus, priorities
- note (one specific working observation)
Person 1: [Name, Role, and what you know about how they work]
Person 2: [Name, Role, and what you know about how they work]
Person 3: [Name, Role, and what you know about how they work]
After adding all three, generate a person brief for each of them.
Rate each brief: is the guidance specific enough to change how
you approach an interaction? What would make it more specific?
What you are learning: The exercise of rating brief specificity teaches you what makes a person entry useful. When you rate a brief as too generic, trace the gap back to the entry — usually the communication field is under-specified. Improving person entries is an ongoing practice, not a one-time setup.
Apply: Use people memory for a difficult conversation.
I need to have a difficult conversation with someone on my team
about a missed deadline. Here is what I know about them:
[Describe their role, communication style, how they respond to feedback,
any relevant context about the missed deadline and why it occurred]
Add them to memory as a person entry. Then give me a brief for this
specific interaction: what should I know, how should I approach it,
what should I avoid, and what outcome should I be aiming for?
What you are learning: People memory is most valuable in high-stakes interactions — not just routine delegation. The brief does not tell you what to say. It tells you how this specific person is likely to receive what you say, which is the foundation for any difficult conversation.
Flashcards Study Aid
Continue to Lesson 5: Projects and Priorities →