MindPalaceX · The Workshop
The Idea Wall →

On the Idea Wall · ▲ 0 votes

Sentinel — the spec

Sentinel is an AI-powered time-intelligence agent for solo business owners. It runs passively in the background, tracking how time is actually spent across apps and tools, then surfaces daily and weekly coaching to help owners shift attention toward higher-value work. It does not ask for manual input; it watches, categorises, reports, and proposes — so the owner can decide what to delegate, block, or batch.

---

Decisions locked

QuestionAnswer
ShapeAI agent / automation
AudienceSolo business owners
Core loopOwner reviews AI coaching on time patterns, then delegates or batches a flagged task
Must-have v1 scopePassive tracking, category breakdown, daily attention report, calendar-blocking suggestions, weekly ROI comparison
Success metric at 30 days10 people using it every week

---

The core loop

  1. Sentinel passively captures activity across connected apps (calendar, email, browser, project tools) throughout the workday.
  2. Each activity is classified into a business category (sales, admin, fulfilment, strategy, etc.) in real time.
  3. At end of day, the owner receives a Daily Attention Report — time spent by category, number of context switches, and what demanded focus versus what drained it.
  4. The AI coach surfaces one or two specific suggestions: a calendar block to protect tomorrow's focus time, or a task cluster worth batching.
  5. The owner accepts, edits, or dismisses suggestions; accepted blocks are pushed directly to the calendar.
  6. At the end of each week, a Weekly ROI Comparison shows how this week's time allocation moved relative to the owner's baseline, with a plain-language verdict.

---

v1 scope

Passive Activity Tracking — Connects to the owner's existing tools via OAuth and API integrations and records activity without requiring manual input. Lives on the Integrations setup screen and runs silently in the background.

Time-Allocation Breakdown by Category — Automatically maps each captured activity to a business category using AI classification. Displayed as a visual breakdown on the Dashboard, updated throughout the day.

Daily Attention Report — Delivers a concise end-of-day summary via email or SMS showing time-by-category, total context switches, and the sharpest attention drains. Generated nightly and accessible in the Reports screen.

Calendar-Blocking Suggestions — The AI coach proposes specific focus blocks based on the day's pattern and the owner's priorities. Suggestions appear in a feed on the Dashboard; accepted blocks push to the connected calendar via API.

Weekly ROI Comparison — Compares this week's time allocation against the owner's rolling baseline, highlighting drift and improvement. Delivered as a structured report on the Reports screen and optionally via email.

---

Deliberately later

  • Revenue impact scoring per activity — requires enough transaction data to correlate reliably; meaningful only after several weeks of baseline exist.
  • Recommended task batching or delegation — depends on pattern history that v1 is still accumulating; coaching quality rises sharply with longitudinal data.
  • Custom alert thresholds by owner — adds configuration surface area that would slow onboarding; better introduced once owners understand their own patterns from v1 reports.

---

Data model sketch

  • users — id, name, email, phone, subscription_status, onboarded_at
  • integrations — id, user_id, provider, oauth_token, scopes, last_synced_at
  • activity_events — id, user_id, source_integration, raw_label, category, started_at, ended_at, duration_seconds
  • categories — id, name, description, owner_defined (bool)
  • daily_reports — id, user_id, date, category_breakdown (JSON), context_switch_count, top_drain, delivered_at
  • suggestions — id, user_id, type (block/batch), body, status (pending/accepted/dismissed), created_at
  • calendar_blocks — id, user_id, suggestion_id, start_time, end_time, pushed_at, external_event_id
  • weekly_snapshots — id, user_id, week_start, allocation_json, baseline_json, roi_delta

---

Screens

  • Onboarding / Integrations — connects the owner's apps and tools, sets business categories, establishes baseline preferences.
  • Dashboard — live time-allocation view, today's context-switch count, and the AI coach suggestion feed.
  • Daily Report View — full breakdown for any selected day, readable as a narrative plus category chart.
  • Weekly Report View — baseline comparison, drift analysis, and plain-language verdict for the selected week.
  • Calendar Suggestions Panel — review, edit, and confirm pending focus blocks before they are pushed.
  • Account & Billing — subscription management, notification preferences (email vs. SMS), and connected integration status.

---

How to use this document

  • Bring it into Mind Palace, where the guided platform and its AI coach pick up from exactly this document and build the product with you step by step.
  • Hand it to a developer as the complete v1 brief — scope, data model, and screen list are ready to act on immediately.
  • Keep it as the single source of truth while you build, so every decision about what to add or cut is measured against what is written here.

This is Kairos's. Build yours.

A spec like this is step 3 of Mind Palace's nine — the other six take a weekend: working software, a real offer, first customers.

Want to build this with Kairos?

Tell us who you are and we'll broker the introduction — their contact details stay private until they say yes.