| Level | Data Required | AI – Predicting & Managing Demand | AI – Scheduling & Allocating Supply |
|---|---|---|---|
| 1 – Basic | - Org team headcounts • via directory integration or CSV • manual input (CSV) | - Maintain accurate, up-to-date headcounts • No forecasting or intelligent sharing at this level | - Auto-update space allocations, assignments or neighbourhoods to match current headcounts • Instant block & stack planning |
| 2 – Weekly Patterns | - Weekly attendance counts (anonymized) • individual IDs matched to teams • count of days attended per week | - Identify team-level in-office patterns • e.g. 50% attend 1 day/week, 30% attend 2 days, 20% attend 3 days - Classify “office personas” • e.g. Type A: 5 days every other week (avg 2.5); Type B: avg 2.5 days/week - Limit: no day-specific ratios → sub-optimal allocations | - Load-balance demand across weekdays • Recommend which days employees should come in - Match complementary schedules • e.g. pair 1-day/5-day patterns with 4-day/2-day to maximize co-attendance |
| 3 (Recommended minimum) | - Daily attendance (anonymized) • did / did not attend each day • IDs matched to teams | - Calculate daily sharing ratios per team • e.g. Team Alpha: 2.7 people/desk (Mon), 2.1 (Tue), 1.3 (Wed), 1.9 (Thu), 3.2 (Fri) - Predict peak weekly demand • e.g. Team Alpha peak 1.3 ratio → 8 desks for 10 people - Compute team co-attendance scores (same-day % of team onsite) | - Auto-allocate the right number of desks per team, per day - Optimize neighbourhood assignments - Recommend day-swaps to boost same-day attendance |
| 4 – Real-Time Insights | - Real-time attendance & location (per user) • IDs matched to teams | - Live building capacity monitoring • Enables on-the-fly re-provisioning - Individual co-attendance recs • Detect low co-attendance patterns & suggest schedule tweaks - Preferred space types • Track Wi-Fi/MAC to learn seating preferences | - Direct people to nearby desks or neighbourhoods based on real-time demand - Push in-office day recs to individuals for better collaborator alignment - Update future allocations by learned preferences |
| 5 – Collaboration Driven | - Collaboration data (MS Graph, Google, Slack) | - Identify informal teams based on interaction frequency • Reveal cross-org project groups - Surface preferred collaborators • Correlate meeting locations with Wi-Fi/MAC footprint | - Auto-suggest new custom teams based on communication patterns - When users span multiple teams, learn their true “home” teams via shared-wifi behavior |