Skip to content

AI Strategy KPI Baselines

Purpose

The AI Strategy commits Feoda to ambitious 18-month targets — including Level 4 (Native) AI maturity across all seven departments. Without baselines, those targets are unmeasurable.

This document establishes:

  1. The measurement methodology for each KPI in Section 11 of the AI Strategy.
  2. The baseline reading (the value as of the strategy approval quarter, 2026 Q2).
  3. The data source and owner for each measurement.

Baselines are captured once per KPI. Subsequent readings are recorded in the quarterly review (see AI Strategy Section 14).

Status

Draft — methodology defined for all KPIs; numerical baselines pending the first measurement cycle (target: 2026-05-31).

For each KPI marked Pending, the responsible owner is committed to capturing the baseline value before the date noted.

1. Measurement principles

  • One source of truth per KPI. Each measurement has exactly one named system or named owner producing the number. No duplicate measurement.
  • Define before measuring. A KPI is not "measured" until its definition, formula, data source, frequency, and owner are recorded in this document.
  • Same instrument across the horizon. Once a KPI's measurement methodology is fixed, it does not change without a documented amendment. Methodology changes invalidate trend comparisons.
  • Quarterly cadence. All KPIs are reported quarterly to the AI Strategy Steering Committee. High-risk KPIs (Section 4) are reviewed monthly.
  • Audited annually. A sample of measurements is audited annually by the Head of Technology against the underlying data source.

2. Adoption KPIs

ID KPI Definition Data source Owner Cadence Baseline (2026 Q2) Notes
A-01 Departments with ≥1 production AI workflow Count of distinct departments (of 7) where at least one AI workflow is approved by the relevant Department Head and used in production for ≥30 consecutive days. Department workflow registry (to be created under AP-04). Head of Technology Quarterly Pending — target capture 2026-05-31 Target: 7/7 by month 12.
A-02 Employees with active AI agent access % of employees in roles deemed "AI-relevant" (per the role classification published with AP-01) who have an active, identity-bound, named AI account on at least one approved tool, measured against the staff register. Approved Provider admin consoles + HR staff register. Head of Technology + HR Quarterly Pending — depends on AP-01 + AP-03 Target: 100% of relevant roles by month 6.
A-03 Out-of-policy AI usage incidents Count of confirmed incidents in the period where an employee used an unapproved AI provider, anonymous account, or accessed AI with confidential data outside approved channels. Audit log (AP-05) + HR/IT incident reports. Head of Technology Monthly 0 confirmed (no audit log yet — see caveat below) Target: zero by month 6. Caveat: baseline is "no detected incidents" because the audit log infrastructure is not yet built (AP-05). True baseline will be re-stated once AP-05 ships.
A-04 Foundational AI literacy training completion % of all current employees who have completed the 8-module foundational AI literacy curriculum (AP-02) and passed the knowledge check. LMS / training register (to be selected under AP-02). HR Monthly during rollout, then quarterly 0% — curriculum not yet built (AP-02) Target: 100% by month 6.
A-05 Role-specific advanced training completion % of department members who have completed the role-specific advanced AI training defined for their department. Measured per department. LMS / training register. HR + Department Heads Quarterly Pending — curricula to be defined per department in Phase 2 Target: 100% by month 12.
A-06 Active weekly adoption % of eligible-role employees with at least one logged interaction with an approved AI tool in the rolling 7-day window. Approved Provider usage telemetry + audit log (AP-05). Department Heads (consolidated by Head of Technology) Quarterly (with monthly internal tracking) Pending — depends on AP-03 + AP-05 Target: ≥90% by month 12, sustained.
A-07 AI competency integrated into hiring/onboarding/performance review Boolean per department: has AI competency been added to (a) job descriptions, (b) onboarding curriculum, (c) performance review template? HR documentation. HR Quarterly 0/7 departments (pending Phase 4) Target: 7/7 complete by month 18.

3. Outcome KPIs

ID KPI Definition Data source Owner Cadence Baseline (2026 Q2) Notes
O-01 Median story-to-deployment time (engineering) Median elapsed wall-clock time from "story moved to In Progress" to "merged + deployed to production", computed over the rolling quarter. Excludes stories in won't-do or cancelled. GitHub + CI deployment timestamps. Head of Technology Quarterly Pending — to be captured from Q1 2026 GitHub data by 2026-05-31 Target: material reduction vs baseline. "Material" = ≥25% by month 12.
O-02 Median client-implementation cycle time Median elapsed wall-clock time from contract signed to client go-live. Computed across the rolling 12-month window of completed implementations. Client implementation register (clients/). Head of Delivery Quarterly Pending — to be back-computed from Pymble, Saint Edwards, Al Faisal data by 2026-05-31 Target: reduction vs baseline. Currently 3 data points — small-sample caveat applies.
O-03 Median support resolution time Median elapsed time from a support ticket being raised to it being marked Resolved, computed over the rolling quarter. Support ticketing system (to be confirmed — current system in catalogue). Head of Support Quarterly Pending — system identification + baseline capture by 2026-05-31 Target: reduction vs baseline.
O-04 Documentation compliance score Composite score: (% of docs with complete frontmatter × 0.4) + (% of docs with all internal links resolving × 0.4) + (% of months with a CHANGELOG entry × 0.2). Computed by the AI-audit job once AP-04 ships. Documentation repository + AI audit job (AP-04). Head of Technology Quarterly Manual one-off baseline: ~85% (estimate) — to be replaced with automated reading once AP-04 ships. Target: ≥98% by month 9. Methodology: reuse the audit pattern from the 2026-04-20 compliance audit.

4. Risk KPIs (reviewed monthly)

ID KPI Definition Data source Owner Cadence Baseline (2026 Q2) Notes
R-01 Confidentiality incidents Count of confirmed incidents involving AI handling of Confidential or Restricted data outside approved channels (per AI Strategy data classification, Section 8). Audit log (AP-05) + Security/Legal incident register. Head of Technology + Legal Monthly 0 known Target: zero across strategy horizon. Caveat: as with A-03, baseline reflects detection capability, not absence of incidents, until AP-05 ships.
R-02 Audit log coverage % of AI interactions on Restricted-tier data that are captured in the audit log, measured by sampled comparison against provider-side admin logs. Audit log (AP-05) + provider admin consoles. Head of Technology Quarterly 0% (audit log not yet built) Target: 100% by month 6 (AP-05 deliverable).
R-03 Quarterly access review completion % of quarterly access reviews completed on time (where "on time" = within 14 days of quarter end). Each provider × each role group counts as one review. Access review register (to be created under AP-01). Head of Technology Quarterly N/A — first review due 2026-09-30 Target: 100% on time, every quarter.

5. Maturity KPIs

ID KPI Definition Data source Owner Cadence Baseline (2026 Q2) Notes
M-01 Departments at Level 4 (Native) Count of departments (of 7) assessed at Level 4 against the AI maturity model (AI Strategy Section 6) by the Steering Committee. Quarterly self-assessment + Steering Committee ratification. AI Strategy Steering Committee Quarterly 0/7 Target: 7/7 by month 18.
M-02 Departments at Level 3 or above As above, threshold Level 3 or higher. As above. As above. Quarterly Pending — baseline assessment scheduled 2026-05-15 Target: 7/7 by month 12.
M-03 Cross-cutting AI workflows in production Count of AI workflows that span ≥2 departments and are in production for ≥30 consecutive days. Department workflow registry (AP-04). Head of Technology Quarterly 0 Target: ≥3 by month 12.
M-04 Core processes redesigned AI-first per department Count of department-owned processes that have been (a) explicitly redesigned with AI as a first-class component, (b) approved by the Department Head, (c) documented under company/processes/ or the relevant solution folder. Process registry (company/processes/ index). Department Heads Quarterly 0/7 departments with a redesigned process Target: ≥1 per department by month 9; full redesign by month 18.

6. Baseline-capture plan

The following items must be completed to convert all Pending rows above into hard numbers. Owner and target date are listed.

Action Owner Target Blocked by
Compute O-01 baseline from GitHub history (Q1 2026 data) Head of Technology 2026-05-31 None
Compute O-02 baseline from Pymble/Saint Edwards/Al Faisal records Head of Delivery 2026-05-31 Confirm dates in client overview.md files
Identify support ticketing system + extract Q1 2026 ticket data for O-03 Head of Support 2026-05-31 None
Run a one-off documentation compliance audit for O-04 baseline (manual, mirroring 2026-04-20 audit) Head of Technology 2026-05-15 None
Conduct first maturity self-assessment per department (M-01, M-02) AI Strategy Steering Committee 2026-05-15 None
Publish role classification (which roles are "AI-relevant") for A-02, A-06 Head of Technology + HR 2026-05-31 None
Restate A-03, R-01, R-02 baselines once AP-05 (audit logging) ships Head of Technology 2026-10-31 (month 6) AP-05
Restate O-04 with automated reading once AP-04 ships Head of Technology 2027-01-31 (month 9) AP-04

7. Reading log

This table records each quarterly reading once captured. Baseline column above is never updated — only this log grows.

Quarter Date captured Captured by KPIs updated Link to report
2026 Q2 Pending Pending Initial baselines per Section 6 plan TBD

8. Status log

Date Author Change
2026-04-22 Head of Technology Initial draft — KPI definitions, methodology, baseline-capture plan. Numerical baselines pending the first measurement cycle.