AI Business Strategist

The charter identifies a specific service outcome, users, current process, decision owner, scope, exclusions, and a credible non-AI alternative.

The capability classification is technically sound and separates likely uses, unsupported claims, uncertainty, human responsibility, and failure conditions.

The value statement links a sourced baseline to a measurable target, timeframe, beneficiary, calculation, and guardrail without double counting.

Material assumptions, affected groups, possible harms, evidence gaps, and the conditions for stopping or reframing the opportunity are explicit.

Arabic and English summaries communicate the same decision, figures, limits, and ownership in concise language suitable for sponsors.

Cards use a consistent level of detail and trace each opportunity to a real service need, baseline, owner, affected groups, and credible alternative.

Criteria, scales, direction, weights, missing-data handling, and calculations are documented and applied consistently without false precision.

The selected mix reflects value, feasibility, risk, learning, dependencies, and capacity rather than simply choosing the highest raw scores.

Evidence gaps and disputed assumptions are visible, sensitivity tests identify unstable rankings, and high-risk opportunities receive defer or stop criteria.

The bilingual narrative communicates start, investigate, defer, and stop decisions with consistent figures, rationale, ownership, and next review.

Required inputs and outcomes are traceable to owners and sources, with purpose, permission, quality, representativeness, access, retention, and gaps recorded.

Every score has an evidence reference and defined scale; unknowns remain visible and material disagreement is recorded.

Decision rights and accountable owners are assigned for data access, quality, model use, user impact, risk acceptance, and escalation.

Actions address root gaps, dependencies and resources are realistic, owners and dates are named, and completion has verifiable evidence and decision consequences.

Arabic and English summaries state ready, conditionally ready, or not ready, with identical blockers, owners, evidence, and next review.

Classification considers consequence, scale, sensitivity, autonomy, reversibility, vulnerable groups, and existing decision rights, with evidence for each rating.

The register covers foreseeable technical, data, human, service, exclusion, misuse, and operational harms without collapsing causes, events, and consequences.

Preventive, detective, and responsive controls are specific, proportionate, owned, testable, and linked to evidence and residual-risk ratings.

The design gives qualified people timely information and authority to review, override, stop, explain, correct, and handle appeal without rubber-stamping.

The bilingual record names the decision, authority, evidence, dissent, conditions, review date, escalation route, and prohibited uses consistently.

Volumes, unit measures, service outcomes, affected groups, and benefit formulas are sourced, avoid double counting, and distinguish cashable from non-cash value.

The workbook includes lifecycle cost categories, timing, units, formulas, contingency, recurring costs, and exit assumptions with no material formula error.

AI, non-AI, procurement, and do-nothing options are compared on common assumptions, with base, upside, downside, break-even, and sensitivity results.

The recommendation follows the evidence, respects governance constraints, identifies irreducible uncertainty, and defines funding stages, stop conditions, and owner.

Arabic and English memo summaries use identical figures, clearly label estimates, and make the requested decision and its consequences easy to review.

The charter defines one decision question, eligible population, setting, duration, intervention, comparison, exclusions, capacity, owner, and boundary against wider rollout.

Metrics map to hypotheses, use valid denominators and collection timing, define segments and missing data, and avoid causal claims unsupported by the design.

Baseline, target, guardrail, minimum evidence, uncertainty, and proceed, adapt, pause, or stop thresholds are explicit and linked to the value case.

Roles, staffing, support, monitoring, complaints, incident response, pause, rollback, and communications are realistic and the tabletop exposes and resolves gaps.

Arabic and English notices explain purpose, AI role, data use, human review, limitations, feedback, complaint, and alternative service consistently.

Every lifecycle decision and incident has one accountable owner, feasible responsible roles, clear escalation, and no critical ownership gap or conflict.

Processes cover intake through retirement, state inputs, decisions, evidence, service levels, handoffs, controls, and records, and align with governance gates.

Required skills and capacity are compared with current evidence, gaps have proportionate build, buy, partner, or stop actions, and ongoing support is funded and owned.

The plan targets behavior and service outcomes, addresses resistance and workload, includes feedback and recourse, and preserves accessible alternatives for affected groups.

Arabic and English messages consistently explain what changes, what does not, why, when, expected action, support, limits, and accountable contact.

The recommendation resolves or exposes dependencies and contradictions across value, readiness, governance, pilot evidence, operations, and alternatives.

Horizons contain outcome-based work, realistic dependencies, capacity and cost, accountable owners, evidence outputs, dates, and measurable gates rather than feature lists.

Material claims and figures trace to current sources or calculations, facts and assumptions are separated, uncertainty is visible, and alternatives remain comparable.

Governance, affected-group safeguards, human authority, complaints, monitoring, incident response, review, pause, stop, and exit are embedded in the roadmap.

Arabic and English materials request the same decision, use consistent figures and terms, distinguish evidence from uncertainty, and support informed challenge.