🎯 Discovery & Strategy — Objectives¶
Purpose
Objectives, key activities and deliverables of Phase 1: identifying the right problem and assessing feasibility for an AI project.
🎯 Objective¶
The primary objective of the Discovery phase is to identify the right problem and verify that we are ready to start an AI project. We define the challenge from the end user's perspective, not from the technology's perspective, and establish whether AI is genuinely the appropriate solution.
Key result: A clearly defined problem with a substantiated hypothesis that AI is the right solution, including an initial risk inventory, a drafted Project Charter, and an agreed Collaboration Mode.
[!TIP] The Fast Lane For projects with a Minimal Risk level and an Instrumental/Advisory mode (Mode 1 & 2) we offer an accelerated route. Following a positive Risk Pre-Scan (Gate 1), a limited Validation pilot can be started directly. See Fast Lane for details.
✅ Entry Criteria (Definition of Ready)¶
Before this phase starts, the following conditions must be met:
- A business sponsor is in place who recognises the problem and has allocated budget.
- The problem cannot be trivially solved with existing tools or processes.
- There is willingness to share data and processes for analysis.
- The team has identified at least one concrete use case to investigate.
Do not start Discovery without a sponsor
Projects without an identified business sponsor lack the mandate to access data, interview stakeholders, and make decisions. If no sponsor exists, the initiative remains an idea — not a project.
⚙️ Core Activities¶
1. Problem Exploration¶
We articulate the challenge from the user's perspective and assess whether AI is truly the right approach.
- Question Articulation: What is the real problem? What are the pain points? Who experiences them daily?
- AI Suitability Assessment: Is AI truly the right solution here? Can the problem be solved more simply with existing tools, process changes, or automation?
- Success Indicators: How do we measure whether we have solved the problem? Define measurable outcomes before any technical work begins.
- Project Type Classification: Determine whether this is a Type A project (building with AI as a development tool) or Type B project (AI in the product for end users). See Project Type Classification for details.
2. Data Evaluation¶
We analyse the required information across three dimensions before committing to a solution.
- Access: Are we legally permitted and technically able to access the data? Check legal rights, API availability, database access, and security requirements.
- Quality: Is the data complete and consistent? Assess completeness, accuracy, currency, and the presence of duplicates or inconsistencies.
- Relevance: Does the data contain the answer to the question? Evaluate correlation with the objective and representativeness of the target population.
3. Risk Inventory¶
We perform an initial scan for legal, ethical, and organisational obstacles.
- EU AI Act Classification: Does the system fall under the high-risk category? Use the Risk Pre-Scan to determine the risk level.
- Privacy & GDPR: Which personal data is being processed? Document data flows and identify lawful bases for processing.
- Ethical Questions: Can the system discriminate or cause harm? Consult the Guardian for an initial ethical assessment.
- Organisational Risks: Do we have the right people, skills, and resources? Identify gaps that need to be filled before proceeding.
4. Collaboration Mode Assessment¶
We determine the intended human-AI relationship for this project. The Collaboration Mode drives governance requirements throughout the lifecycle.
- Review the five Collaboration Modes — from Instrumental (Mode 1) to Autonomous (Mode 5).
- Select the intended mode based on risk level, task complexity, and organisational readiness.
- Record the mode in the Project Charter as a design constraint.
- Guideline: Start low, scale up. Begin in Mode 2 (Advisory) to collect data and build trust before transitioning to higher autonomy modes.
Do not skip directly to Mode 4 or 5
Jumping to Delegated or Autonomous mode without intermediate learning phases leads to uncontrolled risk exposure. Always validate human oversight procedures before increasing autonomy.
👥 RACI¶
| Role | Responsibility in Discovery |
|---|---|
| AI Product Manager | Accountable: Owner of the business case, problem articulation, and Project Charter. |
| Data Scientist | Responsible: Performing the Data Evaluation across Access, Quality, and Relevance. |
| Business Sponsor | Consulted: Validates the problem statement and the value hypothesis. |
| Guardian (Ethicist) | Consulted: Conducts the initial ethical and legal scan, advises on EU AI Act classification. |
| Stakeholders | Informed: Are kept informed of findings and participate in problem exploration interviews. |
✅ Exit Criteria (Gate 1 — Go/No-Go Discovery)¶
The Discovery phase closes when all of the following are satisfied:
- Project Charter is drafted with clear scope, objectives, and intended Collaboration Mode.
- Data Evaluation is completed with a positive result on Access, Quality, and Relevance.
- Risk Pre-Scan is completed and risk level is classified (Minimal, Limited, or High).
- Collaboration Mode is assessed and recorded in the Project Charter.
- Gate 1 review is conducted with the Business Sponsor and Guardian.
- Go/No-Go decision is documented.
Collaboration Mode: [Mode X — Name] as recorded in the Project Charter. Required validation for this mode: → See Evidence Standards for the full requirements.
📦 Deliverables¶
The following artefacts are produced during this phase:
- Project Charter — scope, objectives, stakeholders, and intended Collaboration Mode.
- Risk Pre-Scan — initial risk classification under the EU AI Act framework.
- Data Evaluation Report — assessment of Access, Quality, and Relevance.
- Collaboration Mode Assessment — documented rationale for the selected mode.
Next step: Complete the Project Charter and run the Risk Pre-Scan. → Use the Project Charter as a starting point. → See also: Activities | Collaboration Mode Assessment
Version: 1.1 Date: 07 May 2026 Status: Final