A pre-project process assesses whether an ML project is possible and useful before funding and delivery time are secured. It frames the opportunity, business requirements, stakeholders, data, infrastructure, tools, security, privacy, ethics, expected performance, scope, and required effort. section 2.0
Use the process to reduce delivery and organizational risk:
- Clarify the project type, stakeholder expectations, scale, structure, and general requirements. section 2.0
- Examine available data assets and delivery infrastructure before judging whether an ML approach is viable. section 2.0
- Account for security, privacy, and ethical considerations early, because omissions can create liability and reputational harm. section 2.0
- Turn requirements into a hypothesis describing the expected outcome and key challenges, then estimate effort using the delivery environment and team structure. section 3.7
- Review and sign off project documentation before discussing customer costs; use the resulting case to support, reject, or reshape the proposal. section 3.6 section 3.7
Pre-project work is deliberately shallower than funded delivery: deep customer-data investigation and high-performance infrastructure access may not be available before contracts are signed. The aim is a factual, proportionate basis for price, timeframe, staffing, and go/no-go decisions—not an exhaustive implementation. section 2.0 section 3.6
