A complex system does not necessarily have to appear all at once. Components may originally perform different functions. Genes can be duplicated and subsequently modified. Existing proteins can be recruited into new systems. Structures can acquire additional components and functions over time.
This process is often called co-option. It is important because an evolutionary pathway does not require every part of a modern molecular machine to have served its present function from the beginning.
Comparative studies of bacterial flagella and secretion systems identify homologous components and relationships among molecular systems. Research has proposed stepwise histories involving gene duplication, modification, recruitment, loss, and divergence. Similar principles of reuse and modification occur throughout evolutionary biology.
Historical reconstruction has limits. Similarity and homology can identify relationships among components, but the precise sequence of ancient events may be difficult to recover. A proposed pathway must also account for selectable functions at relevant stages rather than merely arrange modern components into a possible sequence.
For particular molecular systems, how much of the proposed evolutionary pathway is supported by comparative evidence, how much has been demonstrated experimentally, and how much remains reconstruction?
Intelligent Design should engage co-option and evolutionary reconstruction directly. Evidence that components can be reused or modified is important. The remaining question is whether these mechanisms adequately account for the origin and integration of the complete systems under investigation.