graph TD
    A[Identify the event A whose probability is needed] --> B[Find a partition $B_1$ $B_2$ ... $B_n$ of sample space]
    B --> C[Determine prior probabilities P of $B_i$]
    C --> D[Determine likelihoods P of A given $B_i$]
    D --> E[Compute P of A = sum of P of A given $B_i$ times P of $B_i$]
    E --> F{Need reverse conditional?}
    F -->|Yes| G[Apply Bayes theorem: P of $B_i$ given A]
    F -->|No| H[Answer is P of A]

The total probability theorem is used when the event A can occur through multiple mutually exclusive "paths" (the partition). It is always the first step before applying Bayes' theorem.