Claude Certified Engineers logoCCE

CCAO-F · Study Guide

← Domain 7: Troubleshooting and Optimization

7.3 · Lesson 3 of 3

Optimize Workflows

What you need to know

Once a workflow works, make it cheaper and faster without lowering quality. Share what works through Projects, standardise good prompts, and use the smallest model that still meets the bar.

Equal quality, lower cost

  • Optimise only after quality is acceptable.
  • Try a smaller, faster model on simple steps and keep it if quality holds.
  • Trim unnecessary context — concise, relevant input is faster and often more accurate.
  • Remove or merge steps whose output nobody uses.

Share what works

Put proven instructions and reference knowledge into a shared Project and turn repeated prompts into templates so improvements reach the whole team. Savings from repeated setup add up. Re-check periodically — tools, models, and needs change.

Worked example

Time one multi-step run and note the model and retyped setup. Move repeated context into a Project, test a smaller model on simple steps against a rubric, remove unused steps, then rerun. The optimised run is faster or cheaper and still passes the same quality rubric — with the strong model kept only where the smaller tier failed.

Exam traps

  • Optimize for speed alone

    Equal quality is the condition. Speed gained by lowering quality is not optimisation.

  • Cut review or verification to save time

    Saving time by dropping safeguards is not optimisation.

  • Optimise before quality is acceptable

    Fix quality first; then make the working workflow cheaper and faster.

  • Assume a smaller model is fine everywhere without a rubric check

    Keep the stronger model on steps where the smaller tier fails the quality bar.

Practice scenario

A weekly workflow already meets quality. Setup is retyped every time and a strong model is used on simple reformatting steps. What is the right optimisation?

Choose one answer

Build exercise

Make a Workflow Faster and Cheaper Without Losing Quality

7.3 · Advanced · ~45 min · claude.ai

What you will practise

  • How shared Projects remove repeated setup.
  • How a smaller sufficient model reduces cost and wait time.
  • How to cut steps that add no value.
  • How to verify quality did not drop.

Complete the Domain 7 optimisation exercise so the new run beats baseline on time or cost and still passes the rubric.

Open Domain 7 exercises

Path: /learn/claude-associate/exercises/7-troubleshooting-optimization

Sources

View progress