Troubleshooting and Optimization

Question 1 · 7.1

A policy officer attaches a 90-page consultation document and asks for the five main objections raised. The answer is well written, but two of the five objections do not appear anywhere in the document.

What is the most effective next step?

• AAdd three examples of well-written objection summaries to the prompt.
• BAsk for exact quotes about objections first, then summarise only from those.
• CRephrase the request more precisely and specify a 500-word limit.
• DTell Claude to be accurate and not to make anything up in its answer.

Question 2 · 7.1

A team lead’s prompt asks Claude to summarise a project update, list risks, suggest mitigations, draft a client email and propose an agenda. The replies consistently cover the first three and skip the last two.

What is the best diagnosis and fix?

• AThe model is too small for the task; switch to a more capable model.
• BThe request bundles five deliverables; split it into separate steps and check each.
• CThe prompt needs stronger wording, such as telling Claude it must complete all five.
• DThe conversation is too long; start a new chat with the same prompt.

Question 3 · 7.1

An internal communications manager keeps getting updates in heavy bullet points when she wants flowing prose. She has already tried adding “do not use bullet points” to her prompt.

Which two changes are most likely to help? (Select 2.)

• ADescribe the output positively — complete paragraphs of flowing prose.
• BWrite her own prompt in plain paragraphs rather than bullets.
• CAttach the source documents above her question.
• DAsk for step-by-step reasoning before the final answer.
• ERepeat the instruction three times in the prompt.

Question 4 · 7.2

Across a 30-person consultancy, partners report that Claude-drafted client summaries consistently miss the firm’s standard risk paragraph. Each consultant writes their own prompts in their own style.

What is the most appropriate adjustment?

• AAsk each consultant to add a line about the risk paragraph to their prompts.
• BPut the requirement and two example summaries into the shared project setup.
• CGive the thumbs-down on any summary that omits the paragraph.
• DStop using Claude for client summaries until the model improves.

Question 5 · 7.2

A fundraising manager has rewritten her donor-appeal prompt four times. Each version reads better to her and to her colleagues. Donations per appeal have not moved at all.

What does this pattern most strongly suggest?

• AThe prompt needs a fifth revision with more specific instructions.
• BShe is iterating against colleague approval rather than the outcome the appeal exists for.
• CDonation rates are outside anyone's control, so the appeals are fine.
• DThe task should be handed back to an external fundraising agency.

Question 6 · 7.2

An operations lead has found a phrasing that reliably produces good incident write-ups. She wants the improvement to survive her being on leave.

Which two placements are appropriate for this improvement? (Select 2.)

• AThe instructions of the project her team uses for incident work.
• BA saved template the team can start from for each incident.
• CThe organisation instructions, so every chat in the company follows it.
• DA message to Anthropic support describing what worked.
• EHer own chat history, which colleagues can search.

Question 7 · 7.3

Six analysts each produce a monthly client report. Every month, each of them uploads the same methodology note, glossary and template into a new chat, then re-explains the report’s structure before asking for a draft.

What is the most effective optimisation?

• AAsk each analyst to write a longer, more detailed first prompt.
• BPut the methodology, glossary and template into a shared project.
• CHave all six analysts work in one long shared conversation.
• DSwitch to a model that offers a substantially larger context window.

Question 8 · 7.3

A programme manager has kept one conversation open since March for everything to do with a transformation programme. Answers have become vaguer, and she notices she is reaching her usage limit sooner than colleagues doing similar work.

Which two statements best explain what is happening? (Select 2.)

• AA very long thread carries every earlier wrong turn alongside the useful material.
• BConversations that trigger automatic context management consume more of the usage limit.
• CUsage limits are calculated per conversation, so one long chat is penalised.
• DClaude forgets everything older than a fixed number of messages.
• EStarting a new chat would lose all the programme context permanently.

Question 9 · 7.3

A team has cut its tender-response workflow to a single message with no attachments and no review pass, and is pleased with how quickly answers now appear.

What is the strongest objection?

• ASingle messages always produce worse writing than multi-turn conversations.
• BTenders are high-value work that earns a review-and-refine pass and verification.
• CThey should attach more documents to make fuller use of the context window.
• DRemoving attachments breaches the usage policy for business documents.