What is the best prompt to write a client status report from email?

The five-section prompt in this guide is the most reliable structure. It asks the model for progress by workstream, risks with severity, decisions made, items awaiting client action, and next steps with owners and dates. The two rules that matter most are 'use only facts from the threads' and 'write None found if a section is empty.' Both stop the model from filling a section with plausible-sounding content that the threads do not support. Paste the threads oldest-message-first and run on any capable model: GPT-4o, Claude Sonnet, Gemini 1.5 Pro, or Copilot on a work account. Treat the first output as a draft, not a finished document — verify the risk section manually before you send.

How do you write a status report when the email threads are too long to paste at once?

Chunk by topic rather than by date. Group progress threads together, risk threads together, and decision threads together. Summarize each group separately using the relevant section of the prompt — progress only, then risks only — and then merge the section outputs into a single report in a final pass, asking the model to reconcile any conflicts. This is more work than a single paste but produces a more accurate result than hoping the model holds context across a very long input. Trim repetitive quoted history from each thread before chunking — removing the same block quoted under every reply can halve the length without losing a single fact.

What should a client status report include?

The five sections that matter most are: what moved forward or was blocked (progress), what could affect the project negatively (risks), what was agreed (decisions), what the client still needs to provide or approve (awaiting client), and what happens next and who owns it (next steps). Monthly or milestone reports may warrant an executive summary at the top — a two-sentence statement of overall project health — which you can add by asking the model to prepend it after generating the five sections. For weekly reports, the five sections alone are usually enough.

Is it safe to paste client email threads into a consumer AI tool?

Treat it as a disclosure decision, not a convenience. When you paste a thread into a consumer AI tool, the content leaves your device and is processed on a third-party server. On free and consumer tiers, your inputs may be used to train the model unless you disable that setting — and the control to do so is easy to forget under time pressure. For threads that contain client names, financial details, or contractually sensitive language, use an enterprise or team tier, which typically carries stronger data-handling commitments, or redact the sensitive content before pasting. The cleanest option is a tool that processes threads inside your own mailbox so nothing is copied out at all.

How do you add a risk rating when the thread does not state one explicitly?

The model assigns severity based on the language in the thread. If the thread says 'urgent' or 'this is blocking us,' it tends to rate the risk High. If it says 'something to watch' or 'a minor concern,' it tends to rate it Medium or Low. When you want to override the model's read — because you have context the thread does not contain, from a call or a conversation — edit the severity in the output directly and note the reason in the mitigation column. Adding 'rate as High only when the thread signals urgency explicitly' to the prompt also reduces the most common problem, which is overstated severities rather than understated ones.

How do you handle a project risk that was discussed verbally and never appeared in email?

Add it manually after the model produces the draft. The model can only extract risks from the threads you paste; anything discussed in a call or a meeting that never made it into email is invisible to it. Before running the prompt, scan your call notes for any outstanding risk or decision that did not reach email. Add it as a plain-text note above the thread paste, with a line telling the model it came from a call rather than a thread. The model will include it in the appropriate section and flag it as coming from a source outside the pasted threads.

Can AI Emaily generate client status reports automatically?

AI Emaily is an AI-native email client that reads your project threads inside your mailbox and surfaces structured output — progress, risks, decisions, awaiting items, and next steps — without a copy-paste loop. The draft is ready for your review and approval before anything goes to the client, with an undo and audit trail on every action the agent takes. It works across Gmail, Outlook, iCloud, Fastmail, Proton, and IMAP, and your email is not used to train models. You can start a 7-day free trial at aiemaily.com.

Blog/ AI email prompts & use-cases

AI email prompts & use-cases

AI Prompt: Write a Client Status Report From Email

Nafiul HasanAugust 4, 2026· 11 min read

AI prompt extracting a five-section client status report — progress, risks, decisions, awaiting items, and next steps — from project email threads

The short answer

Paste the relevant project email threads into a capable AI model and ask it to extract progress by workstream, risks with severity ratings, decisions made, items awaiting client input, and next steps with owners. Use the five-section prompt in this guide, paste threads oldest-first, and tell the model to flag anything uncertain.

Turn project email threads into a client status report using one AI prompt — progress, risks, decisions, and next steps extracted in minutes.

On this page

01What do you need before pasting threads into an AI?
02How to write a client status report from email threads with AI
03The five-section status report prompt
04How does this prompt behave on different AI platforms?
05What do you do when the status report comes out wrong?
06A faster way to extract status reports from every project thread

A client status report is only as good as the information behind it. When it arrives vague — 'good progress this week' — the client reaches for the one reply no project manager wants: 'Can you be more specific?' When it arrives with real risks flagged, severity rated, decisions logged, and next steps named with a date, the client feels informed and the project relationship stays intact. The information to write that kind of report already exists. It is in your email.

An ai prompt to write a status report from emails does exactly what the phrase says: it reads the threads, extracts the relevant facts, and assembles them into the five sections a defensible status report needs. This guide gives you the prompt, explains how to prepare your threads so the model gets accurate input, covers how the same prompt behaves across the major platforms, and addresses the most common ways the output falls short and what to do about each one.

What do you need before pasting threads into an AI?#

The model can only report on what you give it. A thin paste of one or two top-level replies produces a thin report; a complete paste of the right threads produces something defensible. Before you open a chat window, do three things: decide which threads belong in this report, expand every thread to show its full history, and decide what to redact.

Choosing the threads is the judgment the model cannot make for you. Pull the chains that touch delivered or blocked work, risks that have surfaced, decisions made with or by the client, and anything the client is still waiting on. Leave out internal-only context the client does not need — a frank exchange about an internal problem, a team member's unfiltered opinion. The model will faithfully include whatever you paste, so anything that belongs only inside your organization should stay out of the paste.

Order matters more than most people expect. Expand all quoted history so the model sees the full conversation, not just the last reply, and arrange each thread oldest-message-first. A model reading a thread newest-first can mistake the latest reply for the starting context and build the report backwards. Finally, treat the paste as a disclosure: the threads leave your device when you submit them, so mask any figures or language that should not travel to a third-party server.

The report is only as complete as the input

If a risk surfaced in a meeting or a Slack thread and never made it into email, the model cannot report it. Before pasting, ask whether the project's real state is fully captured in the threads you have selected. Gaps in the input become gaps in the report.

How to write a client status report from email threads with AI#

The prompt below produces a five-section status report: progress by workstream, risks with severity ratings, decisions made, items the client still needs to provide or approve, and next steps with owners and dates. The steps below explain how to use it reliably.

1
Collect and prepare the threads
Gather the email threads that cover delivered work, emerging risks, decisions, and open client questions. Expand all quoted history, arrange messages oldest-first within each thread, and remove any content that should not leave your organization.
2
Copy the prompt verbatim
Use the five-section prompt in the next section exactly as written, filling in your client and project name. Do not paraphrase the instructions — models follow explicit format rules more consistently than vague ones, and the section headers are what keep the output scannable.
3
Paste the threads and run
Add the threads below the prompt, each separated by an === marker. Run on a capable model: GPT-4o, Claude Sonnet, Gemini 1.5 Pro, or Copilot on a work account. Treat the output as a first draft — not a finished document.
4
Verify risks and calibrate severity
Read the risk section carefully. Models filling a severity field tend toward High when Medium is more accurate. Downgrade ratings that are overstated, add any risk the model missed, and layer in your own read on probability — that is a judgment the model cannot make from text alone.
5
Edit for voice before you send
Raw model output is slightly more formal than most client relationships call for. Read the draft once for tone, cut any phrase that would make a client reach for a dictionary, and add the one or two pieces of context the threads did not contain — a previous verbal commitment, a known client preference, the name of whoever made an informal decision in a call.

The five-section status report prompt#

Prompt: five-section client status report from email threads

ContextYou are a project manager preparing a status report for client: [CLIENT NAME], project: [PROJECT NAME].

TaskRead the email threads pasted below. Extract information for each of the five sections, in order.

Section 1PROGRESS — what moved forward or was blocked this period, by workstream. One bullet per item.

Section 2RISKS — every risk identified. Format: Risk | Severity (High / Medium / Low) | Owner | Mitigation if stated in the threads.

Section 3DECISIONS MADE — what was agreed or resolved. Include who decided and the date where the thread states them.

Section 4AWAITING CLIENT — every item the client still needs to approve, answer, or provide. Include the original due date if one was given.

Section 5NEXT STEPS — what happens next, named owner, and target date. Flag any step that has no owner or no date.

RulesUse only facts from the threads. Write 'None found' if a section is empty. If anything is uncertain, say so inline — do not invent details.

Threads[paste threads here, separated by ===, oldest message first in each thread]

How does this prompt behave on different AI platforms?#

The five-section format works on all the major AI platforms, but each model handles the task differently. The table below covers what to expect — and what to adjust — on the four platforms most professionals use. Verify current features and context-window limits against the vendor's own documentation, since both change frequently.

Platform	Strengths on this prompt	Where to adjust
ChatGPT (GPT-4o)	Follows multi-section formats reliably; handles long thread pastes across a large context window	Tends toward High severity; add the rule 'flag as High only when the thread signals urgency explicitly'
Claude (Sonnet or Opus)	Produces clear, readable prose in the progress section; good at flagging uncertain items inline	May ask clarifying questions rather than completing the draft; tell it to finish the report and note uncertainties within it
Gemini 1.5 Pro or Advanced	Strong at locating dates and named commitments across multiple threads in one paste	Output is verbose by default; add 'use bullets throughout, avoid paragraphs' to keep the report scannable
Microsoft Copilot (work account)	Can reference Microsoft 365 emails directly without a manual paste in some configurations	Section structure is looser; restate the five headers explicitly if the output merges sections

Rate severity conservatively

Left to itself, a model tends to fill a severity field with High because it is the most defensible label. Adding 'rate as High only when the thread uses words like urgent, blocking, or at risk' aligns the ratings with what the threads actually say.

What do you do when the status report comes out wrong?#

The output falls short in four predictable ways. Each traces to something specific and has a specific fix.

A flat progress section means the pasted threads were themselves light — weekly check-in emails saying 'on track' with no named deliverable. Pull the threads that contain actual work: the one where the engineer confirmed the feature shipped, the one where the designer shared the revised screens.
An empty or overloaded risk section traces to how risks were phrased in the threads. Risks described as 'something to keep an eye on' often go unclassified; risks described as 'a serious problem' may come out rated High when Medium is accurate. Fix the first with the rule 'flag as a risk anything described as delayed, blocked, or a concern.' Fix the second with 'rate as High only if urgency is stated explicitly.'
A next-steps section with no owners or dates is usually faithful to the threads — the threads did not name owners or dates. Before pasting, scan for any commitment phrased as 'I will handle that' or 'we will follow up by Thursday.' Add the rule: 'infer owner from context where it is obvious; flag as owner TBD where it is not.'
A report that leaks internal-only language means you included threads that should have been excluded. Re-read each thread before pasting and ask whether you would forward it to the client. If not, exclude or redact the relevant exchange before it enters the prompt.

Internal threads are not client threads

The fastest way to damage a client relationship with a status report is to include a thread where your team spoke candidly about a client problem. Before pasting, re-read each thread and ask: would I forward this to the client? If not, exclude it.

A faster way to extract status reports from every project thread#

The steps above work, and they produce a defensible first draft from raw email. The friction is the paste loop itself: identifying the right threads, expanding their history, ordering them, deciding what to redact, copying everything into a chat window, and repeating the whole sequence for every reporting cycle on every project.

AI Emaily is an AI-native email client that runs this process where the project threads already live — inside your inbox, without the paste. The relevant context is assembled from the real conversation; the five-section draft is ready for review and one approval away from sending. AI Emaily uses a user-set Context brain and per-client profiles, so the voice of the draft reflects how you actually write to that client, not a generic register. We build AI Emaily. Start a 7-day free trial at aiemaily.com — no commitment if you cancel before day seven. Pricing and plan details are at aiemaily.com/pricing.

Frequently asked

See it in AI Emaily

AI DraftingReplies written in your voice, not a robot's Smart Search & AskAsk your inbox a question, get an answer

Keep reading

AI Prompt: Write a Weekly Update From Your Sent Mail AI Email Prompts for Project Managers: 20 Prompts AI Email Prompts for Account Managers: 20 Prompts Turn an Email Thread Into a Project Brief With AI

Sources

Written by

Nafiul Hasan

Nafiul Hasan is an entrepreneur and AI automation system builder with 10+ years of experience turning messy, manual workflows into reliable automated systems. He designs and ships AI enterprise solutions end-to-end — the agent logic, the data plumbing, and the product people actually use — and founded AI Emaily to give busy professionals their attention back. He writes here from the builder's seat: what works, what breaks, and how to put AI to work without giving up control.

EntrepreneurAI Automation System BuilderAI EnthusiastBuilds AI Enterprise Solutions10+ years experience