AI at Work

Writing Someone Will Sign

AI can produce a fluent first draft. The signed version still needs your voice, your facts, and a reviewer who will put their name on it.

Keywordsai for professional writingBlogai for workai at workworkplace aiai for professionalshow to use ai at workai productivity

Related searcheschatgpt for professional writinghow to edit ai drafts at workai writing you can put your name onprofessional email drafts with aihow to review ai generated writingai for professional writing

Fluency is not a finished document

Professional writing used to be slow because the writer was also discovering the argument. Language models remove that delay. They will give you a structured draft in the tone you asked for, with headings, a recommendation, and a closing that sounds like a person who has done this before. That is useful. It is also the main risk. A fluent page is easy to send. A signed page is a commitment.

The work is therefore not “get AI to write it.” The work is to separate the draft from the signature. The draft can be fast and imperfect. The signature cannot. Emails, memos, board notes, client letters, policies and performance language all follow the same split, with a higher bar as the audience gets more external, more regulated, or more personal.

If you would not put your name under a sentence, do not let a model put your name under it either. Tone can be borrowed. Responsibility cannot.

Brief the draft the way you would brief a colleague

A weak prompt produces generic prose: “I hope this email finds you well,” a restatement of the obvious, and a call to action that could apply to anyone. A professional brief names the reader, the decision you want from them, the facts they may use, the facts they must not see, and the tone that fits the relationship. Role, task, context, constraints, output and quality checks still apply. Writing is just the output format.

Role: You are drafting for a director who will sign this note.
Task: Write a first draft of a client update from the bullets I will paste.
Context: The client already knows the project. They need status, risk, and a decision.
Constraints:
- Do not invent dates, fees, or legal positions.
- Do not apologise unless I have said we were at fault.
- Keep it under 400 words. No throat-clearing.
- Mark any sentence that is an inference rather than a fact in the bullets.
Output: Draft + a short list of claims that need a source.
Quality checks: Read it as a sceptical client. List what you might have got wrong.

Paste the facts. Do not ask the model to remember last quarter. If a number is not in the brief, it should not appear in the draft. That single rule prevents most of the confident errors that survive because they read well.

Edit for voice, not for polish

The first pass after generation is not copy-editing. It is ownership. Cut the sentences you would never say. Restore the specific noun the model replaced with a category. Put back the hedge that is honest (“we do not yet know”) and remove the hedge that is cowardly (“it might perhaps be worth considering”). Check every name, date and figure against the source, not against the draft’s own fluency.

Kind of writingWhat the model may draftWhat you must still own
Internal updateStructure, recap, suggested next stepsPriority, politics, and what not to put in writing
Client or customer letterClear status and optionsCommitments, tone, and anything that sounds like advice
Board or leadership packNarrative skeleton and optionsThe recommendation and the numbers
Policy or legal languageA first cut against a templateThe actual rule, exceptions, and sign-off
People languageNothing that judges a named personPerformance, hiring and discipline stay human

Do not use a model to write a performance review, a disciplinary letter, or a characterisation of a named colleague as if the output were an observation. It will produce fluent judgements from incomplete notes. That is not a draft. That is a risk.

A review pass that stays shorter than the original writing time

Ask the model, before you start reading, to list claims that are inferred rather than stated, and to produce a short “what I might have got wrong” note. Compare the draft to a checklist that existed before the model: audience, ask, facts, risk, next step. If you only compare the draft to itself, you will approve polish. The checklist is how review stays cheaper than writing from a blank page without becoming theatrical.

When the piece will be forwarded, assume the forward will strip context. The signed version has to stand alone. That is another reason the model’s helpful recap of “as we discussed” is often the first thing to delete.

What to do this week

Take one document you will send under your own name. Brief the draft properly. Generate once. Edit for voice and facts before you edit for commas. Keep the prompt if it earned its place. If you cannot name the human who will be accountable when a sentence is wrong, you do not yet have a writing process. You have a way of producing text without an owner.

FAQ: Writing Someone Will Sign

Common questions about this page.

What is the StudyGrid blog?

The StudyGrid blog covers using artificial intelligence for productivity, data analysis, decision-making, and business transformation. Each essay includes frameworks, charts, and professional prompts.

Who is the blog for?

It is written for professionals who use AI in knowledge work: managers, analysts, operators, and specialists who must combine human judgement with model output. You do not need to be a machine-learning engineer.

How should I read the blog essays?

Start at The AI Opportunity and follow Next in order, or open a single essay if you need a briefing on prompting, hallucination, RAG, agents or governance.

Does the blog replace the Vibe Coding course?

No. The blog is about using AI across knowledge work. Vibe Coding is the software-building playbook. Read the blog for judgement, prompting, and governance. Open Vibe Coding when you want to ship code with an agent.

Is the blog free?

Yes. The full blog on StudyGrid (studygrid.in) is free. Open Blog from the header and follow Next through the essays.