Can we outsource our content without losing our voice?

You know the content needs writing. You also know you are not going to write it, because you have a business to run. So you hire a freelancer or an agency, and six weeks later you get a draft that is grammatically perfect, says nothing you would say, and could have been published by any competitor with the logo swapped.

That draft is worse than no draft. Buyers notice, and it gives AI assistants nothing to quote. In a Princeton and IIT Delhi study presented at KDD 2024, adding quotations from relevant people raised a source's visibility in AI-generated answers by about 41%, adding statistics by about 34% and citing sources by about 30%. It was tested on a research engine, so treat it as directional. Google's guide to its generative AI features (updated July 2026) says the same in plainer words: commodity content "is often based on common knowledge, which could originate from anyone, and typically adds little unique insight for readers".

Buyers read it the same way. In the 2025 Edelman and LinkedIn B2B Thought Leadership Impact Report, a survey of US executives, 64% of the "hidden" decision makers who shape a purchase said they trust thought leadership more than marketing materials when judging what a supplier can do.

You do not have to write everything yourself. You do have to give whoever writes for you three things they cannot invent: what you think, what you have seen, and a named expert who checks the draft. Here is how, in a few hours a month.

#Why does outsourced content so often sound like nobody?

Because the writer was given a topic and nothing else. "Write 1,500 words on cash flow forecasting for construction firms" is a topic, not a brief. A competent writer will read what everyone else has published and produce a tidy summary of it. That is the generic article, and it is what you asked for.

What the writer did not have was your view. For example, you may think most construction firms forecast monthly when they should forecast weekly, because you have watched clients run short of cash between month-end reports. You may remember a firm that nearly failed because one main contractor paid late for four months. None of that is on the internet. Unless it reaches the writer, it will not reach the page.

Voice is also more than word choice: it is which trade-offs you admit, which examples you reach for and how blunt you are. And incentives differ: a freelancer's job is to get a draft approved, while yours is to be named when a buyer asks an assistant who knows this subject. What makes ChatGPT treat content as expert rather than marketing covers what that takes: first-hand experience, specific claims and a named author.

#What should a brand voice guide for a B2B firm include?

A plain statement of what your firm thinks and how it talks, three to five pages long, written by a partner. It pays back every time a piece is written, whoever or whatever drafts it.

• What you believe about your field, in three or four sentences.
• How you sound, shown with sentences you like and sentences you would never publish.
• How specific you are: numbers, real situations, worked examples.
• How you handle trade-offs, including when the answer is "it depends".
• Who the reader is: a busy, sceptical managing director with no patience for jargon.
• What to avoid: subjects, claims, competitors and words.
• Who signs off each subject, so the writer knows whose view to follow.

Example: a page like this, with your own sentences in it, tells a writer more than any list of adjectives.

In our experience, when this document is delegated to a marketing coordinator it comes back as a description of what the firm would like to sound like, which is different from what it thinks. Two hours of a partner's time, once, is the cost.

#How do you brief each piece so it comes back sounding like you?

Give the writer the buyer's question, your answer in two sentences, the point where you disagree with the usual advice, a real example and the length. Do not give them an outline: provide the substance and let the writer find the form. A brief that works fits on one page; the examples are illustrative.

Length should follow the question; see how long an article should be for ChatGPT and AI Overviews. Then add the most valuable input of all: a twenty-minute recorded call in which the writer asks you the question and you answer it out loud. In our experience writers get more usable material from twenty minutes of you talking than from any written brief. The recording is the source. The brief is the frame.

The same discipline works on service and sector pages. When we rebuilt ID-Pal's industry pages, the case study records: "Five industry pages rebuilt: fund administration, wealth management, legal, accountancy and estate agency. Each opens with the question a compliance lead types, names the regulator, and answers inside the first ninety words." The ID-Pal case study also reports: "Where a page has been rebuilt, ID-Pal appears. Where it has not yet, it does not, which tells us exactly what to build next."

#Where can AI tools help, and what must stay human?

AI tools can do the mechanical steps: transcribing the call, turning a transcript into a first draft, and tidying grammar. The angle, the examples and the final check for accuracy must come from a person who knows the subject, because a model knows what the internet says about your field and nothing about what you think.

Will Google penalise content drafted this way? Not for using AI as such. Google's guidance on AI-generated content (February 2023) says "Appropriate use of AI or automation is not against our guidelines" and "Using AI doesn't give content any special gains. It's just content." The line it draws is purpose and scale: its scaled content abuse policy, announced in March 2024, covers many pages made mainly to manipulate rankings "no matter whether content is produced through automation, human efforts, or some combination of human and automated processes".

The bigger risk is sameness: an AI draft made without a voice guide and a transcript is the most generic version of the piece there is, as our guide to AI-written and keyword-led content shows.

#Who should review the draft, and how many rounds should we expect?

The expert first, for substance. The editor second, for polish. Expect two, sometimes three, rounds for the first few pieces, and fewer once the writer knows you.

The expert reads the draft and asks: is this accurate against what we see with clients, did the writer use our angle, does it sound like us, and does every claim have an example, a reason or a source? If not, the piece goes back with a note that names the problem plainly, such as "the section on retentions needs to say what happens when the main contractor disputes the final account." Only then does the editor tidy it.

Make the expert's part cheap: thirty minutes reading with a pen, saying what is wrong rather than rewriting. It gets much quicker after half a dozen pieces. In our experience, twelve excellent pieces a year, each answering a question a buyer asks, do more for whether a firm gets named than twenty-four that are nearly right.

#Should we hire a freelance writer, a specialist or a content agency?

It depends on the piece, and most firms end up using two sources. The cost column is relative, and the revision rounds are what we typically see at the start, not a rule.

A freelancer suits a firm with someone in-house to brief, chase and edit; an agency suits a firm that wants volume and has nobody to manage writers. Every option needs the same inputs: your voice guide, your briefs and your expert's review. If an agency says it needs none of that, you have your answer.

#How do you ghostwrite a founder's LinkedIn posts without them sounding fake?

Use the same inputs as the articles, and have the founder approve every post. Posts sound fake when a writer invents opinions, anecdotes or enthusiasm the founder does not have. They sound like the founder when each one starts from something the founder said out loud that week.

LinkedIn matters for AI search as well as for buyers. A Semrush study of 89,000 LinkedIn URLs (March 2026) found LinkedIn in about 11% of AI answers across ChatGPT search, Google AI Mode and Perplexity. In ChatGPT search and AI Mode, 59% of LinkedIn citations came from individual members rather than company pages, and around three quarters of the authors of cited posts had posted more than five times in the previous four weeks.

#Step 1: Record the raw material

Fifteen to twenty minutes a week: what happened with clients, what the founder disagreed with, what they would tell a peer.

#Step 2: Draft from the transcript

One post per idea, in the founder's own words, drafted by the writer or by an AI tool given the transcript and the voice guide. No invented stories.

#Step 3: Have the founder approve and answer

The founder changes anything they would not say in a meeting, approves every post, and answers the comments personally. That is where conversations with buyers start.

Founder LinkedIn is part of our Reputation That Wins service. If you keep meaning to post and never do, this page describes the pattern.

#How do we stop the voice drifting over time?

Every three to six months, read the last six or eight pieces in one sitting. Voice drifts a little at a time towards the writer's habits or the last round of feedback, and nobody notices piece by piece. If it has become more cautious, more salesy or more general, update the voice guide and share it with the writer.

#What to do this month

1. Write the voice guide: a partner, two hours, three to five pages, including the sentences you would never publish.
2. Pick the one buying question you most want to be named for and write your two-sentence answer.
3. Brief one piece with the template above and record a twenty-minute call answering the question out loud.
4. Decide in writing which steps AI tools may do, and name the expert who reviews for substance and the editor who polishes.
5. Book a short weekly call to gather the founder's LinkedIn material.
6. When the piece is live, add its question to a fixed list of buyer questions you check every month.

#Frequently asked questions

#Can we use ChatGPT to write articles in our own voice, and will Google penalise it?

You can use it to draft, and it gets closer to your voice when you give it the voice guide, the brief and the call transcript. Google says appropriate use of AI is not against its guidelines; what breaks its rules is producing many pages mainly to manipulate rankings, however they are written. Have the expert review every draft and check every figure it adds.

#How much does a B2B content writer cost per article in the UK?

It varies widely with specialism, research, interviews and revision rounds, and we have not found a reliable, sourced UK benchmark to quote. Compare quotes like for like: a price per piece should cover reading your voice guide, a briefing call and two revision rounds, and very cheap content usually skips all three. Our own fees are on what it costs.

#Should the article carry the partner's name if a freelancer wrote it?

Yes, if the partner supplied the thinking and approved the draft; most professional writing works this way. Google's guidance on creating helpful content asks whether it is "self-evident to your visitors who authored your content" and encourages accurate bylines. Use the partner's real name, role and a short biography, and keep them consistent on LinkedIn and in the trade press.

#Does the partner really need to be involved, or can marketing handle it?

Marketing can run the process, chase the writer and polish the drafts. The partner needs to write or dictate the voice guide once, then spend twenty to thirty minutes per piece on the briefing call and the substance review. Without that, you are paying for content that says what everyone else says.

#How do we know whether the content is working?

Keep a fixed list of 20 to 50 buyer questions and ask them in ChatGPT, Gemini, Claude and Copilot several times a month in a fresh session, because answers vary from run to run; count a win only when it repeats. Then listen in sales calls for prospects mentioning something you wrote. Finding out whether ChatGPT recommends your firm shows how to make this routine.

Outsourcing the writing is sensible; the thinking has to stay with you. Help with the voice guide, briefing, founder LinkedIn and review is part of Reputation That Wins, or start with a free Growth Review: 30 minutes with David Adams and three fixes in writing.

#Sources