Paul Maguire

Editorial & Content Specialist

Thought leadership, editorial development, website content and digital publishing support.More than 20 years across energy, commodities, technology and professional services.

• Email Paul

Have a content project in mind?

Send a brief with the intended audience, format and stage the content has reached.

• Submit a Brief
• Email
• LinkedIn
• Upwork

SELECTED CONTENT STRATEGY AND STORYTELLING PORTFOLIO

From Brief to Publication

Executive interviewing, film scripting, thought leadership, customer storytelling and employee communications, with public examples and confidentiality-safe project summaries.

Brief → Narrative → Format → Publication

• Open portfolio PDF, 344 KB, in a new tab

Insights

Reflections on editorial judgement, AI-assisted content, expert interviews and digital publishing.

2. A Transcript Is Source Material, Not an Article — Why expert conversations still need selection, verification, structure and editorial judgement before they become publishable thought leadership.

• Read the Insight

1. Why I started Medialingo — What more than 20 years in editorial taught me about clear content in the age of AI.

• Read the Insight

Have a content project in mind?

Send a brief with the intended audience, format and stage the content has reached.

• Submit a Brief
• Email
• LinkedIn
• Upwork

2. A Transcript Is Source Material, Not an Article

What it takes to turn an expert conversation into clear, credible and publishable content.

• Read the Insight

1. Why I started Medialingo

What more than 20 years in editorial taught me about clear content in the age of AI.

• Read the Insight

Why I started Medialingo

Paul Maguire · July 2026

What more than 20 years in editorial taught me about clear content in the age of AI

I’ve worked in editorial and content for more than 20 years, and almost every part of producing content has become faster during that time.Interviews can be recorded without specialist equipment. Transcripts arrive within minutes. AI can summarise a long conversation, suggest an angle and produce a reasonable-looking draft before the meeting has properly finished.I use those tools myself. It would be odd not to.What they don’t do is remove the need for someone to make sense of the material.A transcript can be accurate and still not contain an article. An AI draft can read well and still miss the point, flatten someone’s voice or include something that should never make it into the final copy.That gap between producing words and making sound editorial decisions is a large part of the thinking behind Medialingo.

The tools have changed. The questions haven’t

My career has covered journalism, digital publishing, content marketing, website content, thought leadership, podcasts, video and other forms of multimedia.Much of that work has been in energy, commodities, technology and professional services, where the subject matter can be complicated and accuracy matters. Sometimes the audience already understands the subject in depth. Sometimes the challenge is explaining it to people who don’t, without oversimplifying it or talking down to them.The formats and platforms have changed repeatedly, but the basic editorial questions have stayed much the same.What is the actual point? Who is this for? What does the reader need to understand? Is the claim supported? What can be removed? Does the structure help? Does the finished piece still sound like the person or organisation behind it?Those questions applied when I was editing and publishing news. They applied when producing podcasts, videos and webinars. They also applied during a large website rebuild involving several hundred pages.Rewriting and moving the content was only part of the work. Pages had to be rebuilt for new templates, published, reviewed and tested. Links, metadata and the live pages all needed checking before and after launch.It reinforced that content quality is not only about the words. It also depends on systems, ownership and attention to detail.A page can be well written and still fail because it is in the wrong template, contains an old link or has not been checked properly after publication. A migration is not complete simply because the content has moved from one system to another.

An interview is not an article

Some of the work I’ve found most interesting has involved developing articles from conversations with senior leaders and subject matter experts.The starting point might be a structured interview, an informal recorded discussion, a transcript or notes gathered over several conversations.AI can help organise that material. It can identify repeated themes, compare answers and suggest possible structures. Used carefully, it can save time.But people don’t speak in finished articles. They repeat themselves, qualify points, change direction and sometimes make their strongest observation almost in passing.A transcript captures the conversation. It doesn’t tell you which idea is worth developing or whether there is a meaningful argument underneath it.That still takes editorial judgement.You have to understand what the person is trying to say, decide what is genuinely distinctive and work out what the reader will care about. You may need to check a claim, question an assumption or go back with a follow-up.You also need to improve the structure without removing the characteristics that make the person sound like themselves. The aim is not to make every executive or expert sound polished in exactly the same way.I’ve seen AI-assisted drafts that were tidy, grammatical and almost entirely forgettable. They had headings, sensible transitions and the appearance of a finished article. What they lacked was the thing the expert actually believed.That is one of the areas I want to explore through Medialingo Insights. Not whether AI can produce copy, because it clearly can, but what still needs to happen before that copy is ready to publish.

Technology has affected how I work in a more personal way

My interest in useful technology is not limited to content production.I’ve lived with essential tremor since my teens. It affects more than my hands. It can also cause movement in my head and neck and make my voice tremble.For a long time, it had a significant effect on my confidence. Presentations were particularly difficult. Five years ago, the prospect of speaking in front of a room could bring me out in a cold sweat, even when I knew the subject and had prepared properly.Part of the problem was that the physical signs of tremor could look like nerves. I was already anxious about presenting, then became more anxious about how the tremor might be perceived. It was easy to focus less on what I wanted to say and more on what my hands, head or voice might be doing.Remote working helped more than I expected.Presenting through Teams didn’t remove the tremor, but it made the situation more manageable. I could work from a familiar environment, arrange my notes and screen properly and concentrate more on the material.Recordings and transcripts also became useful in practical ways. Taking detailed notes during a meeting can be difficult when your hands are shaking. Being able to return to a recording or work from a transcript reduces the risk of missing something important.An Occupational Health assessment in a previous role helped identify adjustments that could reduce some of those barriers. These included meeting transcription, suitable workstation support, flexibility around office attendance and the ability to present from home where the work allowed.None of those adjustments changed what I knew or what I was capable of doing. They changed some of the conditions around how I was expected to do it.In October 2025, after waiting around six years, I underwent a focused ultrasound procedure to reduce the tremor on my right side.It was a major event for me. I had waited a long time for the opportunity and, when an earlier date became available, I took it.While the treatment made a difference to the tremor on that side, I underestimated the effects of the procedure and the recovery afterwards.It made me wary of simple transformation stories. Technology and medical treatment can improve someone’s life without producing a neat before-and-after. Something can help significantly and still have consequences you did not fully anticipate.The same caution applies, on a much less serious level, to some of the claims made about AI.A useful tool is not necessarily a complete solution.

Where AI fits for me

I’m not against AI-generated content, and I don’t think every sentence that has involved AI is automatically poor.I use AI for research, organisation, transcript analysis, early outlines, comparisons and repurposing. It can help me work through a large amount of source material and identify connections worth examining more closely.It can also be wrong.It can misunderstand context, invent certainty where there isn’t any and smooth out the features that make someone’s voice recognisable. It can produce a confident answer without knowing whether the source material supports it.A detailed prompt helps. So does a good knowledge base. Clear guidance on audience, tone and terminology can improve the result considerably.But those things don’t make editorial judgement unnecessary.Someone still needs to recognise when the argument is weak, when a fact needs checking, when the tone is wrong or when the most important point has been left out.In complex B2B content, the risk is not always obvious nonsense. More often, it is copy that looks professional enough to pass without close attention.

Why Medialingo

Medialingo grew out of this mix of experience.It brings together the editorial work I’ve done throughout my career, my experience of digital publishing and my interest in how technology can support good work without replacing the thinking behind it.I’m particularly interested in the material organisations already have but often struggle to use properly: interviews, transcripts, presentations, webinars, technical documents, draft articles and AI-assisted copy.There may be a worthwhile article hidden inside an hour-long conversation. There may be a clear argument buried in an awkward first draft. A website may contain useful information that has become difficult to find, inconsistent or out of date.The answer is not always to produce more.Sometimes it is to understand the material better, make firmer choices and give more attention to what is already there.Through these Insights, I plan to write about the parts of content work that are often overlooked when the focus is on speed. That includes finding the real idea in an interview, preserving someone’s voice, checking whether an AI-assisted draft is reliable and making sure content still works once it reaches a website.I’ll also write occasionally about accessibility, confidence and the ways technology can remove barriers without pretending those barriers never existed.The tools will keep changing. That is one of the things that makes this work interesting.But someone will still need to decide what matters, whether it is true and whether it is ready to publish.

About Paul Maguire

Paul Maguire is an editorial and content specialist with more than 20 years’ experience across journalism, digital publishing, content marketing, thought leadership, web content and multimedia.

• ← Back to Insights

• Return to Medialingo

Medialingo

A Transcript Is Source Material, Not an Article

Paul Maguire · August 2026

What it takes to turn an expert conversation into clear, credible and publishable content

A good expert interview can contain insight, experience and a distinctive point of view. It can also contain repetition, unfinished thoughts, tangents, untested assumptions and answers that only become useful after the third follow-up question.That is not a flaw in the interview. It is how people speak when they are thinking through a subject properly.The mistake comes later, when a transcript is treated as though it were almost ready to publish. Tidying the grammar, removing a few filler words and adding subheadings may produce cleaner copy, but it does not necessarily produce an article.A transcript records what was said. Editorial development determines what matters.

Find the core argument

An interview rarely arrives with its strongest idea neatly stated at the beginning.The expert may circle around the real point, qualify it halfway through or reveal it in response to a challenge. Several answers may contain different parts of the same argument. Other comments may be interesting but irrelevant to the intended reader.The editor’s first job is therefore not to rewrite sentences. It is to decide what the article is actually about.Consider this simple source comment:“We’ve got around 2,000 pages that need to move.”That sounds like a useful starting point for an article about website migration. But the number itself is not yet the insight.The editorial questions are more important:   • Does the organisation genuinely need 2,000 pages on the new site?
   • Which pages still meet a user need?
   • Which should be revised, consolidated or removed?
   • What does the page count reveal about ownership and governance?The stronger publishable point becomes:   The page count describes the existing estate, not what deserves to survive.The transcript supplied the material. Editorial judgement found the argument inside it.

Check and clarify

Experts know their subjects, but that does not mean every statement should pass directly into publication.A conversation may include:   • figures recalled from memory
   • broad statements that need qualifying
   • technical claims outside the speaker’s immediate responsibility
   • references to research that need checking
   • assumptions that make sense internally but may confuse an external readerGood editing separates experience, opinion and verifiable fact.It also preserves uncertainty where it matters. If an expert says a result depends on the organisation, audience or technical environment, the article should not turn that into an absolute rule simply because certainty sounds more authoritative.This is particularly important in technical, regulated or fast-changing subjects. A polished sentence can still be inaccurate, outdated or misleading.Editorial QA therefore begins well before proofreading. It includes checking the reasoning behind the copy.

Write for the reader

An interview follows the logic of a conversation. An article needs the logic of a reader.The final structure may bear little resemblance to the order in which topics arose. A useful observation from minute 24 may become the opening. The original first answer may disappear entirely. Three separate exchanges may need to be combined into one section.That is not a distortion of the expert’s voice. It is the work required to make the expertise accessible.The editor must decide:   • what the reader already knows
   • what needs explaining
   • which detail earns its place
   • where evidence is required
   • what sequence makes the argument easiest to follow
   • what the reader should understand or do nextThe result should feel natural, but that clarity is constructed.

Preserve the voice without preserving every word

A verbatim transcript is not the same as an authentic voice.Spoken language contains false starts, repetition and shorthand that may be perfectly clear in conversation but awkward on the page. Preserving every word can make an expert sound less articulate than they actually are.At the other extreme, excessive rewriting can flatten the speaker into generic corporate language.The balance lies in preserving:   • distinctive observations
   • meaningful qualifications
   • characteristic turns of phrase
   • the expert’s level of confidence
   • the way they distinguish one problem from anotherThe finished article should sound like the expert on a clear day, not like a cleaned-up recording or a standardised brand template.

Where AI can support the process

AI can help organise a long transcript, group repeated themes and identify possible structures. It can produce a first-pass summary or draft that gives an editor something to interrogate.But it can also remove the very things that made the conversation valuable.It may:   • promote the most repeated point rather than the most important one
   • turn a qualified observation into a general claim
   • smooth away disagreement or uncertainty
   • impose a familiar but unsuitable article structure
   • replace specific language with plausible-sounding abstractionsThe useful model is tool-supported organisation followed by accountable editorial judgement.AI can help process the source material. It should not decide, without scrutiny, what the expert meant, what the audience needs or what is safe to publish.

The article is the result of decisions

Turning an interview into thought leadership is not primarily a transcription task.It involves selecting, questioning, verifying, structuring and refining. It requires judgement about what to include, what to leave out and what needs further evidence.The final article should respect the expert’s voice, serve the reader and stand up to scrutiny.That is the editorial difference between capturing a conversation and publishing something worth reading.

Related work: See From Plausible to Publishable, a research-led case study showing how source material and a plausible first draft were developed into a verified B2B article.

About Paul Maguire

Paul Maguire is an editorial and content specialist with more than 20 years’ experience across journalism, digital publishing, content marketing, thought leadership, web content and multimedia.

• ← Back to Insights

• Return to Medialingo

Medialingo