The human factor – why AI can’t fact-check an engineer and tell a story

We’ve all seen the headlines: “AI is coming for the copywriters”. I’ve previously discussed how AI can be a powerful tool for productivity. It’s great at summarising meeting notes, proofreading copy (in British English – please!), or helping to brainstorm new ideas if your brain is suffering from “blank page” syndrome.

But the threshold where AI hits a wall is facts. In my world of technology, complex engineering and precision manufacturing, close enough is simply not good enough, and accuracy and in-depth insight is required.  

This is why the human factor is still so important for creating compelling PR content and effective storytelling. The ability to sit across from a lead engineer and craft what they get up to in their day-job into an interesting thought leadership piece that will have media appeal is more valuable than ever.

It’s behind you!

AI looks through a rearview mirror as it uses existing internet data. It has never stood on a factory floor, seen products being tested, or visited a datacentre.

When I speak to a client, I’m not just looking for written content, I dig out the story that isn’t written down yet. AI can’t sniff out a story in an EMC chamber or on the factory floor. It can’t spot the prototype that the engineers are very excited about, but which they haven’t even thought to mention to the marketing department.

PR in the technical sector isn’t about output, it’s about observation. It’s this human premium that connects the dots between something that a client may mention in passing and the topics editors are focusing on this year. LLMs can see what’s been, but can’t see what’s coming.

If you ask an AI to explain a technical patent, it will give you a confident and readable summary, but it might be completely wrong. LLMs predict the next most likely word in a sentence, but they don’t understand fluid dynamics, thermal conductivity or the nuances of certification. If an engineer hands me a technical white paper, my job is to make it sound PR-friendly, but at the same time ensure that it remains accurate and the technical integrity stays intact.

An engineer will respect a PR pro who asks questions: “Wait, if this tolerance is that tight, how does that affect the assembly time?” They will not respect a press release generated by a bot that confuses microns with millimetres. To fact-check an engineer, you must speak their language. You have to be able to look at their datasheet and say: “Explain this to me like I’m a customer (or your Mum) and not one of your technical colleagues.”

So what?

AI is brilliant for the “what” as it can describe a product or service perfectly, but it fails when it comes to the “so what?”.

It doesn’t ask the questions from the end-user perspective – why does this specific valve matter to a buyer in the oil and gas sector right now? How does this software update solve a particular pain point in the supply chain?

Technical PR is about not being afraid to ask questions, while taking raw, often overwhelming information from an engineering team and refining it into the language of end-user benefits – answering the “so what?” You must also be able to use engineering judgement when information is presented to you, rather than take everything at face value.

I guess you could say that while AI provides the volume, humans provide the value. The truth is in the details, and the details are still a human speciality – for now at least!

(And I did ask AI to proofread this for me, but I didn’t agree with all its suggestions 😊)