Tempted To Delve Into AI Prop Writing?

Over the Summer marking season, I noted a Uni philosophy lecturer had found a pattern in too many of the papers he was judging.

Exeter's Dr Edward Skidelsky had found one give-away word;

delve.

In general such output, "is typical of the purple but empty prose that ChatGPT produces", he rightly proclaimed.

Among my first thoughts, were that word would spread, and students simply now order their AI bot not to use said incriminating verb.

Yet that wouldn't remove the overall 'tell', merely that one alone.

For there are both other words, and more systemic betrayers.

Apparently syntax red flags currently include other over-used rarer vocab (crucial, enhancing, exhibited, potential, showcasing, underscore), mechanical repetition of unusual phrases, and submitted text wholly without grammatical error.

Not to mention the factual mishaps such as incomplete or inaccurate referencing, with the odd hallucination, through to full-blown regurgitation of 'slop'.

I have had sellers proudly show me their AI-fuelled Props, presentations and email prose.

I am yet to see one that passed muster.

I was reminded of this lately when coders I know told me how some interviewers in their domain are insisting on 'No AI' when providing live evidence of their work to a given task. Yet their point was how misguided such stipulation is. If you want to know truly what they can do, then let them show you what they produce with AI as their partner. The winners will blow all-comers away.

I can fully picture that.

A key issue with AI-bot speak is that it is not in your voice. You may get output which is not 'wrong' as such, but it is likely bland and, at some stage, a prospect would notice.

We can also learn from the recent mass culling of fake reviews by Google and Tripadvisor through identification of AI red flags.

Ensure your voice comes through. Go into the 'nitty gritty details'. Sprinkle liberally with language the prospect themselves uses.

Conduct these health checks of any AI-assisted output and maybe then you'll use your magic helper to your advantage.

There's also another angle. What you might call SDR; sales data reuse.

There's much collateral you may on occasion supply those whom you pitch that is fairly standard.

Whether confirmation of a delighted client experience, resource expectations and service promises.

You can definitely fine-tune this through conversing with your bot of choice.

I know plenty who chat away when they must write a difficult email, for instance.

It not only saves you time in the now, but also eradicating the wasteful mental exertion and prevarication that the hanging, uncompleted task infects you with. Apparently a concept already known to researchers. As they've found that the mere act of starting on sorting such task makes us feel better from the off, and 'talking through' what you want becomes a kind of burden-lifting act.

I recently wrote a tricky mail, all by myself. In which from my draft I swapped out the word 'exasperated'. Deliberately replacing it with another because I specifically did not want the whiff of AI to foul my message. The word I decided that AI would never use? Miffed.

Make sure your prospect feels the polar opposite of said emotion.