At O’Brien Media, we often use AI tools to help draft and polish emails more quickly as well as supporting team members, and customers, who may need assistance with understanding messages or need further details, or summaries, of content in emails and online.

While these technologies can jump-start your writing and keep your brand voice consistent, they don’t always understand the subtle emotional tone or the full context of what you’re trying to say. That is why taking a moment to read and review every AI-generated draft is so important.

Even the best AI can sometimes misread sarcasm, miss cultural nuances, or introduce awkward phrasing. It may also include proprietary details or personal information that shouldn’t be shared outside your business. By carefully reading each sentence, you can ensure that your message truly reflects your intent and that nothing sensitive slips through, and that you don’t unknowingly agree to something you didn’t intend to.

When you review an AI-assisted email, think about how the recipient will perceive your tone. Ask yourself if the message sounds as warm, professional, or persuasive as you intended. This is especially crucial when discussing sensitive subjects like technical support requests, complaints, or contract and legal matters. A quick read-through can help you catch unintended implications before they become problems.

When you review an AI-assisted email, think about how the recipient will perceive your tone. Share on X

It’s also essential to double-check all facts and figures. Verify that names, dates, statistics, and links are correct and up to date. An AI might pull in an outdated number or the wrong date, and a human eye is your best defense against those kinds of errors. AI has even been known to “hallucinate” where it makes up facts, figures, and even people and quotes – so careful review is essential.

Please don’t be offended if we ask you to confirm something in your own words – we use AI detection so that we can take a closer look at emails that might contain content created by an AI and we may want to check that what’s written is what you intended to convey.

Similarly, please don’t assume we will double check everything you send us, we’re only human after all and in a lot of cases AI can be instructed to mask that it’s creating content so we may not always know the words aren’t yours if you send us messages written by an AI.

Our position is that we consider instructions sent by you to be yours, regardless of the technology used to compose them.