HomeBlogAutomationPhrases that give away a bot

Phrases that give away a bot

Читать по-русски

The wordings that instantly mark a message as automated, and what to replace them with so a conversation reads as human.

Phrases that give away a bot

There is a set of phrases after which the reader stops reading. They are not bad in themselves — people simply do not write them in chat.

Misplaced politeness

«Thank you for your question», «great question», «thanks for reaching out». Fine in a formal letter, wrong in a chat: people do not thank you for a question, they answer it.

Empty connectives

«It is worth noting», «it should be emphasised», «it is important to understand», «as a rule». They carry no meaning and take up space — the signature of text generated to fill a length.

Stock endings

«Hope this helps», «feel free to ask if you have questions», «happy to help». The conversation ends there: the reader has understood they are talking to a script.

Excessive completeness

«There are several approaches to solving this task» instead of «depends on volume». In chat people are shorter and more specific.

Officialese

«Utilise», «aforementioned», «in connection with», «at the present moment». Almost absent from live speech.

What to use instead

A direct answer from the first word. Instead of «thanks for the question, it all depends on the situation» — «depends on volume: under a thousand, one is enough». Shorter, more specific, and it does not read as a template.

Put the list into the prompt as a ban — models avoid what is named. How to phrase that is in the prompt piece.

Topic: Automation

More on this topic: The tool for this

Read next

Writing prompts that do not sound robotic

What to add to a model instruction so replies read as human: length, unevenness, banned clichés and rules about emoji.

How to answer «are you a bot?»

Why people ask, why both obvious answers are wrong and how to run a conversation so the question never comes up.

When AI comments come out off topic

Why a model writes generic filler instead of answering the post, and the three prompt changes that fix it fastest.

Ready to put this to work?

Neurogram automates parsing, commenting and DM conversations on Telegram. Three days free — no card required.

Try it free