- Key insight: Learn how treating AI like a colleague improves prompt quality and output accuracy.
- Expert quote: Riabova - Soft skills shape model responses; contextual prompts unlock AI reasoning.
- Supporting data: Survey: 86% of 2,000 office workers use "please" or "thank you" with AI.
Source: Bullets generated by AI with editorial review
Employees saying "please" and "thank you" to AI goes beyond good manners — it can also yield better work results.
In a March survey of 2,000 office workers, 86% said they use these phrases when
The study, commissioned by tech education technology company TripleTen and conducted by Talker Research, found that 87% of those who speak to AI as they would a human view its role as similar to a personal assistant. When employees adopt this approach, their comfort level with the technology is increased and they can better train today's increasingly conversational models to produce
"We have to understand not only how to talk to the models technically, like how to prompt it, but also what you trigger in that model as well, because it was trained in behavioral patterns of humans," she said. "This means that if you have good soft skills … you will be able to communicate the right context for the model so that it can actually figure out how to help you get to your goal, because modern reasoning models can do that — they're much smarter in this way."
Read more:
It's a similar mindset to training a new team member, Riabova said. "Whatever model you use, you're integrating it into your process. It's similar to onboarding; I like to think of it as a very smart intern who knows a lot about many things, but knows little to nothing about what your product is, so teaching it in the right way makes you much more powerful in what you can do with it."
More approachable and adoptable
The TripleTen team is currently using Claude, and often refers to the AI technology as if it were a coworker when discussing its work contributions. Interestingly, Riabova said she's noticed that these types of interactions with AI have improved how team members communicate with one another.
"When we began focusing more on learning how to communicate with AI, we started to be clear in how we communicate between each other, too," she said. "Once you figure out that if you provide unclear input to the model, it won't understand you … and the result won't match what you expect, you understand how people feel sometimes when you give them unclear tasks or share unclear context. So we now try to be more accurate in how we communicate within the team, too."
Read more:
Conversational communication with AI can also make it more approachable and easier to adopt, said TripleTen brand director Ksenia Petriaeva, especially when coworkers are encouraged to share what's working for them — as well as what isn't.
"The more you hear your teammates talk about using AI and funny jokes about what outcome they get unexpectedly, or anything like that, this actually helps you adopt it better, and without any fear, because you see others also struggling sometimes," she said.
The more leadership can ease the transition to increased AI use, the better: Recent research from AI platform Writer and research agency Workplace Intelligence shows a lack of AI strategy is
When it comes to proper and successful AI use, employees will be better equipped if they know the best practices for prompting, and human conversation components are part of this, Riabova said. "The focus is in understanding how you can make these tools work for you … and [part of that] is how you talk to models correctly."
