Hi Team,
I know many of you are using ChatGPT — heavily. Perhaps you’ve noticed that, when compared to other conversational AI tools, ChatGPT has a proclivity to end every response with a question:
Should I run an A/B test on both these responses?
How about I add a border to that chart?
If you want, I can pull a few local options that may be more relevant?
These are what ChatGPT refers to as a conversational nudge—a simple hand-off designed to keep the energy of the conversation alive and prevent the thought that caused it from feeling closed.
For the sake of this article, let’s assume ChatGPT isn’t just trying to keep us on their platform forever—away from sunshine and eternal hope—but wants to ensure it fully answers our prompts.
Let’s analyze when and if these conversational nudges are something we should bring into our own communication practices.
Are we conversational colleagues or agents of action?
ChatGPT is conversational AI—it waits, responds, waits some more. It's designed to keep talking because talking is all it can do. But in the real world, we have two types of collaborators:
Conversational collaborators: Love to discuss, explore, and process everything together. Valuable for complex problems, but can slow things down.
Action-oriented collaborators: Give them direction and they execute. Less chatty, more results.
Some people are conversational collaborators who need to talk through everything, while others are action-takers.Their conversations and how they end them, often reflects their style.
How to end a conversation
ChatGPT is a conversational collaborator. It has no place to be. Its only purpose is to wait for someone to tell it what to do and to learn from that interaction. It will converse with you forever if you allow it—but conversing isn’t always collaborating. Any two people with five minutes to spare and nowhere to be can keep a conversation going. But these people aren’t collaborating, they're talking.
Good communicators and leaders don't just carry on. They prioritize the relevant conversations. They cut the cord on those that are likely to have little impact. They know how to properly end a conversation.
Most of our work conversations don't end because we've reached some natural conclusion. They end because Kevin needs the conference room, or it's 5 PM, or both. That makes how we close these conversations even more critical.
Good leaders don't just check if everyone's satisfied or find ways to exit—they call the next play. They don't ask if you have questions; they make sure the right questions get answered.
To prompt or not to prompt
We’re training ourselves to run nearly every question and assertion through AI prompts. This may seem fine on the surface—after all AI has nowhere to be. But if it's creating a feedback loop where we become accustomed to external validation for decisions we're actually capable of making independently, it may be worth reconsidering.
Ask yourself if the proclivity to approach an AI prompt on nearly everything is drifting into your collaboration with physical members of your team:
Do you consistently need input on decisions you're supposedly equipped to make?
Are you consistently the one initiating discussions?
Are you genuinely lacking resources or insight, or are you just paralyzed by the idea of making a call solo?
Would you have arrived at the same outcome without all the prompting? If yes, your collaboration might actually be consulting, which is usually more one-sided.
Next Steps
To ensure your interactions with AI aren’t impacting your ability to be a good communicator, here are some steps you can take starting tomorrow:
Come up with some of your own conversational nudges, knowing the intent isn’t for a person to continue to interact with you but to keep the collaboration open ended, don't require an immediate response, and create permission for the person to come back with new information. A few examples:
See what questions this brings up as you sit with it…
Let me know if the data starts telling a different story…
Let this idea breathe for a bit and see what else emerges...
Track Your "Input-to-Action" Ratio: Tally up how many times do you ask for input versus how many times you just make the call and move forward? There's no perfect ratio, but if you're asking for input on 80% of your decisions, you might be over-collaborating. If you're making unilateral calls 90% of the time, you might be missing valuable perspectives.
Develop an openness test: Before you create an opening for collaboration, ask "Am I actually prepared to hear something that might change my mind, delay my timeline, or complicate my plan?" If the answer is no, don't feign teamwork and act like you’re creating an opening.
Good collaboration is inherently open ended, but it is not vague or compulsory. Practice brevity. Tell me where you’re going, what you need, and who it impacts.
I hope this newsletter impacted you. If you feel like it can be improved, please use the survey below or email me here to let me know what you felt this piece would be about.
As Ever,
Paul Nyhart
Chief Connection Officer - Hi Team
If this was forwarded by a friend you can subscribe below for free👇