Podcasts genreated by NotebookLM.
EN
DE
NL
Ever feel like you’re going it alone when it comes to crafting the perfect message? Whether it’s a difficult email, a professional introduction, or any situation where your words truly matter, finding the right tone and phrasing can be a challenge. That’s where Large Language Models (LLMs) come in – not to replace you, but to act as your ultimate peer coach and sounding board.
More Than Just a Text Generator
We’ve seen how LLMs can automate tasks and manipulate text, but their true power lies in helping us refine our intent and improve our communication. Think of an LLM as a coach that provides:
- Guidance and direction
- Skill development
- Personalized support
- Accountability
- Motivation and encouragement
They help us be more true to ourselves, facilitating our growth without making us feel like we have to do it all alone. We still do the work, but with an intelligent partner by our side.
The Art of Saying “No” (and Everything Else)
Let’s take a common challenge: saying “no.” As an editor, rejecting a paper that someone poured their heart into is tough. You need to be consistent, professional, and get straight to the point without being harsh. Simply asking an LLM to “write a rejection letter” often results in something bland and impersonal – basically spam.
The key? Don’t just tell the LLM to “do it for me.” Instead, treat it like a coach.
- Define your role and audience: “I’m an editor for a journal, and this submission is out of scope.”
- Provide a draft (even a rough one): This gives the LLM a starting point.
- Specify your desired tone and length: Do you want to be succinct? Encouraging? Both?
By providing this context, you transform the LLM from a simple text generator into a sophisticated tool that helps you explore the “language space.” You can ask for multiple alternatives and have the LLM explain why each option makes sense. This helps you understand the subtle differences in language that convey vastly different meanings, something you might not have time to explore on your own.
For example, a rejection letter could be:
- Direct and professional for high-volume situations.
- More personal and encouraging to maintain a positive relationship.
- Brief and to-the-point when brevity is essential.
You get to choose the option that aligns perfectly with your intention and desired impact.
Beyond Rejection: Shaping Your Voice and Tone
The coaching goes beyond just difficult messages. Imagine you need an introduction for a new team member or a presentation. You can provide existing information (like a resume or biography) and ask the LLM to draft something. Then, you can refine it.
Don’t like the tone? Ask for alternatives! The LLM can suggest options like:
- Inspirational
- Authoritative
- Casual
- Collegial
- Visionary
You can even drill down further. If “visionary” feels too broad, ask for subsets like “inspirational visionary” or “forward-looking visionary.” This iterative process empowers you to articulate your message with greater precision and impact.
Prompt Engineering: An Agile Journey
This process of refining your requests and exploring options with an LLM highlights a crucial point: prompt engineering is inherently agile. Just like in agile development, where you iterate and refine based on feedback, interacting with an LLM is a continuous loop of:
- Providing initial input (your “user story” for the communication).
- Receiving a draft (your “sprint deliverable”).
- Reviewing and providing feedback (your “retrospective”).
- Iterating with more specific instructions (your next “sprint goal”).
You’re not just writing a single prompt and expecting magic. You’re engaging in an iterative dance, constantly checking the results against your intent and adjusting your “requirements” (the prompts) until you achieve the desired outcome. This also means you need to critically assess what the LLM produces. If something is incorrect or doesn’t fit, call it out! The LLM learns from your guidance.
The Crucial Takeaway: Be You, Empowered by AI
The most important aspect of using LLMs as a peer coach is that you remain in control. The output should be something you’re proud of, something that truly expresses your style and intent. You provide the substance, the direction, and the final assessment.
As our instructor wisely puts it, if you’re not comfortable disclosing that you used an LLM to help craft something, you probably shouldn’t be using it that way. Transparency and integrity are key.
So, next time you’re facing a communication challenge, remember to use AI as your coach, your sounding board, and your partner in exploring the vast possibilities of language. It’s about empowering you to communicate exactly what you want, with clarity and impact.
Some Information about this post
This post is based on the transcript of a video from the Vandebilt course about Integrating Generative AI into Project Management
This course is really interesting and egangng also because Professor Benett Landman has an engaging way, to explain topics . It’s really fun 🙂 Part of agile is sharing, and I like to share my new learnings on this blog. Sharing is caring 😉
I copy the transcript in Google Gemini and ass some extra requirements in the prompt. During this current video, it came up to my mind, that prompting is an iterative task and therefore agile. I have attended a Vanderbilt Course about Pronpt Engineering in which Professor Jules White has been my teacher. I also wrote in which tone I want my post to be. In this case, I wrote enganging and detailed.
I have written several postings about this topic and two came up to my mind, when I was listening what was explained. Here are the two relevant posts:
https://robvanlinda.digital/prompt-engineering-is-agile/
https://robvanlinda.digital/prompt-engineering-compared-with-requirement-engineering/
I copied both URLs into my prompt in Gemini, together with the transcript, and I asked Gemini to be so kind to write an enganging post based on the information I delivered. After the post was generated, I bought an image on Adobe Stock and I inserted the text and uploaded the photo in a new post in my WordPress backend. Then I opened Google NotebookLM and generated a podcst, based on the post text.
This is just a demonstration of my way how I use the advantages of AI tools. Like a Facilitator during a workshop about AI once said “Be lazy, but effective and productive” 😉
