- Almost Intelligent
- Posts
- AI and Communications Skills
AI and Communications Skills
AI doesn't replace the need to communicate effectively. If anything, it not only taxes your baseline communications skills but taps into any innate managerial or didactic instincts.
Connie Connors
September 29, 2025 • Estimated Reading Time: 5 minutes
Say what? I’ll use AI to write “it.”
One assignment for my college course on Public Relations and Storytelling is to write a press release using an LLM of choice (ChatGPT, Gemini, CoPilot, Grok, Claude, or other) and then edit the press release. I love this kind of assignment because the learning outstrips the feeling that you’re learning. The first struggle is often the prompt. What are you asking of the LLM, and how much information are you giving it? Garbage in, garbage out? Of note, LLMs are good at knowing the structure for almost any writing form. In general, press releases are five paragraphs, so it takes the guessing out of that part.
LLMs ingest information across a broad spectrum of sources. Because of this, LLMs sometimes mimic bad habits, such as misspelling and what I call wasted words, which I teach my students to avoid. Keep this in mind in your own AI writing. It’s easier to catch these words when we’re speaking, so I also tell my students to read their writing out loud. |  Sample of wasted words. |
Basic Tips in Prompt Engineering*
Ask for options
Part of the assignment for my students is to provide an overview of their experience. Did they think the AI-generated press release was good? Did they try more than one request, or did they re-frame the request? Instead of waiting for one response to a question, ask for more versions up front, for example, “Give me two responses: one serious and one more lively.”
Tell it what you don’t want
I’ve found that ChatGPT — at least for me — can get quite cheeky, even campy. Again, I reiterate the tone I’m looking for up front.
Give it a limit
Tell the LLM, “Summarize this in under 500 words, or make it into a 90-second speech.”
Comparisons are great
I rely on AI to provide comparisons, like asking a friend or a consultant. “Compare Park City, Snowmass, and Jackson Hole for skiing in 2026 and include prices, availability of Airbnbs, weather predictions, and anything else I haven’t thought of.” And don’t be afraid to ask for a full-blown competitive chart. Downloading a chart makes it easier to share: “here’s why we’re skiing in Snowmass next year.”
Prescriptives: State of AI
Section AI is a relatively new educational entity devoted to teaching employees of larger companies how to adopt and integrate AI. I look forward to its annual AI: ROI, which took place last week. What struck me is that global companies really haven’t progressed in both acceptance and usage of AI.
Only 23% of employees report their company has a formal AI strategy. Most hear vague mandates like "we need to be AI first" without understanding what that means or why it matters.
Most companies still have no clear “why” for using AI, no real plan to measure its value, and employees are too nervous or too busy to use it. According to the conference speakers, the biggest problems aren’t technical — they’re human. The fix? Treat AI rollouts like a product launch. Market it internally, make tools easy to use, and find early champions who get other people excited.
“At $30/month per person, basic AI access costs the same as lunch. The fundamental question isn’t financial return but whether you trust and empower your people with tools that obviously provide value.”
And, most of the speakers agreed, when you get pushback from a friend worried about job loss, the goal isn’t cutting jobs — it’s freeing up brainpower. Basic AI tools are like internet access: everyone should have them, and the real magic happens when you build unique applications on top.
“Prompt engineering* is the new literacy requirement. Writing for machines will become as fundamental as reading and arithmetic. Companies need systematic training in prompt writing and context engineering, not just AI tool access. The “curse of knowledge" problem – experts who can’t communicate effectively with less technical stakeholders – is a major scaling barrier.”
And yes, it continues to boil down to communication. And good communication.
Until Next Time
Stay curious and know that you could master prompt engineering, but sometimes, almost AI literate is enough.
Connie
UP NEXT: When, how, or why is AI involved in our everyday shopping behavior?